wmllint

#!/usr/bin/env python3
# -*- coding: utf-8 -*-
#
# wmllint -- check WML for conformance to the most recent dialect
#
# By Eric S. Raymond April 2007.
#
#                              PURPOSE
#
# wmllint is a tool for porting and debugging WML (Wesnoth Markup Language).
# It can do much of the work of upgrading old content to a new version of
# Battle for Wesnoth, automatically.
#
# While the script is at it, it checks for various incorrect and dodgy WML
# constructs, including:
#   * unbalanced tags
#   * strings that need a translation mark and should not have them
#   * strings that have a translation mark and should not
#   * translatable strings containing macro references
#   * filter references by id= not matched by an actual unit
#   * abilities or traits without matching special notes, or vice-versa
#   * consistency between recruit= and recruitment_pattern= instances
#   * unknown unit types in recruitment lists
#   * double space after punctuation in translatable strings.
#   * unknown races or movement types in units
#   * unknown base units
#   * misspellings in message and description strings
#
#                              BASIC PROCEDURE
#
# Takes any number of directories as arguments.  Each directory is converted.
# If no directories are specified, acts on the current directory.
#
# The recommended procedure is this:
# 1. Run it with --dryrun first to see what it will do.
# 2. If the messages look good, run without --dryrun; the old content
#    will be left in backup files with a -bak extension.
# 3. Eyeball the changes with the --diff option.
# 4. Use wmlscope, with a directory list including the Wesnoth mainline WML
#    as first argument, to check that you have no unresolved references.
# 5. Test the conversion.
# 6. Use either --clean to remove the -bak files or --revert to
#    undo the conversion.
#
# Compatibility with pre-1.4 versions has been removed; to port very old
# content, first run wmllint-1.4 to convert to "1.4", before running this
# one.
#
#
#                              MAGIC COMMENTS
#
# Note: You can shut wmllint up about custom terrains by having a comment
# on the same line that includes the string "wmllint: ignore" or
# "wmllint: noconvert". The same magic comments will also disable checking
# of translation marks, which may alternatively be toggled for line blocks
# with "wmllint: markcheck off" and "wmllint: markcheck on".
#
# You can also prevent description insertions with "wmllint: no-icon".
#
# You can force otherwise undeclared characters to be recognized with
# a magic comment containing the string "wmllint: recognize".
# The rest of the line is stripped and treated as the name of a character
# who should be recognized in descriptions.  This will be useful,
# for example, if your scenario follows a continue so there are
# characters present who were not explicitly recalled.  It may
# also be useful if you have wrapped unit-creation or recall markup in
# non-core macros and wmllint cannot recognize it.
#
# If you use custom macros to create (or recall) a named unit, you can tell
# wmllint which field contains the id with the string, "wmllint: whofield
# <macro> <number>".  After the macro's last use, *make sure* to remove it
# with "wmllint: whofield clear <macro>"; if "clear" is not followed by a
# macro name, all macros will be cleared from the list.  (If the <number>
# string was not a non-zero number, wmllint will also attempt to remove the
# specified macro.)
#
# For macros used in multiple scenarios to field characters, you can tell
# wmllint to recognize them with another magic comment:
#     wmllint: who <macro> is <character(s)>
# <Macro> should be the macro name; <character(s)> would be the ids of the
# recalled (or created) characters, comma-separated if there were more than
# one.  If more characters attach to a macro in later scenarios, they can be
# appended with additional "who" magic comments (though this requires scenarios
# to be numbered or otherwise sorted to work properly, as wmllint traverses
# files in alphabetical order).  If there is a character that drops out,
# preceding that person's entry with a double-minus will cause that character
# to be removed, e.g., "-- Garak".
#
# At the end of the campaign, *make sure* you use:
#     wmllint: unwho ALL
# This will clear the list of "who" magic comments and tell wmllint to stop
# checking for them in later files and directories.  If a specific macro is
# not used in subsequent scenarios, put "unwho <macro>" in its last scenario.
#
# Similarly, it is possible to explicitly declare a unit's usage class
# with a magic comment that looks like this:
#    wmllint: usage of <unit> is <class>
# Note that <unit> must be a string wrapped in ASCII doublequotes.  This
# declaration will be useful if you are declaring units with macros that
# include a substitutable formal in the unit name; there were examples in
# UtBS, but they have since been converted to use the [base_unit] tag.
#
# If a mismatch between a recruit list and recruitment pattern involves a
# usage type outside the five core types, the warning message will include a
# note that a non-standard usage class is involved.  If you are using custom
# usage classes and would like wmllint to be aware of them, you can insert
# the magic comment, "wmllint: usagetype <class>".  This comment will take a
# comma-separated list, and can also be pluralized to "usagetypes".
#
# You can disable stack-based malformation checks with a comment
# containing "wmllint: validate-off" and re-enable with "wmllint: validate-on".
#
# You can prevent filename conversions with a comment containing
# "wmllint: noconvert" on the same line as the filename.
#
# You can suppress complaints about files without an initial textdomain line
# by embedding the magic comment "# wmllint: no translatables" in the file.
# of course, it's a good idea to be sure this assertion is actually true.
#
# You can skip checks on unbalanced WML (e.g. in a macro definition) by
# bracketing it with "wmllint: unbalanced-on" and "wmllint: unbalanced-off".
# Note that this will also disable stack-based validation on the span
# of lines they enclose.
#
# You can suppress warnings about newlines in messages (and attempts to
# repair them) with "wmllint: display on", and re-enable them with
# "wmllint: display off".  The repair attempts (only) may also be
# suppressed with the --stringfreeze option.
#
# A special comment "# wmllint: notecheck off" will disable checking unit types
# for consistency between abilities/weapon specials and usage of special notes
# macros in their descriptions.
# The comment "# wmllint: notecheck on" will re-enable this check.
#
# A special comment "# wmllint: deathcheck off" will disable the check whether
# dying units speak in their death events.
# The comment "# wmllint: deathcheck on" will re-enable this check.
#
# A magic comment of the form "wmllint: general spellings word1
# word2..."  will declare the tokens word1, word2, etc. to be
# acceptable spellings for anywhere in the Wesnoth tree that the
# spellchecker should never flag.  If the keyword "general" is
# replaced by "local", the spelling exceptions apply only in the
# current file. If the keyword "general" is replaced by "directory",
# the spelling exceptions apply to all files below the parent
# directory.
#
# A comment containing "no spellcheck" disables spellchecking on the
# line where it occurs.
#
# A comment of the form
#
# #wmllint: match {ABILITY_FOO} with {SPECIAL_NOTES_IOO}
#
# will declare an ability macro and a special-notes macro to be tied
# together for reference-checking purposes.
#
# In 1.11.5 and 1.11.6 respectively, the special ellipses for leaders and units
# without a ZoC are automatically handled by the C++ engine. wmllint warns if such
# ellipses are found in your WML and asks to remove them; at the same time it
# notifies of custom ellipses that may need to be updated (updating means that
# -nozoc, -leader and -leader-nozoc variations of your ellipse need to be added).
# You can disable this sanity check for the current line with the comment
# "# wmllint: no ellipsecheck".
#
#                              DEVELOPER INFORMATION
#
# All conversion logic for lifting WML and maps from older versions of the
# markup to newer ones should live here.  This includes resource path changes
# and renames, also map format conversions.
#
# Note: Lift logic for pre-1.4 versions has been removed; if you need it,
# use wmllint-1.4 to lift before running this one.  I did this for a policy
# reason; I wanted to kill off the --oldversion switch.  It will *not*
# be restored; in future, changes to WML syntax *must* be forward
# compatible in such a way that tags from old versions can be
# unambiguously recognized (this will save everybody heartburn).  As a
# virtuous side effect, this featurectomy cuts wmllint's code
# complexity by over 50%, improves performance by about 33%, and
# banishes some annoying behaviors related to the 1.2 map-conversion
# code.
#

import sys, os, re, argparse, string, copy, difflib, time, gzip, codecs
from wesnoth.wmltools3 import *
from wesnoth.wmliterator3 import *

# Changes meant to be done on maps and .cfg lines.
mapchanges = (
    ("^Voha", "^Voa"),
    ("^Voh",  "^Vo"),
    ("^Vhms", "^Vhha"),
    ("^Vhm",  "^Vhh"),
    ("^Vcha", "^Vca"),
    ("^Vch",  "^Vc"),
    ("^Vcm",  "^Vc"),
    ("Ggf,",  "Gg^Emf"),
    ("Qv,",  "Mv"),
    )

# Base terrain aliasing changes.
aliaschanges = (
    # 1.11.8:
    ("Ch",   "Ct"),
    ("Ds",   "Dt"),
    ("Hh",   "Ht"),
    ("Mm",   "Mt"),
    ("Ss",   "St"),
    ("Uu",   "Ut"),
    ("Ww",   "Wst"),
    ("Wo",   "Wdt"),
    ("Wwr",  "Wrt"),
    ("^Uf",  "Uft"),
    # Vi -> Vit in 1.11.8, Vit -> Vt in 1.11.9.
    ("Vit",  "Vt"),
    # 1.11.9:
    ("Vi",   "Vt"),
    )

# Global changes meant to be done on all lines.  Suppressed by noconvert.
linechanges = (
        ("canrecruit=1", "canrecruit=yes"),
        ("canrecruit=0", "canrecruit=no"),
        # Fix a common typo
        ("agression=", "aggression="),
        # These changed just after 1.5.0
        ("[special_filter]", "[filter_attack]"),
        ("[wml_filter]", "[filter_wml]"),
        ("[unit_filter]", "[filter]"),
        ("[secondary_unit_filter]", "[filter_second]"),
        ("[attack_filter]", "[filter_attack]"),
        ("[secondary_attack_filter]", "[filter_second_attack]"),
        ("[special_filter_second]", "[filter_second_attack]"),
        ("[/special_filter]", "[/filter_attack]"),
        ("[/wml_filter]", "[/filter_wml]"),
        ("[/unit_filter]", "[/filter]"),
        ("[/secondary_unit_filter]", "[/filter_second]"),
        ("[/attack_filter]", "[/filter_attack]"),
        ("[/secondary_attack_filter]", "[/filter_second_attack]"),
        ("[/special_filter_second]", "[/filter_second_attack]"),
        ("grassland=", "flat="),
        ("tundra=", "frozen="),
        ("cavewall=", "impassable="),
        ("canyon=", "unwalkable="),
        # This changed after 1.5.2
        ("advanceto=", "advances_to="),
        # This changed after 1.5.5, to enable mechanical spellchecking
        ("sabre", "saber"),
        ("nr-sad.ogg", "sad.ogg"),
        # Changed after 1.5.7
        ("[debug_message]", "[wml_message]"),
        ("[/debug_message]", "[/wml_message]"),
        # Changed just before 1.5.9
        ("portraits/Alex_Jarocha-Ernst/drake-burner.png",
         "portraits/drakes/burner.png"),
        ("portraits/Alex_Jarocha-Ernst/drake-clasher.png",
         "portraits/drakes/clasher.png"),
        ("portraits/Alex_Jarocha-Ernst/drake-fighter.png",
         "portraits/drakes/fighter.png"),
        ("portraits/Alex_Jarocha-Ernst/drake-glider.png",
         "portraits/drakes/glider.png"),
        ("portraits/Alex_Jarocha-Ernst/ghoul.png",
         "portraits/undead/ghoul.png"),
        ("portraits/Alex_Jarocha-Ernst/mermaid-initiate.png",
         "portraits/merfolk/initiate.png"),
        ("portraits/Alex_Jarocha-Ernst/merman-fighter.png",
         "portraits/merfolk/fighter.png"),
        ("portraits/Alex_Jarocha-Ernst/merman-hunter.png",
         "portraits/merfolk/hunter.png"),
        ("portraits/Alex_Jarocha-Ernst/naga-fighter.png",
         "portraits/nagas/fighter.png"),
        ("portraits/Alex_Jarocha-Ernst/nagini-fighter.png",
         "portraits/nagas/fighter+female.png"),
        ("portraits/Alex_Jarocha-Ernst/orcish-assassin.png",
         "portraits/orcs/assassin.png"),
        ("portraits/Emilien_Rotival/human-general.png",
         "portraits/humans/general.png"),
        ("portraits/Emilien_Rotival/human-heavyinfantry.png",
         "portraits/humans/heavy-infantry.png"),
        ("portraits/Emilien_Rotival/human-ironmauler.png",
         "portraits/humans/iron-mauler.png"),
        ("portraits/Emilien_Rotival/human-lieutenant.png",
         "portraits/humans/lieutenant.png"),
        ("portraits/Emilien_Rotival/human-marshal.png",
         "portraits/humans/marshal.png"),
        ("portraits/Emilien_Rotival/human-peasant.png",
         "portraits/humans/peasant.png"),
        ("portraits/Emilien_Rotival/human-pikeman.png",
         "portraits/humans/pikeman.png"),
        ("portraits/Emilien_Rotival/human-royalguard.png",
         "portraits/humans/royal-guard.png"),
        ("portraits/Emilien_Rotival/human-sergeant.png",
         "portraits/humans/sergeant.png"),
        ("portraits/Emilien_Rotival/human-spearman.png",
         "portraits/humans/spearman.png"),
        ("portraits/Emilien_Rotival/human-swordsman.png",
         "portraits/humans/swordsman.png"),
        ("portraits/Emilien_Rotival/transparent/human-general.png",
         "portraits/humans/transparent/general.png"),
        ("portraits/Emilien_Rotival/transparent/human-heavyinfantry.png",
         "portraits/humans/transparent/heavy-infantry.png"),
        ("portraits/Emilien_Rotival/transparent/human-ironmauler.png",
         "portraits/humans/transparent/iron-mauler.png"),
        ("portraits/Emilien_Rotival/transparent/human-lieutenant.png",
         "portraits/humans/transparent/lieutenant.png"),
        ("portraits/Emilien_Rotival/transparent/human-marshal.png",
         "portraits/humans/transparent/marshal.png"),
        ("portraits/Emilien_Rotival/transparent/human-marshal-2.png",
         "portraits/humans/transparent/marshal-2.png"),
        ("portraits/Emilien_Rotival/transparent/human-peasant.png",
         "portraits/humans/transparent/peasant.png"),
        ("portraits/Emilien_Rotival/transparent/human-pikeman.png",
         "portraits/humans/transparent/pikeman.png"),
        ("portraits/Emilien_Rotival/transparent/human-royalguard.png",
         "portraits/humans/transparent/royal-guard.png"),
        ("portraits/Emilien_Rotival/transparent/human-sergeant.png",
         "portraits/humans/transparent/sergeant.png"),
        ("portraits/Emilien_Rotival/transparent/human-spearman.png",
         "portraits/humans/transparent/spearman.png"),
        ("portraits/Emilien_Rotival/transparent/human-swordsman.png",
         "portraits/humans/transparent/swordsman.png"),
        ("portraits/James_Woo/assassin.png",
         "portraits/humans/assassin.png"),
        ("portraits/James_Woo/dwarf-guard.png",
         "portraits/dwarves/guard.png"),
        ("portraits/James_Woo/orc-warlord.png",
         "portraits/orcs/warlord.png"),
        ("portraits/James_Woo/orc-warlord2.png",
         "portraits/orcs/warlord2.png"),
        ("portraits/James_Woo/orc-warlord3.png",
         "portraits/orcs/warlord3.png"),
        ("portraits/James_Woo/orc-warlord4.png",
         "portraits/orcs/warlord4.png"),
        ("portraits/James_Woo/orc-warlord5.png",
         "portraits/orcs/warlord5.png"),
        ("portraits/James_Woo/troll.png",
         "portraits/trolls/troll.png"),
        ("portraits/Jason_Lutes/human-bandit.png",
         "portraits/humans/bandit.png"),
        ("portraits/Jason_Lutes/human-grand-knight.png",
         "portraits/humans/grand-knight.png"),
        ("portraits/Jason_Lutes/human-halberdier.png",
         "portraits/humans/halberdier.png"),
        ("portraits/Jason_Lutes/human-highwayman.png",
         "portraits/humans/highwayman.png"),
        ("portraits/Jason_Lutes/human-horseman.png",
         "portraits/humans/horseman.png"),
        ("portraits/Jason_Lutes/human-javelineer.png",
         "portraits/humans/javelineer.png"),
        ("portraits/Jason_Lutes/human-knight.png",
         "portraits/humans/knight.png"),
        ("portraits/Jason_Lutes/human-lancer.png",
         "portraits/humans/lancer.png"),
        ("portraits/Jason_Lutes/human-paladin.png",
         "portraits/humans/paladin.png"),
        ("portraits/Jason_Lutes/human-thug.png",
         "portraits/humans/thug.png"),
        ("portraits/Kitty/elvish-archer.png",
         "portraits/elves/archer.png"),
        ("portraits/Kitty/elvish-archer+female.png",
         "portraits/elves/archer+female.png"),
        ("portraits/Kitty/elvish-captain.png",
         "portraits/elves/captain.png"),
        ("portraits/Kitty/elvish-druid.png",
         "portraits/elves/druid.png"),
        ("portraits/Kitty/elvish-fighter.png",
         "portraits/elves/fighter.png"),
        ("portraits/Kitty/elvish-hero.png",
         "portraits/elves/hero.png"),
        ("portraits/Kitty/elvish-high-lord.png",
         "portraits/elves/high-lord.png"),
        ("portraits/Kitty/elvish-lady.png",
         "portraits/elves/lady.png"),
        ("portraits/Kitty/elvish-lord.png",
         "portraits/elves/lord.png"),
        ("portraits/Kitty/elvish-marksman.png",
         "portraits/elves/marksman.png"),
        ("portraits/Kitty/elvish-marksman+female.png",
         "portraits/elves/marksman+female.png"),
        ("portraits/Kitty/elvish-ranger.png",
         "portraits/elves/ranger.png"),
        ("portraits/Kitty/elvish-ranger+female.png",
         "portraits/elves/ranger+female.png"),
        ("portraits/Kitty/elvish-scout.png",
         "portraits/elves/scout.png"),
        ("portraits/Kitty/elvish-shaman.png",
         "portraits/elves/shaman.png"),
        ("portraits/Kitty/elvish-shyde.png",
         "portraits/elves/shyde.png"),
        ("portraits/Kitty/elvish-sorceress.png",
         "portraits/elves/sorceress.png"),
        ("portraits/Kitty/human-dark-adept.png",
         "portraits/humans/dark-adept.png"),
        ("portraits/Kitty/human-dark-adept+female.png",
         "portraits/humans/dark-adept+female.png"),
        ("portraits/Kitty/human-mage.png",
         "portraits/humans/mage.png"),
        ("portraits/Kitty/human-mage+female.png",
         "portraits/humans/mage+female.png"),
        ("portraits/Kitty/human-mage-arch.png",
         "portraits/humans/mage-arch.png"),
        ("portraits/Kitty/human-mage-arch+female.png",
         "portraits/humans/mage-arch+female.png"),
        ("portraits/Kitty/human-mage-light.png",
         "portraits/humans/mage-light.png"),
        ("portraits/Kitty/human-mage-light+female.png",
         "portraits/humans/mage-light+female.png"),
        ("portraits/Kitty/human-mage-red.png",
         "portraits/humans/mage-red.png"),
        ("portraits/Kitty/human-mage-red+female.png",
         "portraits/humans/mage-red+female.png"),
        ("portraits/Kitty/human-mage-silver.png",
         "portraits/humans/mage-silver.png"),
        ("portraits/Kitty/human-mage-silver+female.png",
         "portraits/humans/mage-silver+female.png"),
        ("portraits/Kitty/human-mage-white.png",
         "portraits/humans/mage-white.png"),
        ("portraits/Kitty/human-mage-white+female.png",
         "portraits/humans/mage-white+female.png"),
        ("portraits/Kitty/human-necromancer.png",
         "portraits/humans/necromancer.png"),
        ("portraits/Kitty/human-necromancer+female.png",
         "portraits/humans/necromancer+female.png"),
        ("portraits/Kitty/troll-whelp.png",
         "portraits/trolls/whelp.png"),
        ("portraits/Kitty/undead-lich.png",
         "portraits/undead/lich.png"),
        ("portraits/Kitty/transparent/elvish-archer.png",
         "portraits/elves/transparent/archer.png"),
        ("portraits/Kitty/transparent/elvish-archer+female.png",
         "portraits/elves/transparent/archer+female.png"),
        ("portraits/Kitty/transparent/elvish-captain.png",
         "portraits/elves/transparent/captain.png"),
        ("portraits/Kitty/transparent/elvish-druid.png",
         "portraits/elves/transparent/druid.png"),
        ("portraits/Kitty/transparent/elvish-fighter.png",
         "portraits/elves/transparent/fighter.png"),
        ("portraits/Kitty/transparent/elvish-hero.png",
         "portraits/elves/transparent/hero.png"),
        ("portraits/Kitty/transparent/elvish-high-lord.png",
         "portraits/elves/transparent/high-lord.png"),
        ("portraits/Kitty/transparent/elvish-lady.png",
         "portraits/elves/transparent/lady.png"),
        ("portraits/Kitty/transparent/elvish-lord.png",
         "portraits/elves/transparent/lord.png"),
        ("portraits/Kitty/transparent/elvish-marksman.png",
         "portraits/elves/transparent/marksman.png"),
        ("portraits/Kitty/transparent/elvish-marksman+female.png",
         "portraits/elves/transparent/marksman+female.png"),
        ("portraits/Kitty/transparent/elvish-ranger.png",
         "portraits/elves/transparent/ranger.png"),
        ("portraits/Kitty/transparent/elvish-ranger+female.png",
         "portraits/elves/transparent/ranger+female.png"),
        ("portraits/Kitty/transparent/elvish-scout.png",
         "portraits/elves/transparent/scout.png"),
        ("portraits/Kitty/transparent/elvish-shaman.png",
         "portraits/elves/transparent/shaman.png"),
        ("portraits/Kitty/transparent/elvish-shyde.png",
         "portraits/elves/transparent/shyde.png"),
        ("portraits/Kitty/transparent/elvish-sorceress.png",
         "portraits/elves/transparent/sorceress.png"),
        ("portraits/Kitty/transparent/human-dark-adept.png",
         "portraits/humans/transparent/dark-adept.png"),
        ("portraits/Kitty/transparent/human-dark-adept+female.png",
         "portraits/humans/transparent/dark-adept+female.png"),
        ("portraits/Kitty/transparent/human-mage.png",
         "portraits/humans/transparent/mage.png"),
        ("portraits/Kitty/transparent/human-mage+female.png",
         "portraits/humans/transparent/mage+female.png"),
        ("portraits/Kitty/transparent/human-mage-arch.png",
         "portraits/humans/transparent/mage-arch.png"),
        ("portraits/Kitty/transparent/human-mage-arch+female.png",
         "portraits/humans/transparent/mage-arch+female.png"),
        ("portraits/Kitty/transparent/human-mage-light.png",
         "portraits/humans/transparent/mage-light.png"),
        ("portraits/Kitty/transparent/human-mage-light+female.png",
         "portraits/humans/transparent/mage-light+female.png"),
        ("portraits/Kitty/transparent/human-mage-red.png",
         "portraits/humans/transparent/mage-red.png"),
        ("portraits/Kitty/transparent/human-mage-red+female.png",
         "portraits/humans/transparent/mage-red+female.png"),
        ("portraits/Kitty/transparent/human-mage-silver.png",
         "portraits/humans/transparent/mage-silver.png"),
        ("portraits/Kitty/transparent/human-mage-silver+female.png",
         "portraits/humans/transparent/mage-silver+female.png"),
        ("portraits/Kitty/transparent/human-mage-white.png",
         "portraits/humans/transparent/mage-white.png"),
        ("portraits/Kitty/transparent/human-mage-white+female.png",
         "portraits/humans/transparent/mage-white+female.png"),
        ("portraits/Kitty/transparent/human-necromancer.png",
         "portraits/humans/transparent/necromancer.png"),
        ("portraits/Kitty/transparent/human-necromancer+female.png",
         "portraits/humans/transparent/necromancer+female.png"),
        ("portraits/Kitty/transparent/troll-whelp.png",
         "portraits/trolls/transparent/whelp.png"),
        ("portraits/Kitty/transparent/undead-lich.png",
         "portraits/undead/transparent/lich.png"),
        ("portraits/Nicholas_Kerpan/human-poacher.png",
         "portraits/humans/poacher.png"),
        ("portraits/Nicholas_Kerpan/human-thief.png",
         "portraits/humans/thief.png"),
        ("portraits/Other/brown-lich.png",
         "portraits/undead/brown-lich.png"),
        ("portraits/Other/cavalryman.png",
         "portraits/humans/cavalryman.png"),
        ("portraits/Other/human-masterbowman.png",
         "portraits/humans/master-bowman.png"),
        ("portraits/Other/scorpion.png",
         "portraits/monsters/scorpion.png"),
        ("portraits/Other/sea-serpent.png",
         "portraits/monsters/sea-serpent.png"),
        ("portraits/Pekka_Aikio/human-bowman.png",
         "portraits/humans/bowman.png"),
        ("portraits/Pekka_Aikio/human-longbowman.png",
         "portraits/humans/longbowman.png"),
        ("portraits/Philip_Barber/dwarf-dragonguard.png",
         "portraits/dwarves/dragonguard.png"),
        ("portraits/Philip_Barber/dwarf-fighter.png",
         "portraits/dwarves/fighter.png"),
        ("portraits/Philip_Barber/dwarf-lord.png",
         "portraits/dwarves/lord.png"),
        ("portraits/Philip_Barber/dwarf-thunderer.png",
         "portraits/dwarves/thunderer.png"),
        ("portraits/Philip_Barber/saurian-augur.png",
         "portraits/saurians/augur.png"),
        ("portraits/Philip_Barber/saurian-skirmisher.png",
         "portraits/saurians/skirmisher.png"),
        ("portraits/Philip_Barber/undead-death-knight.png",
         "portraits/undead/death-knight.png"),
        ("portraits/Philip_Barber/transparent/dwarf-dragonguard.png",
         "portraits/dwarves/transparent/dragonguard.png"),
        ("portraits/Philip_Barber/transparent/dwarf-fighter.png",
         "portraits/dwarves/transparent/fighter.png"),
        ("portraits/Philip_Barber/transparent/dwarf-lord.png",
         "portraits/dwarves/transparent/lord.png"),
        ("portraits/Philip_Barber/transparent/dwarf-thunderer.png",
         "portraits/dwarves/transparent/thunderer.png"),
        ("portraits/Philip_Barber/transparent/saurian-augur.png",
         "portraits/saurians/transparent/augur.png"),
        ("portraits/Philip_Barber/transparent/saurian-skirmisher.png",
         "portraits/saurians/transparent/skirmisher.png"),
        ("portraits/Philip_Barber/transparent/undead-death-knight.png",
         "portraits/undead/transparent/death-knight.png"),
        # Changed just before 1.5.11
        ("titlescreen/landscapebattlefield.jpg",
         "story/landscape-battlefield.jpg"),
        ("titlescreen/landscapebridge.jpg",
         "story/landscape-bridge.jpg"),
        ("titlescreen/landscapecastle.jpg",
         "story/landscape-castle.jpg"),
        ("LABEL_PERSISTANT", "LABEL_PERSISTENT"),
        # Changed just before 1.5.13
        ("targetting", "targeting"),
        # Changed just after 1.7 fork
        ("[stone]", "[petrify]"),
        ("[unstone]", "[unpetrify]"),
        ("[/stone]", "[/petrify]"),
        ("[/unstone]", "[/unpetrify]"),
        ("WEAPON_SPECIAL_STONE", "WEAPON_SPECIAL_PETRIFY"),
        ("SPECIAL_NOTE_STONE", "SPECIAL_NOTE_PETRIFY"),
        (".stoned", ".petrified"),
        ("stoned=", "petrified="),
        # Changed at rev 37390
        ("swing=", "value_second="),
        # Changed just before 1.7.3
        ("Drake Gladiator", "Drake Thrasher"),
        ("gladiator-", "thrasher-"),
        ("Drake Slasher", "Drake Arbiter"),
        ("slasher-", "arbiter-"),
        # Changes after 1.7.5
        ("portraits/nagas/fighter+female.png", "portraits/nagas/fighter.png"),
        # Changes after 1.8rc1
        ("portraits/orcs/warlord.png", "portraits/orcs/transparent/warlord.png"),
        #("portraits/orcs/warlord2.png","portraits/orcs/transparent/warlord.png"), // see 1.11.5
        ("portraits/orcs/warlord3.png","portraits/orcs/transparent/grunt-2.png"),
        #("portraits/orcs/warlord4.png","portraits/orcs/transparent/grunt-2.png"), // see 1.11.5
        ("portraits/orcs/warlord5.png","portraits/orcs/transparent/grunt-3.png"),
        # Changes just before 1.9.0
        ("flat/grass-r8", "flat/grass6"),
        ("flat/grass-r7", "flat/grass5"),
        ("flat/grass-r6", "flat/grass6"),
        ("flat/grass-r5", "flat/grass5"),
        ("flat/grass-r4", "flat/grass4"),
        ("flat/grass-r3", "flat/grass3"),
        ("flat/grass-r2", "flat/grass2"),
        ("flat/grass-r1", "flat/grass1"),
        ("second_value=", "value_second="), # Correct earlier wmllint error
        (".stones", ".petrifies"),
        ("stones=", "petrifies="),
        # Changes just before 1.9.1
        ("[colour_adjust]", "[color_adjust]"),
        ("[/colour_adjust]", "[/color_adjust]"),
        ("colour=", "color="),
        ("colour_lock=", "color_lock="),
        # Changes just before 1.9.2
        ("[removeitem]", "[remove_item]"),
        ("[/removeitem]", "[/remove_item]"),
        # Changes just before 1.11.0
        ("viewing_side", "side"),
        ("duration=level", "duration=scenario"), # Note: this may be removed after 1.11.2, so an actual duration=level can be implemented
        # Changed before 1.11.5 to incorporate 1.9.0 portraits
        ("portraits/orcs/warlord2.png","portraits/orcs/transparent/grunt-5.png"),
        ("portraits/orcs/warlord4.png","portraits/orcs/transparent/grunt-6.png"),

        # Changed before 1.11.8
        ("misc/schedule-dawn.png","misc/time-schedules/default/schedule-dawn.png"),
        ("misc/schedule-morning.png","misc/time-schedules/default/schedule-morning.png"),
        ("misc/schedule-afternoon.png","misc/time-schedules/default/schedule-afternoon.png"),
        ("misc/schedule-dusk.png","misc/time-schedules/default/schedule-dusk.png"),
        ("misc/schedule-firstwatch.png","misc/time-schedules/default/schedule-firstwatch.png"),
        ("misc/schedule-secondwatch.png","misc/time-schedules/default/schedule-secondwatch.png"),

        ("misc/schedule-indoors.png","misc/time-schedules/schedule-indoors.png"),
        ("misc/schedule-underground.png","misc/time-schedules/schedule-underground.png"),
        ("misc/schedule-underground-illum.png","misc/time-schedules/schedule-underground-illum.png"),

        ("misc/tod-schedule-24hrs.png","misc/time-schedules/tod-schedule-24hrs.png"),

        # Changed before 1.13.0 to fix frames for ragged flags
        ('FLAG_VARIANT ragged','FLAG_VARIANT6 ragged'),
        ('FLAG_VARIANT "ragged"','FLAG_VARIANT6 ragged'),

        # Changed in 1.11.15.
        ("fight_on_without_leader=yes","defeat_condition=no_units_left"),
        ("fight_on_without_leader=no","defeat_condition=no_leader_left"),
        ("remove_from_carryover_on_leaders_loss=yes","remove_from_carryover_on_defeat=yes"),
        ("remove_from_carryover_on_leaders_loss=no","remove_from_carryover_on_defeat=no"),

        # Changed in 1.13.2.
        ("[advance]","[advancement]"),
        ("[/advance]","[/advancement]"),
        ("{ABILITY_LEADERSHIP_LEVEL_1}", "{ABILITY_LEADERSHIP}"),
        ("{ABILITY_LEADERSHIP_LEVEL_2}", "{ABILITY_LEADERSHIP}"),
        ("{ABILITY_LEADERSHIP_LEVEL_3}", "{ABILITY_LEADERSHIP}"),
        ("{ABILITY_LEADERSHIP_LEVEL_4}", "{ABILITY_LEADERSHIP}"),
        ("{ABILITY_LEADERSHIP_LEVEL_5}", "{ABILITY_LEADERSHIP}"),
        ("misc/icon-amla-tough.png","icons/amla-default.png"),

        # Changed in 1.13.4: removal of small portraits with black background
        ("portraits/drakes/transparent/blademaster.png", "portraits/drakes/blademaster.png"),
        ("portraits/drakes/transparent/burner.png", "portraits/drakes/burner.png"),
        ("portraits/drakes/transparent/clasher.png", "portraits/drakes/clasher.png"),
        ("portraits/drakes/transparent/enforcer.png", "portraits/drakes/enforcer.png"),
        ("portraits/drakes/transparent/fighter.png", "portraits/drakes/fighter.png"),
        ("portraits/drakes/transparent/flameheart.png", "portraits/drakes/flameheart.png"),
        ("portraits/drakes/transparent/glider.png", "portraits/drakes/glider.png"),
        ("portraits/drakes/transparent/hurricane.png", "portraits/drakes/hurricane.png"),
        ("portraits/drakes/transparent/inferno.png", "portraits/drakes/inferno.png"),
        ("portraits/drakes/transparent/warden.png", "portraits/drakes/warden.png"),
        ("portraits/dwarves/transparent/dragonguard.png", "portraits/dwarves/dragonguard.png"),
        ("portraits/dwarves/transparent/explorer.png", "portraits/dwarves/explorer.png"),
        ("portraits/dwarves/transparent/fighter-2.png", "portraits/dwarves/fighter-2.png"),
        ("portraits/dwarves/transparent/fighter.png", "portraits/dwarves/fighter.png"),
        ("portraits/dwarves/transparent/gryphon-rider.png", "portraits/dwarves/gryphon-rider.png"),
        ("portraits/dwarves/transparent/guard.png", "portraits/dwarves/guard.png"),
        ("portraits/dwarves/transparent/lord.png", "portraits/dwarves/lord.png"),
        ("portraits/dwarves/transparent/runemaster.png", "portraits/dwarves/runemaster.png"),
        ("portraits/dwarves/transparent/scout.png", "portraits/dwarves/scout.png"),
        ("portraits/dwarves/transparent/sentinel.png", "portraits/dwarves/sentinel.png"),
        ("portraits/dwarves/transparent/thunderer.png", "portraits/dwarves/thunderer.png"),
        ("portraits/dwarves/transparent/ulfserker.png", "portraits/dwarves/ulfserker.png"),
        ("portraits/elves/transparent/archer+female.png", "portraits/elves/archer+female.png"),
        ("portraits/elves/transparent/archer.png", "portraits/elves/archer.png"),
        ("portraits/elves/transparent/captain.png", "portraits/elves/captain.png"),
        ("portraits/elves/transparent/druid.png", "portraits/elves/druid.png"),
        ("portraits/elves/transparent/fighter.png", "portraits/elves/fighter.png"),
        ("portraits/elves/transparent/hero.png", "portraits/elves/hero.png"),
        ("portraits/elves/transparent/high-lord.png", "portraits/elves/high-lord.png"),
        ("portraits/elves/transparent/lady.png", "portraits/elves/lady.png"),
        ("portraits/elves/transparent/lord.png", "portraits/elves/lord.png"),
        ("portraits/elves/transparent/marksman+female.png", "portraits/elves/marksman+female.png"),
        ("portraits/elves/transparent/marksman.png", "portraits/elves/marksman.png"),
        ("portraits/elves/transparent/ranger+female.png", "portraits/elves/ranger+female.png"),
        ("portraits/elves/transparent/ranger.png", "portraits/elves/ranger.png"),
        ("portraits/elves/transparent/scout.png", "portraits/elves/scout.png"),
        ("portraits/elves/transparent/shaman.png", "portraits/elves/shaman.png"),
        ("portraits/elves/transparent/shyde.png", "portraits/elves/shyde.png"),
        ("portraits/elves/transparent/sorceress.png", "portraits/elves/sorceress.png"),
        ("portraits/elves/transparent/sylph.png", "portraits/elves/sylph.png"),
        ("portraits/goblins/transparent/direwolver.png", "portraits/goblins/direwolver.png"),
        ("portraits/goblins/transparent/impaler.png", "portraits/goblins/impaler.png"),
        ("portraits/goblins/transparent/pillager.png", "portraits/goblins/pillager.png"),
        ("portraits/goblins/transparent/rouser-2.png", "portraits/goblins/rouser-2.png"),
        ("portraits/goblins/transparent/rouser.png", "portraits/goblins/rouser.png"),
        ("portraits/goblins/transparent/spearman-2.png", "portraits/goblins/spearman-2.png"),
        ("portraits/goblins/transparent/spearman.png", "portraits/goblins/spearman.png"),
        ("portraits/goblins/transparent/wolf-rider.png", "portraits/goblins/wolf-rider.png"),
        ("portraits/humans/transparent/assassin+female.png", "portraits/humans/assassin+female.png"),
        ("portraits/humans/transparent/assassin.png", "portraits/humans/assassin.png"),
        ("portraits/humans/transparent/bandit.png", "portraits/humans/bandit.png"),
        ("portraits/humans/transparent/bowman.png", "portraits/humans/bowman.png"),
        ("portraits/humans/transparent/cavalier.png", "portraits/humans/cavalier.png"),
        ("portraits/humans/transparent/cavalryman.png", "portraits/humans/cavalryman.png"),
        ("portraits/humans/transparent/dark-adept+female.png", "portraits/humans/dark-adept+female.png"),
        ("portraits/humans/transparent/dark-adept.png", "portraits/humans/dark-adept.png"),
        ("portraits/humans/transparent/duelist.png", "portraits/humans/duelist.png"),
        ("portraits/humans/transparent/fencer.png", "portraits/humans/fencer.png"),
        ("portraits/humans/transparent/footpad+female.png", "portraits/humans/footpad+female.png"),
        ("portraits/humans/transparent/footpad.png", "portraits/humans/footpad.png"),
        ("portraits/humans/transparent/general.png", "portraits/humans/general.png"),
        ("portraits/humans/transparent/grand-knight-2.png", "portraits/humans/grand-knight-2.png"),
        ("portraits/humans/transparent/grand-knight.png", "portraits/humans/grand-knight.png"),
        ("portraits/humans/transparent/halberdier.png", "portraits/humans/halberdier.png"),
        ("portraits/humans/transparent/heavy-infantry.png", "portraits/humans/heavy-infantry.png"),
        ("portraits/humans/transparent/horseman.png", "portraits/humans/horseman.png"),
        ("portraits/humans/transparent/huntsman.png", "portraits/humans/huntsman.png"),
        ("portraits/humans/transparent/iron-mauler.png", "portraits/humans/iron-mauler.png"),
        ("portraits/humans/transparent/javelineer.png", "portraits/humans/javelineer.png"),
        ("portraits/humans/transparent/knight.png", "portraits/humans/knight.png"),
        ("portraits/humans/transparent/lancer.png", "portraits/humans/lancer.png"),
        ("portraits/humans/transparent/lieutenant.png", "portraits/humans/lieutenant.png"),
        ("portraits/humans/transparent/longbowman.png", "portraits/humans/longbowman.png"),
        ("portraits/humans/transparent/mage-arch+female.png", "portraits/humans/mage-arch+female.png"),
        ("portraits/humans/transparent/mage-arch.png", "portraits/humans/mage-arch.png"),
        ("portraits/humans/transparent/mage+female.png", "portraits/humans/mage+female.png"),
        ("portraits/humans/transparent/mage-light+female.png", "portraits/humans/mage-light+female.png"),
        ("portraits/humans/transparent/mage-light.png", "portraits/humans/mage-light.png"),
        ("portraits/humans/transparent/mage.png", "portraits/humans/mage.png"),
        ("portraits/humans/transparent/mage-red+female.png", "portraits/humans/mage-red+female.png"),
        ("portraits/humans/transparent/mage-red.png", "portraits/humans/mage-red.png"),
        ("portraits/humans/transparent/mage-silver+female.png", "portraits/humans/mage-silver+female.png"),
        ("portraits/humans/transparent/mage-silver.png", "portraits/humans/mage-silver.png"),
        ("portraits/humans/transparent/mage-white+female.png", "portraits/humans/mage-white+female.png"),
        ("portraits/humans/transparent/mage-white.png", "portraits/humans/mage-white.png"),
        ("portraits/humans/transparent/marshal-2.png", "portraits/humans/marshal-2.png"),
        ("portraits/humans/transparent/marshal.png", "portraits/humans/marshal.png"),
        ("portraits/humans/transparent/master-at-arms.png", "portraits/humans/master-at-arms.png"),
        ("portraits/humans/transparent/master-bowman.png", "portraits/humans/master-bowman.png"),
        ("portraits/humans/transparent/necromancer+female.png", "portraits/humans/necromancer+female.png"),
        ("portraits/humans/transparent/necromancer.png", "portraits/humans/necromancer.png"),
        ("portraits/humans/transparent/outlaw+female.png", "portraits/humans/outlaw+female.png"),
        ("portraits/humans/transparent/outlaw.png", "portraits/humans/outlaw.png"),
        ("portraits/humans/transparent/paladin.png", "portraits/humans/paladin.png"),
        ("portraits/humans/transparent/peasant.png", "portraits/humans/peasant.png"),
        ("portraits/humans/transparent/pikeman.png", "portraits/humans/pikeman.png"),
        ("portraits/humans/transparent/ranger.png", "portraits/humans/ranger.png"),
        ("portraits/humans/transparent/royal-guard.png", "portraits/humans/royal-guard.png"),
        ("portraits/humans/transparent/ruffian.png", "portraits/humans/ruffian.png"),
        ("portraits/humans/transparent/sergeant.png", "portraits/humans/sergeant.png"),
        ("portraits/humans/transparent/spearman-2.png", "portraits/humans/spearman-2.png"),
        ("portraits/humans/transparent/spearman.png", "portraits/humans/spearman.png"),
        ("portraits/humans/transparent/swordsman-2.png", "portraits/humans/swordsman-2.png"),
        ("portraits/humans/transparent/swordsman-3.png", "portraits/humans/swordsman-3.png"),
        ("portraits/humans/transparent/swordsman.png", "portraits/humans/swordsman.png"),
        ("portraits/humans/transparent/thief+female.png", "portraits/humans/thief+female.png"),
        ("portraits/humans/transparent/thief.png", "portraits/humans/thief.png"),
        ("portraits/humans/transparent/thug.png", "portraits/humans/thug.png"),
        ("portraits/humans/transparent/trapper.png", "portraits/humans/trapper.png"),
        ("portraits/humans/transparent/woodsman.png", "portraits/humans/woodsman.png"),
        ("portraits/khalifate/transparent/hakim.png", "portraits/khalifate/hakim.png"),
        ("portraits/merfolk/transparent/enchantress.png", "portraits/merfolk/enchantress.png"),
        ("portraits/merfolk/transparent/fighter.png", "portraits/merfolk/fighter.png"),
        ("portraits/merfolk/transparent/hoplite.png", "portraits/merfolk/hoplite.png"),
        ("portraits/merfolk/transparent/hunter.png", "portraits/merfolk/hunter.png"),
        ("portraits/merfolk/transparent/initiate-2.png", "portraits/merfolk/initiate-2.png"),
        ("portraits/merfolk/transparent/initiate.png", "portraits/merfolk/initiate.png"),
        ("portraits/merfolk/transparent/netcaster.png", "portraits/merfolk/netcaster.png"),
        ("portraits/merfolk/transparent/priestess.png", "portraits/merfolk/priestess.png"),
        ("portraits/merfolk/transparent/spearman.png", "portraits/merfolk/spearman.png"),
        ("portraits/merfolk/transparent/triton.png", "portraits/merfolk/triton.png"),
        ("portraits/monsters/transparent/bat.png", "portraits/monsters/bat.png"),
        ("portraits/monsters/transparent/deep-tentacle.png", "portraits/monsters/deep-tentacle.png"),
        ("portraits/monsters/transparent/giant-mudcrawler.png", "portraits/monsters/giant-mudcrawler.png"),
        ("portraits/monsters/transparent/gryphon.png", "portraits/monsters/gryphon.png"),
        ("portraits/monsters/transparent/ogre.png", "portraits/monsters/ogre.png"),
        ("portraits/monsters/transparent/scorpion.png", "portraits/monsters/scorpion.png"),
        ("portraits/monsters/transparent/sea-serpent.png", "portraits/monsters/sea-serpent.png"),
        ("portraits/monsters/transparent/yeti.png", "portraits/monsters/yeti.png"),
        ("portraits/monsters/transparent/young-ogre.png", "portraits/monsters/young-ogre.png"),
        ("portraits/nagas/transparent/fighter.png", "portraits/nagas/fighter.png"),
        ("portraits/nagas/transparent/myrmidon.png", "portraits/nagas/myrmidon.png"),
        ("portraits/orcs/transparent/archer.png", "portraits/orcs/archer.png"),
        ("portraits/orcs/transparent/assassin.png", "portraits/orcs/assassin.png"),
        ("portraits/orcs/transparent/crossbowman.png", "portraits/orcs/crossbowman.png"),
        ("portraits/orcs/transparent/grunt-2.png", "portraits/orcs/grunt-2.png"),
        ("portraits/orcs/transparent/grunt-3.png", "portraits/orcs/grunt-3.png"),
        ("portraits/orcs/transparent/grunt-4.png", "portraits/orcs/grunt-4.png"),
        ("portraits/orcs/transparent/grunt-5.png", "portraits/orcs/grunt-5.png"),
        ("portraits/orcs/transparent/grunt-6.png", "portraits/orcs/grunt-6.png"),
        ("portraits/orcs/transparent/grunt.png", "portraits/orcs/grunt.png"),
        ("portraits/orcs/transparent/leader-2.png", "portraits/orcs/leader-2.png"),
        ("portraits/orcs/transparent/leader.png", "portraits/orcs/leader.png"),
        ("portraits/orcs/transparent/slayer.png", "portraits/orcs/slayer.png"),
        ("portraits/orcs/transparent/slurbow.png", "portraits/orcs/slurbow.png"),
        ("portraits/orcs/transparent/sovereign.png", "portraits/orcs/sovereign.png"),
        ("portraits/orcs/transparent/warlord.png", "portraits/orcs/warlord.png"),
        ("portraits/orcs/transparent/warrior.png", "portraits/orcs/warrior.png"),
        ("portraits/saurians/transparent/augur.png", "portraits/saurians/augur.png"),
        ("portraits/saurians/transparent/skirmisher.png", "portraits/saurians/skirmisher.png"),
        ("portraits/trolls/transparent/troll-hero-alt.png", "portraits/trolls/troll-hero-alt.png"),
        ("portraits/trolls/transparent/troll-hero.png", "portraits/trolls/troll-hero.png"),
        ("portraits/trolls/transparent/troll.png", "portraits/trolls/troll.png"),
        ("portraits/trolls/transparent/troll-rocklobber.png", "portraits/trolls/troll-rocklobber.png"),
        ("portraits/trolls/transparent/troll-shaman.png", "portraits/trolls/troll-shaman.png"),
        ("portraits/trolls/transparent/troll-warrior.png", "portraits/trolls/troll-warrior.png"),
        ("portraits/trolls/transparent/whelp.png", "portraits/trolls/whelp.png"),
        ("portraits/undead/transparent/ancient-lich.png", "portraits/undead/ancient-lich.png"),
        ("portraits/undead/transparent/archer.png", "portraits/undead/archer.png"),
        ("portraits/undead/transparent/banebow.png", "portraits/undead/banebow.png"),
        ("portraits/undead/transparent/bone-shooter.png", "portraits/undead/bone-shooter.png"),
        ("portraits/undead/transparent/brown-lich.png", "portraits/undead/brown-lich.png"),
        ("portraits/undead/transparent/deathblade.png", "portraits/undead/deathblade.png"),
        ("portraits/undead/transparent/death-knight.png", "portraits/undead/death-knight.png"),
        ("portraits/undead/transparent/draug-2.png", "portraits/undead/draug-2.png"),
        ("portraits/undead/transparent/draug.png", "portraits/undead/draug.png"),
        ("portraits/undead/transparent/ghost.png", "portraits/undead/ghost.png"),
        ("portraits/undead/transparent/ghoul.png", "portraits/undead/ghoul.png"),
        ("portraits/undead/transparent/lich.png", "portraits/undead/lich.png"),
        ("portraits/undead/transparent/nightgaunt.png", "portraits/undead/nightgaunt.png"),
        ("portraits/undead/transparent/revenant.png", "portraits/undead/revenant.png"),
        ("portraits/undead/transparent/shadow.png", "portraits/undead/shadow.png"),
        ("portraits/undead/transparent/skeleton.png", "portraits/undead/skeleton.png"),
        ("portraits/undead/transparent/soulless.png", "portraits/undead/soulless.png"),
        ("portraits/undead/transparent/spectre.png", "portraits/undead/spectre.png"),
        ("portraits/undead/transparent/walking-corpse.png", "portraits/undead/walking-corpse.png"),
        ("portraits/undead/transparent/wraith.png", "portraits/undead/wraith.png"),
        ("portraits/woses/transparent/ancient-wose.png", "portraits/woses/ancient-wose.png"),
        ("portraits/woses/transparent/wose.png", "portraits/woses/wose.png")

        )

def validate_on_pop(tagstack, closer, filename, lineno):
    "Validate the stack at the time a new close tag is seen."
    (tag, attributes) = tagstack[-1]
    ancestors = [x[0] for x in tagstack]
    if verbose >= 3:
        print('"%s", line %d: closing %s I see %s with %s' % (filename, lineno, closer, tag, attributes))
    # Detect a malformation that will cause the game to barf while attempting
    # to deserialize an empty unit.  The final "and attributes" is a blatant
    # hack; some campaigns like to generate entire side declarations with
    # macros.
    if "scenario" in ancestors and closer == "side" and "type" not in attributes and ("no_leader" not in attributes or attributes["no_leader"] != "yes") and "multiplayer" not in ancestors and attributes:
        print('"%s", line %d: [side] without type attribute' % (filename, lineno))
    # This assumes that conversion will always happen in units/ files.
    if "units" not in filename and closer == "unit" and "race" in attributes:
        print('"%s", line %d: [unit] needs hand fixup to [unit_type]' % \
              (filename, lineno))
    if closer in ["campaign", "race"] and "id" not in attributes:
        print('"%s", line %d: %s requires an ID attribute but has none' % \
              (filename, lineno, closer))
    if closer == "terrain" and attributes.get("heals") in ("true", "false"):
        print('"%s", line %d: heals attribute no longer takes a boolean' % \
              (filename, lineno))
    if closer == "unit" and attributes.get("id") is not None and attributes.get("type") is not None and attributes.get("side") is None and not "side" in ancestors:
        print('"%s", line %d: unit declaration without side attribute' % \
              (filename, lineno))
    if closer == "theme" and "id" not in attributes:
        if "name" in attributes:
            print('"%s", line %d: using [theme]name= instead of [theme]id= is deprecated' % (filename, lineno))
        else:
            print('"%s", line %d: [theme] needs an id attribute' % (filename, lineno))
    # Check for user-visible themes that lack a UI name or description.
    if closer == "theme" and ("hidden" not in attributes or attributes["hidden"] not in ("yes", "true")):
        for attr in ("name", "description"):
            if attr not in attributes:
                print('"%s", line %d: [theme] needs a %s attribute unless hidden=yes' % \
                      (filename, lineno, attr))
    if closer == "filter_side":
        ancestor = False
        if "gold" in ancestors:
            ancestor = "gold"
        elif "modify_ai" in ancestors:
            ancestor = "modify_ai"
        if ancestor:
            print('"%s", line %d: %s should have an inline SSF instead of using [filter_side]' % \
                  (filename, lineno, ancestor))
    if closer == "effect":
        if attributes.get("unit_type") is not None:
            print('"%s", line %d: use [effect][filter]type= instead of [effect]unit_type=' % \
                  (filename, lineno))
        if attributes.get("unit_gender") is not None:
            print('"%s", line %d: use [effect][filter]gender= instead of [effect]unit_gender=' % \
                  (filename, lineno))
    if missingside and closer in ["set_recruit", "allow_recruit", "disallow_recruit", "store_gold"] and "side" not in attributes:
        print('"%s", line %d: %s without "side" attribute is now applied to all sides' % \
              (filename, lineno, closer))
    if closer == "variation" and "variation_id" not in attributes:
        print('"%s", line %d: [variation] is missing required variation_id attribute' % \
              (filename, lineno))

def within(tag):
    "Did the specified tag lead one of our enclosing contexts?"
    if type(tag) == type(()): # Can take a list.
        for t in tag:
            if within(t):
                return True
        else:
            return False
    else:
        return tag in [x[0] for x in tagstack]

def under(tag):
    "Did the specified tag lead the latest context?"
    if type(tag) == type(()): # Can take a list.
        for t in tag:
            if within(t):
                return True
        else:
            return False
    elif tagstack:
        return tag == tagstack[-1][0]
    else:
        return False

def standard_unit_filter():
    "Are we within the syntactic context of a standard unit filter?"
    # It's under("message") rather than within("message") because
    # [message] can contain [option] markup with menu item description=
    # attributes that should not be altered.
    return within(("filter", "filter_second",
                   "filter_adjacent", "filter_opponent",
                   "unit_filter", "secondary_unit_filter",
                   "special_filter", "special_filter_second",
                   "neighbor_unit_filter",
                   "recall", "teleport", "kill", "unstone", "store_unit",
                   "have_unit", "scroll_to_unit", "role",
                   "hide_unit", "unhide_unit",
                   "protect_unit", "target", "avoid")) \
                   or under("message")

# Sanity checking

# Associations for the ability sanity checks.
notepairs = [
    ("movement_type=undeadspirit", "{SPECIAL_NOTES_SPIRIT}"),
    ("type=arcane", "{SPECIAL_NOTES_ARCANE}"),
    ("{ABILITY_HEALS}", "{SPECIAL_NOTES_HEALS}"),
    ("{ABILITY_EXTRA_HEAL}", "{SPECIAL_NOTES_EXTRA_HEAL}"),
    ("{ABILITY_UNPOISON}", "{SPECIAL_NOTES_UNPOISON}"),
    ("{ABILITY_CURES}", "{SPECIAL_NOTES_CURES}"),
    ("{ABILITY_REGENERATES}", "{SPECIAL_NOTES_REGENERATES}"),
    ("{ABILITY_STEADFAST}", "{SPECIAL_NOTES_STEADFAST}"),
    ("{ABILITY_LEADERSHIP}", "{SPECIAL_NOTES_LEADERSHIP}"),
    ("{ABILITY_SKIRMISHER}", "{SPECIAL_NOTES_SKIRMISHER}"),
    ("{ABILITY_ILLUMINATES}", "{SPECIAL_NOTES_ILLUMINATES}"),
    ("{ABILITY_TELEPORT}", "{SPECIAL_NOTES_TELEPORT}"),
    ("{ABILITY_AMBUSH}", "{SPECIAL_NOTES_AMBUSH}"),
    ("{ABILITY_NIGHTSTALK}", "{SPECIAL_NOTES_NIGHTSTALK}"),
    ("{ABILITY_CONCEALMENT}", "{SPECIAL_NOTES_CONCEALMENT}"),
    ("{ABILITY_SUBMERGE}", "{SPECIAL_NOTES_SUBMERGE}"),
    ("{ABILITY_FEEDING}", "{SPECIAL_NOTES_FEEDING}"),
    ("{WEAPON_SPECIAL_BERSERK}", "{SPECIAL_NOTES_BERSERK}"),
    ("{WEAPON_SPECIAL_BACKSTAB}", "{SPECIAL_NOTES_BACKSTAB}"),
    ("{WEAPON_SPECIAL_PLAGUE", "{SPECIAL_NOTES_PLAGUE}"), # No } deliberately
    ("{WEAPON_SPECIAL_SLOW}", "{SPECIAL_NOTES_SLOW}"),
    ("{WEAPON_SPECIAL_PETRIFY}", "{SPECIAL_NOTES_PETRIFY}"),
    ("{WEAPON_SPECIAL_MARKSMAN}", "{SPECIAL_NOTES_MARKSMAN}"),
    ("{WEAPON_SPECIAL_MAGICAL}", "{SPECIAL_NOTES_MAGICAL}"),
    ("{WEAPON_SPECIAL_SWARM}", "{SPECIAL_NOTES_SWARM}"),
    ("{WEAPON_SPECIAL_CHARGE}", "{SPECIAL_NOTES_CHARGE}"),
    ("{WEAPON_SPECIAL_DRAIN}", "{SPECIAL_NOTES_DRAIN}"),
    ("{WEAPON_SPECIAL_FIRSTSTRIKE}", "{SPECIAL_NOTES_FIRSTSTRIKE}"),
    ("{WEAPON_SPECIAL_POISON}", "{SPECIAL_NOTES_POISON}"),
    ("{WEAPON_SPECIAL_STUN}", "{SPECIAL_NOTES_STUN}"),
    ]

# This dictionary will pair macros with the characters they recall or create,
# but must be populated by the magic comment, "#wmllint: who ... is ...".
whopairs = {}

# This dictionary pairs macros with the id field of the characters they recall
# or create, and is populated by the comment, "wmllint: whofield <macro> <#>."
whomacros = {}

# This dictionary pairs the ids of stored units with their variable name.
storedids = {}

# This list of the standard recruitable usage types can be appended with the
# magic comment, "#wmllint: usagetype[s]".
usage_types = ["scout", "fighter", "mixed fighter", "archer", "healer"]

# These are accumulated by sanity_check() and examined by consistency_check()
unit_types = []
derived_units = []
usage = {}
sides = []
advances = []
movetypes = []
unit_movetypes = []
races = []
unit_races = []
nextrefs = []
scenario_to_filename = {}

# Attributes that should have translation marks
def is_translatable(key):
    translatables = (
        "abbrev", "cannot_use_message", "caption", "current_player",
        "currently_doing_description", "description", "description_inactive",
        "editor_name", "end_text", "difficulty_descriptions", "female_message",
        "female_name_inactive", "female_names", "help_topic_text", "label",
        "male_message", "male_names", "message", "name", "name_inactive",
        "new_game_title", "note", "option_description", "option_name", "order",
        "plural_name", "prefix", "set_description", "source", "story",
        "summary", "victory_string", "defeat_string", "gold_carryover_string",
        "notes_string", "text", "title", "title2", "tooltip",
        "translator_comment", "user_team_name"
        )
    return key in translatables or key.startswith(("type_", "range_"))

# This is a list of mainline campaigns, used to convert UMC from
# "data/campaigns" to "data/add-ons" while not clobbering mainline.
mainline = ("An_Orcish_Incursion",
            "Dead_Water",
            "Delfadors_Memoirs",
            "Descent_Into_Darkness",
            "Eastern_Invasion",
            "Heir_To_The_Throne",
            "Legend_of_Wesmere",
            "Liberty",
            "Northern_Rebirth",
            "Sceptre_of_Fire",
            "Son_Of_The_Black_Eye",
            "The_Hammer_of_Thursagan",
            "The_Rise_Of_Wesnoth",
            "The_South_Guard",
            "tutorial",
            "Two_Brothers",
            "Under_the_Burning_Suns",
            )

spellcheck_these = (\
    "cannot_use_message=",
    "caption=",
    "description=",
    "description_inactive=",
    "editor_name=",
    "end_text=",
    "help_topic_text=",
    "message=",
    "note=",
    "story=",
    "summary=",
    "text=",
    "title=",
    "title2=",
    "tooltip=",
    "user_team_name=",
    )

# Declare a few common English contractions and ejaculations that pyenchant
# inexplicably knows nothing of.
declared_spellings = {"GLOBAL":["I'm", "I've", "I'd", "I'll",
                                "heh", "ack",
                                # Fantasy/SF/occult jargon that we need
                                "aerie",
                                "aeon",
                                "aide-de-camp",
                                "axe",
                                "ballista",
                                "bided",
                                "crafters",
                                "glaive",
                                "greatsword",
                                "hellspawn",
                                "hurrah",
                                "morningstar",
                                "numbskulls",
                                "overmatched",
                                "spearman",
                                "stygian",
                                "teleport",
                                "teleportation",
                                "teleported",
                                "terraform",
                                "wildlands",
                                # game jargon
                                "melee", "arcane", "day/night", "gameplay",
                                "hitpoint", "hitpoints", "FFA", "multiplayer",
                                "playtesting", "respawn", "respawns",
                                "WML", "HP", "XP", "AI", "ZOC", "YW",
                                "L0", "L1", "L2", "L3", "MC",
                                # archaisms
                                "faugh", "hewn", "leapt", "dreamt", "spilt",
                                "grandmam", "grandsire", "grandsires",
                                "scry", "scrying", "scryed", "woodscraft",
                                "princeling", "wilderlands", "ensorcels",
                                "unlooked", "naphtha", "naïve",
                                # Sceptre of Fire gets spelled with -re.
                                "sceptre",
                                ]}

pango_conversions = (("~", "<b>", "</b>"),
                     ("@", "<span color='green'>", "</span>"),
                     ("#", "<span color='red'>", "</span>"),
                     ("*", "<span size='large'>", "</span>"),
                     ("`", "<span size='small'>", "</span>"),
                     )

def pangostrip(message):
    "Strip Pango markup out of a string."
    # This is all known Pango convenience tags
    for tag in ("b", "big", "i", "s", "sub", "sup", "small", "tt", "u"):
        message = message.replace("<%s>" % tag, "").replace("</%s>" % tag, "")
    # Now remove general span tags
    message = re.sub("</?span[^>]*>", "", message)
    # And Pango specials;
    message = re.sub("&[a-z]+;", "", message)
    return message

def pangoize(message, filename, line):
    "Pango conversion of old-style Wesnoth markup."
    if '&' in message:
        amper = message.find('&')
        if message[amper:amper+1].isspace():
            message = message[:amper] + "&amp;" + message[amper+1:]
    rgb = re.search("(?:<|&lt;)([0-9]+),([0-9]+),([0-9]+)(?:>|&gt;)", message)
    if rgb:
        r, g, b = (min(255, int(c)) for c in rgb.groups())
        hexed = '%02x%02x%02x' % (r, g, b)
        print('"%s", line %d: color spec (%s) requires manual fix (<span color=\'#%s\'>, </span>).' % (filename, line, rgb.group(), hexed))
    # Hack old-style Wesnoth markup
    for (oldstyle, newstart, newend) in pango_conversions:
        if oldstyle not in message:
            continue
        where = message.find(oldstyle)
        if message[where - 1] != '"': # Start of string only
            continue
        if message.strip()[-1] != '"':
            print('"%s", line %d: %s highlight at start of multiline string requires manual fix.' % (filename, line, oldstyle))
            continue
        if '+' in message:
            print('"%s", line %d: %s highlight in composite string requires manual fix.' % (filename, line, oldstyle))
            continue
        # This is the common, simple case we can fix automatically
        message = message[:where] + newstart + message[where + 1:]
        endq = message.rfind('"')
        message = message[:endq] + newend + message[endq:]
    # Check for unescaped < and >
    if "<" in message or ">" in message:
        reduced = pangostrip(message)
        if "<" in reduced or ">" in reduced:
            if message == reduced: # No pango markup
                here = message.find('<')
                if message[here:here+4] != "&lt;":
                    message = message[:here] + "&lt;" + message[here+1:]
                here = message.find('>')
                if message[here:here+4] != "&gt;":
                    message = message[:here] + "&gt;" + message[here+1:]
            else:
                print('"%s", line %d: < or > in pango string requires manual fix.' % (filename, line))
    return message

class WmllintIterator(WmlIterator):
    "Fold an Emacs-compatible error reporter into WmlIterator."
    def printError(self, *misc):
        """Emit an error locator compatible with Emacs compilation mode."""
        if not hasattr(self, 'lineno') or self.lineno == -1:
            print('"%s":' % self.fname, file=sys.stderr)
        else:
            print('"%s", line %d:' % (self.fname, self.lineno+1), end=" ", file=sys.stderr)
        for item in misc:
            print(item, end=" ", file=sys.stderr)
        print("", file=sys.stderr) #terminate line

def local_sanity_check(filename, nav, key, prefix, value, comment):
    "Sanity checks that don't require file context or globals."
    errlead = '"%s", line %d: ' %  (filename, nav.lineno+1)
    ancestors = nav.ancestors()
    in_definition = "#define" in ancestors
    in_call = [x for x in ancestors if x.startswith("{")]
    ignored = "wmllint: ignore" in nav.text
    parent = None
    if ancestors:
        parent = ancestors[-1]
        ancestors = ancestors[:-1]
    # Check for things marked translated that aren't strings
    if "_" in nav.text and not ignored:
        m = re.search(r'[=(]\s*_\s+("?)', nav.text)
        if m and not m.group(1):
            print(errlead + 'translatability mark before non-string')
    # Most tags are not allowed with [part]
    if ("[part]" in ancestors or parent == "[part]") and isOpener(nav.element):
        # FIXME: this should be part of wmliterator's functionality
        if isExtender(nav.element):
            actual_tag = "[" + nav.element[2:]
        else:
            actual_tag = nav.element
        if actual_tag not in ("[part]", "[background_layer]", "[image]", "[insert_tag]",
                              "[if]", "[then]", "[else]", "[switch]", "[case]",
                              "[variable]", "[deprecated_message]"):
            print(errlead + '%s not permitted within [part] tag' % nav.element)
    # Most tags are not permitted inside [if]
    if (len(ancestors) >= 1 and parent == "[if]") or \
       (len(ancestors) >= 2 and parent == "#ifdef" and ancestors[-1] == "[if]"):
        if isOpener(nav.element) and nav.element not in ("[and]",
                           "[else]", "[elseif]", "[frame]", "[have_location]",
                           "[have_unit]", "[not]", "[or]", "[then]",
                           "[variable]") and not nav.element.endswith("_frame]") and not nav.element.startswith("[filter"):
            print(errlead + 'illegal child of [if]:', nav.element)
    # Check for fluky credit parts
    if parent == "[entry]":
        if key == "email" and " " in value:
            print(errlead + 'space in email name')
    # Check for various things that shouldn't be outside an [ai] tag
    if not in_definition and not in_call and "[ai]" not in nav.ancestors() and not ignored:
        if key in ("number_of_possible_recruits_to_force_recruit",
                   "recruitment_ignore_bad_movement",
                   "recruitment_ignore_bad_combat",
                   "recruitment_pattern",
                   "villages_per_scout", "leader_value", "village_value",
                   "aggression", "caution", "attack_depth", "grouping", "advancements"):
            print(errlead + key + " outside [ai] scope")
    # Bad [recruit] attribute
    if parent in ("[allow_recruit]", "[disallow_recruit]") and key == "recruit":
        print(errlead + "recruit= should be type=")
    # Check [binary_path] and [textdomain] paths
    if parent == '[textdomain]' and key == 'path' and '/translations' not in value:
        print(errlead + 'no reference to "/translations" directory in textdomain path')
    if parent == '[binary_path]' and key == 'path':
        if '/external' in value or '/public' in value:
            print(errlead + '"/external" or "/public" image directories should no longer be used')
    # Accumulate data to check for missing next scenarios
    if parent == '[campaign]':
        if key == "first_scenario" and value != "null":
            nextrefs.append((filename, nav.lineno, value))
    if parent == '[scenario]' or parent == None:
        if key == "next_scenario" and value != "null":
            nextrefs.append((filename, nav.lineno, value))
        if key == 'id':
            scenario_to_filename[value] = filename

def global_sanity_check(filename, lines):
    "Perform sanity and consistency checks on input files."
    # Sanity-check abilities and traits against notes macros.
    # Note: This check is disabled on units derived via [base_unit].
    # Also, build dictionaries of unit movement types and races
    in_unit_type = None
    notecheck = True
    trait_note = dict(notepairs)
    note_trait = {p[1]:p[0] for p in notepairs}
    for nav in WmllintIterator(lines, filename):
        if "wmllint: notecheck off" in nav.text:
            notecheck = False
            continue
        elif "wmllint: notecheck on" in nav.text:
            notecheck = True
        #print("Element = %s, text = %s" % (nav.element, repr(nav.text)))
        if nav.element == "[unit_type]":
            unit_race = ""
            unit_id = ""
            base_unit = ""
            traits = []
            notes = []
            has_special_notes = False
            in_unit_type = nav.lineno + 1
            hitpoints_specified = False
            continue
        elif nav.element == "[/unit_type]":
            #print('"%s", %d: unit has traits %s and notes %s' \
            #      % (filename, in_unit_type, traits, notes))
            if unit_id and base_unit:
                derived_units.append((filename, nav.lineno + 1, unit_id, base_unit))
            if unit_id and not base_unit:
                missing_notes = []
                for trait in traits:
                    tn = trait_note[trait]
                    if tn not in notes and tn not in missing_notes:
                        missing_notes.append(tn)
                missing_traits = []
                for note in notes:
                    nt = note_trait[note]
                    if nt not in traits and nt not in missing_traits:
                        missing_traits.append(nt)
                if (notes or traits) and not has_special_notes:
                    missing_notes = ["{SPECIAL_NOTES}"] + missing_notes
                # If the unit didn't specify hitpoints, there is some wacky
                # stuff going on (possibly pseudo-[base_unit] behavior via
                # macro generation) so disable some of the consistency checks.
                if not hitpoints_specified:
                    continue
                if notecheck and missing_notes:
                    print('"%s", line %d: unit %s is missing notes +%s' \
                          % (filename, in_unit_type, unit_id, "+".join(missing_notes)))
                if missing_traits:
                    print('"%s", line %d: unit %s is missing traits %s' \
                          % (filename, in_unit_type, unit_id, "+".join(missing_traits)))
                if notecheck and not (notes or traits) and has_special_notes:
                    print('"%s", line %d: unit %s has superfluous {SPECIAL_NOTES}' \
                         % (filename, in_unit_type, unit_id))
                if "[theme]" not in nav.ancestors() and "[base_unit]" not in nav.ancestors() and not unit_race:
                    print('"%s", line %d: unit %s has no race' \
                         % (filename, in_unit_type, unit_id))
            in_unit_type = None
            traits = []
            notes = []
            unit_id = ""
            base_unit = ""
            has_special_notes = False
            unit_race = None
        if '[unit_type]' in nav.ancestors() and "[filter_attack]" not in nav.ancestors():
            try:
                (key, prefix, value, comment) = parse_attribute(nav.text)
                if key == "id":
                    if value[0] == "_":
                        value = value[1:].strip()
                    if not unit_id and "[base_unit]" not in nav.ancestors():
                        unit_id = value
                        unit_types.append(unit_id)
                    if not base_unit and "[base_unit]" in nav.ancestors():
                        base_unit = value
                elif key == "hitpoints":
                    hitpoints_specified = True
                elif key == "usage":
                    assert(unit_id)
                    usage[unit_id] = value
                elif key == "movement_type":
                    if '{' not in value:
                        assert(unit_id)
                        unit_movetypes.append((unit_id, filename, nav.lineno + 1, value))
                elif key == "race":
                    if '{' not in value:
                        assert(unit_id or base_unit)
                        unit_race = value
                        unit_races.append((unit_id, filename, nav.lineno + 1, unit_race))
                elif key == "advances_to":
                    assert(unit_id or base_unit)
                    advancements = value
                    if advancements.strip() != "null":
                        advances.append((unit_id, filename, nav.lineno + 1, advancements))
            except TypeError:
                pass
            precomment = nav.text
            if '#' in nav.text:
                precomment = nav.text[:nav.text.find("#")]
            if "{SPECIAL_NOTES}" in precomment:
                has_special_notes = True
            for (p, q) in notepairs:
                if p in precomment:
                    traits.append(p)
                if q in precomment:
                    notes.append(q)
    # Detect units that speak in their death events
    filter_subject = None
    die_event = False
    deathcheck = True
    for nav in WmllintIterator(lines, filename):
        if "wmllint: deathcheck off" in nav.text:
            deathcheck = False
            continue
        elif "wmllint: deathcheck on" in nav.text:
            deathcheck = True
        if "[/event]" in nav.text:
            filter_subject = None
            die_event = False
        elif not nav.ancestors():
            continue
        elif "[event]" in nav.ancestors():
            parent = nav.ancestors()[-1]
            if parent == "[event]":
                # Check if it's a death event
                fields = parse_attribute(nav.text)
                if fields:
                    (key, prefix, value, comment) = fields
                    if key == 'name' and value == 'die':
                        die_event = True
            elif die_event and not filter_subject and parent == "[filter]":
                # Check to see if it has a filter subject
                if "id" in nav.text:
                    try:
                        (key,prefix,value,comment) = parse_attribute(nav.text)
                        filter_subject = value
                    except TypeError:
                        pass
            elif die_event and filter_subject and parent == "[message]":
                # Who is speaking?
                fields = parse_attribute(nav.text)
                if fields:
                    (key, prefix, value, comment) = fields
                    if key in ("id", "speaker"):
                        if deathcheck and ((value == filter_subject) or (value == "unit")):
                            print('"%s", line %d: %s speaks in his/her "die" event rather than "last breath"' \
                                  % (filename, nav.lineno+1, value))
    # Collect information on defined movement types and races
    for nav in WmllintIterator(lines, filename):
        above = nav.ancestors()
        if above and above[-1] in ("[movetype]", "[race]"):
            try:
                (key, prefix, value, comment) = parse_attribute(nav.text)
                if above[-1] == "[movetype]" and key == 'name':
                    movetypes.append(value)
                if above[-1] == "[race]" and key == 'id':
                    races.append(value)
            except TypeError:
                pass
    # Sanity-check recruit and recruitment_pattern.
    # This code has a limitation; if there are multiple instances of
    # recruit and recruitment_pattern (as can happen if these lists
    # vary by EASY/NORMAL/HARD level) this code will only record the
    # last of each for later consistency checking.
    in_side = False
    in_ai = in_subunit = False
    recruit = {}
    in_generator = False
    sidecount = 0
    recruitment_pattern = {}
    ifdef_stack = [None]
    for num, line in enumerate((l.strip() for l in lines), start=1):
        if line.startswith("#ifdef") or line.startswith("#ifhave") or line.startswith("#ifver"):
            ifdef_stack.append(line.split()[1])
            continue
        if line.startswith("#ifndef") or line.startswith("#ifnhave") or line.startswith("#ifnver"):
            ifdef_stack.append("!" + line.split()[1])
            continue
        if line.startswith("#else"):
            if ifdef_stack[-1].startswith("!"):
                ifdef_stack.append(ifdef_stack[-1][1:])
            else:
                ifdef_stack.append("!" + ifdef_stack[-1])
            continue
        if line.startswith("#endif"):
            ifdef_stack.pop()
            continue
        if "[generator]" in line:
            in_generator = True
            continue
        elif "[/generator]" in line:
            in_generator = False
            continue
        elif "[side]" in line:
            in_side = True
            sidecount += 1
            continue
        elif "[/side]" in line:
            if recruit or recruitment_pattern:
                sides.append((filename, recruit, recruitment_pattern))
            in_side = False
            recruit = {}
            recruitment_pattern = {}
            continue
        elif in_side and "[ai]" in line:
            in_ai = True
            continue
        elif in_side and "[unit]" in line:
            in_subunit = True
            continue
        elif in_side and "[/ai]" in line:
            in_ai = False
            continue
        elif in_side and "[/unit]" in line:
            in_subunit = False
            continue
        if "wmllint: skip-side" in line:
            sidecount += 1
        if not in_side or in_subunit or '=' not in line:
            continue
        try:
            (key, prefix, value, comment) = parse_attribute(line)
            if key in ("recruit", "extra_recruit") and value:
                recruit[ifdef_stack[-1]] = (num, [x.strip() for x in value.split(",")])
            elif key == "recruitment_pattern" and value:
                if not in_ai:
                    print('"%s", line %d: recruitment_pattern outside [ai]' \
                              % (filename, num))
                else:
                    recruitment_pattern[ifdef_stack[-1]] = (num, [x.strip() for x in value.split(",")])
            elif key == "side" and not in_ai:
                try:
                    if not in_generator and sidecount != int(value):
                        print('"%s", line %d: side number %s is out of sequence (%d expected)' \
                              % (filename, num, value, sidecount))
                except ValueError:
                    pass # Ignore ill-formed integer literals
        except TypeError:
            pass
    # Sanity check ellipses
    # Starting from 1.11.5, units with canrecruit=yes gain automatically a leader ellipse
    # Starting from 1.11.6, units without a ZoC gain automatically a nozoc ellipse
    # Check if ellipse= was used and warn if so
    # Do not warn if misc/ellipse-hero was used, since it isn't automatically assigned by C++
    # and it's assigned/removed with IS_HERO/MAKE_HERO/UNMAKE_HERO
    # magic comment wmllint: no ellipsecheck deactivates this check for the current line
    in_effect = False
    in_unit = False
    in_side = False
    in_unit_type = False
    for num, line in enumerate(lines, start=1):
        if "[effect]" in line:
            in_effect = True
        elif "[/effect]" in line:
            in_effect = False
        elif "[unit]" in line:
            in_unit = True
        elif "[/unit]" in line:
            in_unit = False
        elif "[side]" in line:
            in_side = True
        elif "[/side]" in line:
            in_side = False
        elif "[unit_type]" in line:
            in_unit_type = True
        elif "[/unit_type]" in line:
            in_unit_type = False
        # ellipsecheck magic comment allows to deactivate the ellipse sanity check
        if "wmllint: no ellipsecheck" not in line:
            if in_effect:
                try:
                    (key, prefix, value, comment) = parse_attribute(line)
                    if key == "ellipse" and value in ("misc/ellipse-nozoc", "misc/ellipse-leader"):
                        print('"%s", line %d: [effect] apply_to=ellipse needs to be removed' % (filename, num))
                    elif key == "ellipse" and value not in ("none", "misc/ellipse", "misc/ellipse-hero"):
                        print('"%s", line %d: custom ellipse %s may need to be updated' % (filename, num, value))
                except TypeError: # this is needed to handle tags, that parse_attribute cannot split
                    pass
            elif in_unit or in_side or in_unit_type:
                try:
                    (key, prefix, value, comment) = parse_attribute(line)
                    if key == "ellipse" and value in ("misc/ellipse-nozoc","misc/ellipse-leader"):
                        print('"%s", line %d: %s=%s needs to be removed' % (filename, num, key, value))
                    elif key == "ellipse" and value not in ("none","misc/ellipse","misc/ellipse-hero"):
                        print('"%s", line %d: custom ellipse %s may need to be updated' % (filename, num, value))
                except TypeError: # this is needed to handle tags, that parse_attribute cannot split
                    pass
    # Interpret various magic comments
    for line in lines:
        # Interpret magic comments for setting the usage pattern of units.
        # This coped with some wacky UtBS units that were defined with
        # variant-spawning macros.  The prototype comment looks like this:
        #wmllint: usage of "Desert Fighter" is fighter
        m = re.search('# *wmllint: usage of "([^"]*)" is +(.*)', line)
        if m:
            usage[m.group(1)] = m.group(2).strip()
            unit_types.append(m.group(1))
        # Magic comment for adding non-standard usage types
        m = re.search('# *wmllint: usagetypes? +(.*)', line)
        if m:
            for newusage in m.group(1).split(","):
                usage_types.append(newusage.strip())
        # Accumulate global spelling exceptions
        words = re.search("wmllint: general spellings? (.*)", line)
        if words:
            for word in words.group(1).split():
                declared_spellings["GLOBAL"].append(word.lower())
        words = re.search("wmllint: directory spellings? (.*)", line)
        if words:
            fdir = os.path.dirname(filename)
            if fdir not in declared_spellings:
                declared_spellings[fdir] = []
            for word in words.group(1).split():
                declared_spellings[fdir].append(word.lower())
    # Consistency-check the id= attributes in [side], [unit], [recall],
    # and [message] scopes, also correctness-check translation marks and look
    # for double spaces at end of sentence.
    present = []
    in_scenario = False
    in_multiplayer = False
    subtag_depth = 0
    in_person = False
    in_trait = False
    ignore_id = False
    in_object = False
    in_stage = False
    in_cfg = False
    in_goal = False
    in_set_menu_item = False
    in_clear_menu_item = False
    in_facet = False
    in_sound_source = False
    in_remove_sound_source = False
    in_message = False
    in_option = False
    #last index is true: we're currently directly in an [event]
    #this avoids complaints about unknown [event]id=something, but keeps the check
    #in case some [filter]id= comes in this [event]
    directly_in_event = []
    in_time_area = False
    in_store = False
    in_unstore = False
    in_not = False
    in_clear = False
    storeid = None
    storevar = None
    ignoreable = False
    preamble_seen = False
    sentence_end = re.compile("(?<=[.!?;:])  +")
    capitalization_error = re.compile("(?<=[.!?])  +[a-z]")
    markcheck = True
    translation_mark = re.compile(r'_ *"')
    for i in range(len(lines)):
        if '[' in lines[i]:
            preamble_seen = True
        # This logic looks odd because a scenario can be conditionally
        # wrapped in both [scenario] and [multiplayer]; we mustn't count
        # either as a subtag even if it occurs inside the other, otherwise
        # this code might see id= declarations as being at the wrong depth.
        if "[scenario]" in lines[i]:
            in_scenario = True
            preamble_seen = False
        elif "[/scenario]" in lines[i]:
            in_scenario = False
        elif "[multiplayer]" in lines[i]:
            in_multiplayer = True
            preamble_seen = False
        elif "[/multiplayer]" in lines[i]:
            in_multiplayer = False
        else:
            if re.search(r"\[[a-z]", lines[i]):
                subtag_depth += 1
            if "[/" in lines[i]:
                subtag_depth -= 1
        if "[event]" in lines[i]:
            directly_in_event.append(True)
        elif re.search(r"\[[a-z]", lines[i]):
            directly_in_event.append(False)
        elif "[/" in lines[i]:
            if len(directly_in_event) > 0:
                directly_in_event.pop()
        # Ordinary subtag flags begin here
        if "[trait]" in lines[i]:
            in_trait = True
        elif "[/trait]" in lines[i]:
            in_trait = False
        elif "[object]" in lines[i]:
            in_object = True
        elif "[/object]" in lines[i]:
            in_object = False
        elif "[stage]" in lines[i]:
            in_stage = True
        elif "[/stage]" in lines[i]:
            in_stage = False
        elif "[cfg]" in lines[i]:
            in_cfg = True
        elif "[/cfg]" in lines[i]:
            in_cfg = False
        elif "[goal]" in lines[i]:
            in_goal = True
        elif "[/goal]" in lines[i]:
            in_goal = False
        elif "[set_menu_item]" in lines[i]:
            in_set_menu_item = True
        elif "[/set_menu_item]" in lines[i]:
            in_set_menu_item = False
        elif "[clear_menu_item]" in lines[i]:
            in_clear_menu_item = True
        elif "[/clear_menu_item]" in lines[i]:
            in_clear_menu_item = False
        elif "[facet]" in lines[i]:
            in_facet = True
        elif "[/facet]" in lines[i]:
            in_facet = False
        elif "[sound_source]" in lines[i]:
            in_sound_source = True
        elif "[/sound_source]" in lines[i]:
            in_sound_source = False
        elif "[remove_sound_source]" in lines[i]:
            in_remove_sound_source = True
        elif "[/remove_sound_source]" in lines[i]:
            in_remove_sound_source = False
        elif "[message]" in lines[i]:
            in_message = True
        elif "[/message]" in lines[i]:
            in_message = False
        elif "[/option]" in lines[i]:
            in_option = False
        elif "[option]" in lines[i]:
            in_option = True
        elif "[time_area]" in lines[i]:
            in_time_area = True
        elif "[/time_area]" in lines[i]:
            in_time_area = False
        elif "[label]" in lines[i] or "[chamber]" in lines[i] or "[time]" in lines[i]:
            ignore_id = True
        elif "[/label]" in lines[i] or "[/chamber]" in lines[i] or "[/time]" in lines[i]:
            ignore_id = False
        elif "[kill]" in lines[i] or "[effect]" in lines[i] or "[move_unit_fake]" in lines[i] or "[scroll_to_unit]" in lines[i]:
            ignoreable = True
        elif "[/kill]" in lines[i] or "[/effect]" in lines[i] or "[/move_unit_fake]" in lines[i] or "[/scroll_to_unit]" in lines[i]:
            ignoreable = False
        elif "[side]" in lines[i] or "[unit]" in lines[i] or "[recall]" in lines[i]:
            in_person = True
            continue
        elif "[/side]" in lines[i] or "[/unit]" in lines[i] or "[/recall]" in lines[i]:
            in_person = False
        elif "[store_unit]" in lines[i]:
            in_store = True
        elif "[/store_unit]" in lines[i]:
            if storeid and storevar:
                storedids.update({storevar: storeid})
            in_store = False
            storeid = storevar = None
        elif "[unstore_unit]" in lines[i]:
            in_unstore = True
        elif "[/unstore_unit]" in lines[i]:
            in_unstore = False
        elif "[not]" in lines[i]:
            in_not = True
        elif "[/not]" in lines[i]:
            in_not = False
        elif "[clear_variable]" in lines[i]:
            in_clear = True
        elif "[/clear_variable]" in lines[i]:
            in_clear = False
        if "wmllint: markcheck off" in lines[i]:
            markcheck = False
        elif "wmllint: markcheck on" in lines[i]:
            markcheck = True
        elif 'wmllint: who ' in lines[i]:
            try:
                fields = lines[i].split("wmllint: who ", 1)[1].split(" is ", 1)
                if len(fields) == 2:
                    mac = string_strip(fields[0].strip()).strip('{}')
                    if mac in whopairs:
                        whopairs[mac] = whopairs[mac] + ", " + fields[1].strip()
                    else:
                        whopairs.update({mac: fields[1].strip()})
            except IndexError:
                pass
        elif 'wmllint: unwho ' in lines[i]:
            unmac = lines[i].split("wmllint: unwho ", 1)[1].strip()
            if string_strip(unmac).upper() == 'ALL':
                whopairs.clear()
            else:
                try:
                    del whopairs[string_strip(unmac).strip('{}')]
                except KeyError:
                    print('%s, line %s: magic comment "unwho %s" does not match any current keys: %s' \
                          % (filename, i+1, unmac, ", ".join(whopairs.keys())), file=sys.stderr)
        elif 'wmllint: whofield' in lines[i]:
            fields = re.search(r'wmllint: whofield\s+([^\s]+)(\s+is)?\s*([^\s]*)', lines[i])
            if fields:
                if fields.group(1).startswith('clear'):
                    if fields.group(3) in whomacros:
                        del whomacros[fields.group(3)]
                    else:
                        whomacros.clear()
                elif re.match(r'[1-9][0-9]*$', fields.group(3)):
                    whomacros.update({fields.group(1): int(fields.group(3))})
                else:
                    try:
                        del whomacros[fields.group(1)]
                    except KeyError:
                        print('%s, line %s: magic comment "whofield %s" should be followed by a number: %s' \
                              % (filename, i+1, unmac, fields.group(3)), file=sys.stderr)
        # Parse recruit/recall macros to recognize characters.  This section
        # assumes that such a macro is the first item on a line.
        leadmac = re.match(r'{[^}\s]+.', lines[i].lstrip())
        if leadmac:
            macname = leadmac.group()[1:-1]
            # Recognize macro pairings from "wmllint: who" magic
            # comments.
            if macname in whopairs:
                for who in whopairs[macname].split(","):
                    if who.strip().startswith("--"):
                        try:
                            present.remove(who.replace('--', '', 1).strip())
                        except:
                            ValueError
                    else:
                        present.append(who.strip())
            elif not leadmac.group().endswith('}'):
                # Update 1.4's {LOYAL_UNIT} macro to {NAMED_LOYAL_UNIT}.  Do
                # this here rather than hack_syntax so the character can be
                # recognized.
                if macname == 'LOYAL_UNIT':
                    (args, brack, paren) = parse_macroref(0, leadmac.string)
                    if len(args) == 7:
                        lines[i] = lines[i].replace('{LOYAL_UNIT', '{NAMED_LOYAL_UNIT', 1)
                # Auto-recognize the people in the {NAMED_*UNIT} macros.
                if re.match(r'NAMED_[A-Z_]*UNIT$', macname):
                    (args, brack, paren) = parse_macroref(0, leadmac.string)
                    if len(args) >= 7 and \
                       re.match(r'([0-9]+|[^\s]*\$[^\s]*side[^\s]*|{[^\s]*SIDE[^\s]*})$', args[1]) and \
                       re.match(r'([0-9]+|[^\s]*\$[^\s]*x[^\s]*|{[^\s]*X[^\s]*})$', args[3]) and \
                       re.match(r'([0-9]+|[^\s]*\$[^\s]*y[^\s]*|{[^\s]*Y[^\s]*})$', args[4]) and \
                       len(args[5]) > 0:
                        present.append(args[5])
                elif macname == 'RECALL':
                    (args, brack, paren) = parse_macroref(0, leadmac.string)
                    if len(args) == 2 and brack == 0:
                        present.append(args[1])
                elif macname == 'RECALL_XY':
                    (args, brack, paren) = parse_macroref(0, leadmac.string)
                    if len(args) == 4:
                        present.append(args[1])
                elif macname == 'CLEAR_VARIABLE':
                    (args, brack, paren) = parse_macroref(0, leadmac.string)
                    for arg in [x.lstrip() for x in args[1].split(',')]:
                        if arg in storedids:
                            del storedids[arg]
                elif macname in whomacros:
                    (args, brack, paren) = parse_macroref(0, leadmac.string)
                    present.append(args[whomacros[macname]])
        m = re.search("# *wmllint: recognize +(.*)", lines[i])
        if m:
            present.append(string_strip(m.group(1)).strip())
        if '=' not in lines[i] or ignoreable:
            continue
        parseable = False
        try:
            (key, prefix, value, comment) = parse_attribute(lines[i])
            parseable = True
        except TypeError:
            pass
        if parseable:
            if "wmllint: ignore" in comment:
                continue
            # Recognize units when unstored
            if (in_scenario or in_multiplayer) and in_store:
                if key == 'id' and not in_not:
                    if not storeid == None:
                        storeid == storeid + ',' + value
                    else:
                        storeid = value
                elif key == 'variable' and '{' not in value:
                    storevar = value
            elif in_unstore:
                if key == 'variable':
                    value = value.split("[$")[0]
                    if value in storedids:
                        for unit_id in storedids[value].split(','):
                            present.append(unit_id.lstrip())
                        del storedids[value]
            elif key == 'name' and in_clear:
                for val in value.split(','):
                    val = val.lstrip()
                    if val in storedids:
                        del storedids[val]
            has_tr_mark = translation_mark.search(value)
            if key == 'role':
                present.append(value)
            if has_tr_mark:
                # FIXME: This test is rather bogus as is.
                # Doing a better job would require tokenizing to pick up the
                # string boundaries. I'd do it, but AI0867 is already working
                # on a parser-based wmllint.
                if '{' in value and "+" not in value and value.find('{') > value.find("_"):
                    print('"%s", line %d: macro reference in translatable string'\
                          % (filename, i+1))
                #if future and re.search("[.,!?]  ", lines[i]):
                #    print('"%s", line %d: extraneous space in translatable string'\
                #          % (filename, i+1))
            # Check correctness of translation marks and descriptions
            if key.startswith("#"): # FIXME: parse_attribute is confused.
                pass
            elif key.startswith("{"):
                pass
            elif key == 'letter': # May be led with _s for void
                pass
            elif key in ('name', 'male_name', 'female_name', 'value'): # FIXME: check this someday
                pass
            elif key == "variation_name":
                if markcheck and not has_tr_mark:
                    print('"%s", line %d: %s should be renamed as variation_id and/or marked as translatable' \
                          % (filename, i+1, key))
            elif is_translatable(key):
                if markcheck and has_tr_mark and '""' in line:
                    print('"%s", line %d: %s doesn`t need translation mark (translatable string is empty)' \
                          % (filename, i+1, key))
                    lines[i] = lines[i].replace("=_","=")
                if markcheck and not value.startswith("$") and not value.startswith("{") and not re.match(" +", value) and not has_tr_mark and '""' not in line and not ("wmllint: ignore" in comment or "wmllint: noconvert" in comment):
                    print('"%s", line %d: %s needs translation mark' \
                          % (filename, i+1, key))
                    lines[i] = lines[i].replace('=', "=_ ")
                nv = sentence_end.sub(" ", value)
                if nv != value:
                    print('"%s", line %d: double space after sentence end' \
                          % (filename, i+1))
                    if not stringfreeze:
                        lines[i] = sentence_end.sub(" ", lines[i])
                if capitalization_error.search(lines[i]):
                    print('"%s", line %d: probable capitalization or punctuation error' \
                          % (filename, i+1))
                if key == "message" and in_message and not in_option and not ("wmllint: ignore" in comment or "wmllint: noconvert" in comment):
                    lines[i] = pangoize(lines[i], filename, i)
            else:
                if (in_scenario or in_multiplayer) and key == "id":
                    if in_person:
                        present.append(value)
                    elif value and value[0] in ("$", "{"):
                        continue
                    elif preamble_seen and subtag_depth > 0 and not ignore_id and not in_object and not in_cfg and not in_facet and not in_sound_source and not in_remove_sound_source and not in_stage and not in_goal and not in_set_menu_item and not in_clear_menu_item and not directly_in_event[-1] and not in_time_area and not in_trait:
                        ids = value.split(",")
                        for id_ in ids:
                            # removal of leading whitespace of items in comma-separated lists
                            # is usually supported in the mainline wesnoth lua scripts
                            # not sure about trailing one
                            # also, do not complain about ids if they're referred to a menu item being cleared
                            if id_.lstrip() not in present:
                                print('"%s", line %d: unknown \'%s\' referred to by id' \
                                      % (filename, i+1, id_))
                if (in_scenario or in_multiplayer) and key == "speaker":
                    if value not in present and value not in ("narrator", "unit", "second_unit") and value[0] not in ("$", "{"):
                        print('"%s", line %d: unknown speaker \'%s\' of [message]' \
                              % (filename, i+1, value))
                if markcheck and has_tr_mark and not ("wmllint: ignore" in comment or "wmllint: noconvert" in comment):
                    print('"%s", line %d: %s should not have a translation mark' \
                          % (filename, i+1, key))
                    lines[i] = prefix + value.replace("_", "", 1) + comment + '\n'
    # Now that we know who's present, register all these names as spellings
    declared_spellings[filename] = [x.lower() for x in present if len(x) > 0]
    # Check for textdomain strings; should be exactly one, on line 1
    textdomains = []
    no_text = False
    for num, line in enumerate(lines, start=1):
        if "#textdomain" in line:
            textdomains.append(num)
        elif "wmllint: no translatables":
            no_text = True
    if not no_text:
        if not textdomains:
            print('"%s", line 1: no textdomain string' % filename)
        elif textdomains[0] == 1: # Multiples are OK if first is on line 1
            pass
        elif len(textdomains) > 1:
            print('"%s", line %d: multiple textdomain strings on lines %s' % \
                  (filename, textdomains[0], ", ".join(map(str, textdomains))))
        else:
            w = textdomains[0]
            print('"%s", line %d: single textdomain declaration not on line 1.' % \
                  (filename, w))
            lines = [lines[w-1].lstrip()] + lines[:w-1] + lines[w:]
    return lines

def condition_match(p, q):
    "Do two condition-states match?"
    # The empty condition state is represented by None
    if p is None or q is None or (p == q):
        return True
    # Past this point it's all about handling cases with negation
    sp = p
    np = False
    if sp.startswith("!"):
        sp = sp[1:]
        np = True
    sq = q
    nq = False
    if sq.startswith("!"):
        sq = sp[1:]
        nq == True
    return (sp != sq) and (np != nq)

def consistency_check():
    "Consistency-check state information picked up by sanity_check"
    derivations = {u[2]: u[3] for u in derived_units}
    for (filename, recruitdict, patterndict) in sides:
        for (rdifficulty, (rl, recruit)) in recruitdict.items():
            utypes = []
            for rtype in recruit:
                base = rtype
                if rtype not in unit_types:
                    # Assume WML coder knew what he was doing if macro reference
                    if not rtype.startswith("{"):
                        print('"%s", line %d: %s is not a known unit type' % (filename, rl, rtype))
                    continue
                elif rtype not in usage:
                    if rtype in derivations:
                        base = derivations[rtype]
                    else:
                        print('"%s", line %d: %s has no usage type' % \
                              (filename, rl, rtype))
                        continue
                if not base in usage:
                    print('"%s", line %d: %s has unknown base %s' % \
                          (filename, rl, rtype, base))
                    continue
                else:
                    utype = usage[base]
                utypes.append(utype)
                for (pdifficulty, (pl, recruit_pattern)) in patterndict.items():
                    if condition_match(pdifficulty, rdifficulty):
                        if utype not in recruit_pattern:
                            rshow = ''
                            if rdifficulty is not None:
                                rshow = 'At ' + rdifficulty + ', '
                            ushow = ''
                            if utype not in usage_types:
                                ushow = ', a non-standard usage class'
                            pshow = ''
                            if pdifficulty is not None:
                                pshow = ' ' + pdifficulty
                            print('"%s", line %d: %s%s (%s%s) doesn\'t match the%s recruitment pattern (%s) for its side' \
                                  % (filename, rl, rshow, rtype, utype, ushow, pshow, ", ".join(recruit_pattern)))
            # We have a list of all the usage types recruited at this difficulty
            # in utypes.  Use it to check the matching pattern, if any. Suppress
            # this check if the recruit line is a macroexpansion.
            if recruit and not recruit[0].startswith("{"):
                for (pdifficulty, (pl, recruitment_pattern)) in patterndict.items():
                    if condition_match(pdifficulty, rdifficulty):
                        for utype in recruitment_pattern:
                            if utype not in utypes:
                                rshow = '.'
                                if rdifficulty is not None:
                                    rshow = ' at difficulty ' + rdifficulty + '.'
                                ushow = ''
                                if utype not in usage_types:
                                    ushow = ' (a non-standard usage class)'
                                print('"%s", line %d: no %s%s units recruitable%s' % (filename, pl, utype, ushow, rshow))
    if movetypes:
        for (unit_id, filename, line, movetype) in unit_movetypes:
            if movetype not in movetypes:
                print('"%s", line %d: %s has unknown movement type' \
                      % (filename, line, unit_id))
    if races:
        for (unit_id, filename, line, race) in unit_races:
            if race not in races:
                print('"%s", line %d: %s has unknown race' \
                      % (filename, line, unit_id))
    # Should we be checking the transitive closure of derivation?
    # It's not clear whether [base_unit] works when the base is itself derived.
    for (filename, line, unit_type, base_unit) in derived_units:
        if base_unit not in unit_types:
            print('"%s", line %d: derivation of %s from %s does not resolve' \
                  % (filename, line, unit_type, base_unit))
    # Check that all advancements are known units
    for (unit_id, filename, lineno, advancements) in advances:
        advancements = [elem.strip() for elem in advancements.split(",")]
        known_units = unit_types + list(derivations.keys())
        bad_advancements = [x for x in advancements if x not in known_units]
        if bad_advancements:
            print('"%s", line %d: %s has unknown advancements %s' \
                  % (filename, lineno, unit_id, bad_advancements))
    # Check next-scenario pointers
    #print("Scenario ID map", scenario_to_filename)
    for (filename, lineno, value) in nextrefs:
        if value not in scenario_to_filename:
            print('"%s", line %d: unresolved scenario reference %s' % \
                  (filename, lineno, value))
    # Report stored units never unstored or cleared
    for store in storedids.keys():
        print('wmllint: stored unit "%s" not unstored or cleared from "%s"' % (storedids[store], store))

# Syntax transformations

leading_ws = re.compile(r"^\s*")

def leader(s):
    "Return a copy of the leading whitespace in the argument."
    return leading_ws.match(s).group(0)

def hack_syntax(filename, lines):
    # Syntax transformations go here.  This gets called once per WML file;
    # the name of the file is passed as filename, text of the file as the
    # array of strings in lines.  Modify lines in place as needed;
    # changes will be detected by the caller.
    #
    # Deal with a few Windows-specific problems for the sake of cross-
    # platform harmony. First, the use of backslashes in file paths.
    for i in range(len(lines)):
        if "no-syntax-rewrite" in lines[i]:
            break
        if lines[i].lstrip().startswith("#"):
            pass
        # Looking out for "#" used for color markup
        precomment = re.split(r'\s#', lines[i], 1)[0]
        comment = lines[i][len(precomment):]
        if '\\' in precomment:
            while re.search(r'(?<!\\)\\(?!\\)[^ ={}"]+\.(png|ogg|wav|gif|jpe?g|map|mask|cfg)\b', precomment, flags=re.IGNORECASE):
                backslash = re.search(r'([^ ={}"]*(?<!\\)\\(?!\\)[^ ={}"]+\.)(png|ogg|wav|gif|jpe?g|map|mask|cfg)(?=\b)', precomment, flags=re.IGNORECASE)
                fronted = backslash.group(1).replace("\\","/") + backslash.group(2)
                precomment = precomment[:backslash.start()] + fronted + precomment[backslash.end():]
                print('"%s", line %d: %s -> %s -- please use frontslash (/) for cross-platform compatibility' \
                      % (filename, i+1, backslash.group(), fronted))
        # Then get rid of the 'userdata/' headache.
        if 'userdata/' in precomment:
            while re.search(r'user(data/)?data/[ac]', precomment):
                userdata = re.search(r'(?:\.\./)?user(?:data/)?(data/[ac][^/]*/?)', precomment)
                precomment = precomment[:userdata.start()] + userdata.group(1) + precomment[userdata.end():]
                print('"%s", line %d: %s -> %s -- DO NOT PREFIX PATHS WITH "userdata/"' \
                      % (filename, i+1, userdata.group(), userdata.group(1)))
        lines[i] = precomment + comment
    # Ensure that every attack has a translatable description.
    for i in range(len(lines)):
        if "no-syntax-rewrite" in lines[i]:

            break
        elif "[attack]" in lines[i]:
            j = i;
            have_description = False
            while '[/attack]' not in lines[j]:
                if lines[j].strip().startswith("description"):
                    have_description = True
                j += 1
            if not have_description:
                j = i
                while '[/attack]' not in lines[j]:
                    fields = lines[j].strip().split('#')
                    syntactic = fields[0]
                    comment = ""
                    if len(fields) > 1:
                        comment = fields[1]
                    if syntactic.strip().startswith("name"):
                        description = syntactic.split("=")[1].strip()
                        if not description.startswith('"'):
                            description = '"' + description + '"\n'
                        # Skip the insertion if this is a dummy declaration
                        # or one modifying an attack inherited from a base unit.
                        if "no-icon" not in comment:
                            new_line = leader(syntactic) + "description=_"+description
                            if verbose:
                                print('"%s", line %d: inserting %s' % (filename, i+1, repr(new_line)))
                            lines.insert(j+1, new_line)
                            j += 1
                    j += 1
    # Ensure that every speaker=narrator block without an image uses
    # wesnoth-icon.png as an image.
    need_image = in_message = False
    for i in range(len(lines)):
        if "no-syntax-rewrite" in lines[i]:
            break
        precomment = lines[i].split("#")[0]
        if '[message]' in precomment:
            in_message = True
        if "speaker=narrator" in precomment:
            need_image = True
        elif precomment.strip().startswith("image"):
            need_image = False
        elif '[/message]' in precomment:
            if need_image:
                # This line presumes the code has been through wmlindent
                if verbose:
                    print('"%s", line %d: inserting "image=wesnoth-icon.png"'%(filename, i+1))
                lines.insert(i, leader(precomment) + baseindent + "image=wesnoth-icon.png\n")
            need_image = in_message = False
    # Hack tracking-map macros from 1.4 and earlier.  The idea is to lose
    # all assumptions about colors in the names
    for i in range(len(lines)):
        if "no-syntax-rewrite" in lines[i]:
            break
        if lines[i].lstrip().startswith("#"):
            pass
        elif "{DOT_CENTERED" in lines[i]:
            lines[i] = lines[i].replace("DOT_CENTERED", "NEW_JOURNEY")
        elif "{DOT_WHITE_CENTERED" in lines[i]:
            lines[i] = lines[i].replace("DOT_WHITE_CENTERED", "OLD_JOURNEY")
        elif "{CROSS_CENTERED" in lines[i]:
            lines[i] = lines[i].replace("CROSS_CENTERED", "NEW_BATTLE")
        elif "{CROSS_WHITE_CENTERED" in lines[i]:
            lines[i] = lines[i].replace("CROSS_WHITE_CENTERED", "OLD_BATTLE")
        elif "{FLAG_RED_CENTERED" in lines[i]:
            lines[i] = lines[i].replace("FLAG_RED_CENTERED", "NEW_REST")
        elif "{FLAG_WHITE_CENTERED" in lines[i]:
            lines[i] = lines[i].replace("FLAG_WHITE_CENTERED", "OLD_REST")
        elif "{DOT " in lines[i] or "CROSS" in lines[i]:
            m = re.search("{(DOT|CROSS) ([0-9]+) ([0-9]+)}", lines[i])
            if m:
                n = m.group(1)
                if n == "DOT":
                    n = "NEW_JOURNEY"
                if n == "CROSS":
                    n = "NEW_BATTLE"
                x = int(m.group(2)) + 5
                y = int(m.group(3)) + 5
                lines[i] = lines[i][:m.start(0)] +("{%s %d %d}" % (n, x, y)) + lines[i][m.end(0):]
    # Fix bare strings containing single quotes; these confuse wesnoth-mode.el
    for i in range(len(lines)):
        if "no-syntax-rewrite" in lines[i]:
            break
        elif lines[i].count("'") % 2 == 1:
            try:
                (key, prefix, value, comment) = parse_attribute(lines[i])
                if "'" in value and value[0].isalpha() and value[-1].isalpha() and '"'+value+'"' not in lines[i]:
                    newtext = prefix + '"' + value + '"' + comment + "\n"
                    if lines[i] != newtext:
                        lines[i] = newtext
                        if verbose:
                            print('"%s", line %d: quote-enclosing attribute value.'%(filename, i+1))
            except TypeError:
                pass
    # Palette transformation for 1.7:
    for i in range(len(lines)):
        if "no-syntax-rewrite" in lines[i]:
            break
        if lines[i].lstrip().startswith("#"):
            pass
        # RC -> PAL
        elif "RC" in lines[i]:
            lines[i] = re.sub(r"~RC\(([^=\)]*)=([^)]*)\)",r"~PAL(\1>\2)",lines[i])
    # Rename the terrain definition tag
    in_standing_anim = False
    for i in range(len(lines)):
        if "no-syntax-rewrite" in lines[i]:
            break
        if lines[i].lstrip().startswith("#"):
            pass
        # Ugh...relies on code having been wmlindented
        lines[i] = re.sub(r"^\[terrain\]", "[terrain_type]", lines[i])
        lines[i] = re.sub(r"^\[/terrain\]", "[/terrain_type]", lines[i])
        if "[standing_anim]" in lines[i]:
            in_standing_anim = True
        if "[/standing_anim]" in lines[i]:
            in_standing_anim = False
        if in_standing_anim:
            lines[i] = re.sub(r"terrain([^_])", r"terrain_type\1", lines[i])
    # Rename two attributes in [set_variable]
    in_set_variable = False
    for i in range(len(lines)):
        if "no-syntax-rewrite" in lines[i]:
            break
        if lines[i].lstrip().startswith("#"):
            pass
        if "[set_variable]" in lines[i]:
            in_set_variable = True
        if "[/set_variable]" in lines[i]:
            in_set_variable = False
        if in_set_variable:
            lines[i] = re.sub(r"format(?=\s*=)", r"value", lines[i])
            lines[i] = re.sub(r"random(?=\s*=)", r"rand", lines[i])
    # campaigns directory becomes add-ons
    in_binary_path = in_textdomain = False
    for i in range(len(lines)):
        if "no-syntax-rewrite" in lines[i]:
            break
        if lines[i].lstrip().startswith("#"):
            pass
        # This is done on every line
        if "campaigns/" in lines[i]:
            lines[i] = lines[i].replace("{~campaigns/", "{~add-ons/")
            lines[i] = lines[i].replace("{~/campaigns/", "{~add-ons/")
            lines[i] = lines[i].replace("{@campaigns/", "{~add-ons/")
            # Convert UMC to data/add-ons without clobbering mainline. Each path
            # is checked against a list of mainline campaigns. UMC paths are
            # updated to "data/add-ons/"; mainline path strings are unaltered.
            x = 0
            for dc in re.finditer(r"data/campaigns/(\w[\w'&+-]*)", lines[i]):
                if dc.group(1) in mainline:
                    continue
                # Because start() and end() are based on the original position
                # of each iteration, while each replacement shortens the line
                # by two characters, we must subtract an increment that grows
                # with each substitution.
                lines[i] = lines[i][:dc.start()-x] + 'data/add-ons/' + dc.group(1) + lines[i][dc.end()-x:]
                x = x+2
                print('"%s", line %d: data/campaigns/%s -> data/add-ons/%s'\
                      %(filename, i+1, dc.group(1), dc.group(1)))
        elif "@add-ons/" in lines[i]:
            lines[i] = lines[i].replace("{@add-ons/", "{~add-ons/")
        # Occasionally authors try to use '~' with [textdomain] or [binary_path].
        if "[binary_path]" in lines[i]:
            in_binary_path = True
        if "[/binary_path]" in lines[i]:
            in_binary_path = False
        if "[textdomain]" in lines[i]:
            in_textdomain = True
        if "[/textdomain]" in lines[i]:
            in_textdomain = False
        if in_binary_path or in_textdomain:
            if '~' in lines[i]:
                tilde = re.search('(^\s*path) *= *([^#]{0,5})(~/?(data/)?add-ons/)', lines[i])
                if tilde:
                    lines[i] = tilde.group(1) + '=' + tilde.group(2) + 'data/add-ons/' + lines[i][tilde.end():]
                    print('"%s", line %d: %s -> data/add-ons/ -- [textdomain] and [binary_path] paths do not accept "~" for userdata'\
                          % (filename, i+1, tilde.group(3)))
    # some tags do no longer support default side=1 attribute but may use [filter_side]
    # with a SSF instead
    # (since 1.9.5, 1.9.6)
    if missingside:
        side_one_tags_allowing_filter_side = (
            ("remove_shroud"),
            ("place_shroud"),
            ("gold"),
            ("modify_side"),
            ("modify_ai")
            )
        outside_of_theme_wml = True # theme wml contains a [gold] tag - exclude that case
        in_side_one_tag = False
        side_one_tag_needs_side_one = True
        for i in range(len(lines)):
            if "no-syntax-rewrite" in lines[i]:
                break
            precomment = lines[i].split("#")[0]
            if outside_of_theme_wml:
                if "[theme]" in precomment:
                    outside_of_theme_wml = False
            else:
                if "[/theme]" in precomment:
                    outside_of_theme_wml = True
            if outside_of_theme_wml:
                if not in_side_one_tag:
                    for tag in side_one_tags_allowing_filter_side:
                        if "[" + tag + "]" in precomment:
                            in_side_one_tag = True
                else:
                    if side_one_tag_needs_side_one:
                        if "side=" in precomment:
                            side_one_tag_needs_side_one = False
                        if "[filter_side]" in precomment:
                            side_one_tag_needs_side_one = False
                    for tag in side_one_tags_allowing_filter_side:
                        if "[/" + tag + "]" in precomment:
                            if side_one_tag_needs_side_one:
                                if verbose:
                                    print('"%s", line %d: [%s] without "side" attribute is now applied to all sides'%(filename, i+1, tag))
                                #lines.insert(i, leader(precomment) + baseindent + "side=1\n")
                            in_side_one_tag = False
                            side_one_tag_needs_side_one = True
                            break
    # More syntax transformations would go here.
    return lines

def maptransform(filename, baseline, inmap, y):
    # Transform lines in maps
    for i in range(len(inmap[y])):
        for (old, new) in mapchanges:
            inmap[y][i] = inmap[y][i].replace(old, new)

# Generic machinery starts here

def is_map(filename):
    "Is this file a map?"
    return filename.endswith(".map")

if 0: # Not used, as there are currently no defined map transforms
    class maptransform_error:
        "Error object to be thrown by maptransform."
        def __init__(self, infile, inline, type):
            self.infile = infile
            self.inline = inline
            self.type = type
        def __repr__(self):
            return '"%s", line %d: %s' % (self.infile, self.inline, self.type)

    def maptransform_sample(filename, baseline, inmap, y):
        "Transform a map line."
        # Sample to illustrate how map-transformation hooks are called.
        # The baseline argument will be the starting line number of the map.
        # The inmap argument will be a 2D string array containing the
        # entire map.  y will be the vertical coordinate of the map line.
        # You pass a list of these as the second argument of translator().
        raise maptransform_error(filename, baseline+y+1,
                             "unrecognized map element at line %d" % (y,))

tagstack = [] # For tracking tag nesting

def outermap(func, inmap):
    "Apply a transformation based on neighborhood to the outermost ring."
    # Top and bottom rows
    for i in range(len(inmap[0])):
        inmap[0][i] = func(inmap[0][i])
        inmap[len(inmap)-1][i] = func(inmap[len(inmap)-1][i])
    # Leftmost and rightmost columns excluding top and bottom rows
    for i in range(1, len(inmap)-1):
        inmap[i][0] = func(inmap[i][0])
        inmap[i][len(inmap[0])-1] = func(inmap[i][len(inmap[0])-1])

def translator(filename, mapxforms, textxform):
    "Apply mapxform to map lines and textxform to non-map lines."
    global tagstack
    gzipped = filename.endswith(".gz")
    if gzipped:
        with gzip.open(filename) as content:
            unmodified = content.readlines()
    else:
        with codecs.open(filename, "r", "utf8") as content:
            unmodified = content.readlines()
    # Pull file into an array of lines, CR-stripping as needed
    mfile = []
    map_only = filename.endswith(".map")
    terminator = "\n"
    for line in unmodified:
        if line.endswith("\n"):
            line = line[:-1]
        if line.endswith("\r"):
            line = line[:-1]
            if not stripcr:
                terminator = '\r\n'
        mfile.append(line)
        if "map_data" in line:
            map_only = False
    # Process line-by-line
    lineno = baseline = 0
    cont = False
    validate = True
    unbalanced = False
    newdata = []
    refname = None
    while mfile:
        if not map_only:
            line = mfile.pop(0)
            if verbose >= 3:
                print(line, end=terminator)
            lineno += 1
        # Check for one certain error condition
        if "{" in line and "}" in line:
            refname = line[line.find("{"):line.rfind("}")]
            # Ignore all-caps macro arguments.
            if refname == refname.upper():
                pass
            elif 'mask=' in line and not (refname.endswith("}") or refname.endswith(".mask")):
                print('"%s", line %d: fatal error, mask file without .mask extension (%s)' \
                      % (filename, lineno+1, refname))
                sys.exit(1)
        # Exclude map_data= lines that are just 1 line without
        # continuation, or which contain {}.  The former are
        # pathological and the parse won't handle them, the latter
        # refer to map files which will be checked separately.
        if map_only or (("map_data=" in line or "mask=" in line)
                        and line.count('"') in (1, 2)
                        and '""' not in line
                        and "{" not in line
                        and "}" not in line
                        and not within('time')):
            outmap = []
            have_header = have_delimiter = False
            maskwarn = False
            maptype = None
            if map_only:
                if filename.endswith(".mask"):
                    maptype = "mask"
                else:
                    maptype = "map"
            else:
                leadws = leader(line)
                if "map_data" in line:
                    maptype = "map"
                elif "mask" in line:
                    maptype = "mask"
            baseline = lineno
            cont = True
            if not map_only:
                fields = line.split('"')
                if fields[1].strip():
                    mfile.insert(0, fields[1])
                if len(fields) == 3:
                    mfile.insert(1, '"')
            if verbose >= 3:
                print("*** Entering %s mode on:" % maptype)
                print(mfile)
            # Gather the map header (if any) and data lines
            savedheaders = []
            while cont and mfile:
                line = mfile.pop(0)
                if verbose >= 3:
                    print(line, end=terminator)
                lineno += 1
                # This code supports ignoring comments and header lines
                if len(line) == 0 or line[0] == '#' or '=' in line:
                    if '=' in line:
                        have_header = True
                    if len(line) == 0:
                        have_delimiter = True
                    savedheaders.append(line + terminator)
                    continue
                if '"' in line:
                    cont = False
                    if verbose >= 3:
                        print("*** Exiting map mode.")
                    line = line.split('"')[0]
                if line:
                    if ',' in line:
                        fields = line.split(",")
                    else:
                        fields = [x for x in line]
                    outmap.append(fields)
                    if not maskwarn and maptype == 'map' and "_f" in line:
                        print('"%s", line %d: warning, fog in map file' \
                              % (filename, lineno+1))
                        maskwarn = True
            # Deduce the map type
            if not map_only:
                if maptype == "map":
                    newdata.append(leadws + "map_data=\"")
                elif maptype == "mask":
                    newdata.append(leadws + "mask=\"")
            original = copy.deepcopy(outmap)
            for transform in mapxforms:
                for y in range(len(outmap)):
                    transform(filename, baseline, outmap, y)
            newdata += savedheaders
            if have_header and not have_delimiter:
                newdata.append(terminator)
            for y in range(len(outmap)):
                newdata.append(",".join(outmap[y]) + terminator)
            # All lines of the map are processed, add the appropriate trailer
            if not map_only:
                newdata.append("\"" + terminator)
        elif "map_data=" in line and ("{" in line or "}" in line):
            newline = line
            refre = re.compile(r"\{@?([^A-Z].*)\}").search(line)
            if refre:
                mapfile = refre.group(1)
                if not mapfile.endswith(".map") and is_map(mapfile):
                    newline = newline.replace(mapfile, mapfile + ".map")
            newdata.append(newline + terminator)
            if newline != line:
                if verbose > 0:
                    print('wmllint: "%s", line %d: %s -> %s.' % (filename, lineno, line, newline))
        elif "map_data=" in line and line.count('"') > 1:
            print('wmllint: "%s", line %d: one-line map.' % (filename, lineno))
            newdata.append(line + terminator)
        else:
            # Handle text (non-map) lines.  It can use within().
            newline = textxform(filename, lineno, line)
            newdata.append(newline + terminator)
            fields = newline.split("#")
            trimmed = fields[0]
            destringed = re.sub('"[^"]*"', '', trimmed) # Ignore string literals
            comment = ""
            if len(fields) > 1:
                comment = fields[1]
            # Now do warnings based on the state of the tag stack.
            if not unbalanced:
                for instance in re.finditer(r"\[\/?\+?([a-z][a-z_]*[a-z])\]", destringed):
                    tag = instance.group(1)
                    closer = instance.group(0)[1] == '/'
                    if not closer:
                        tagstack.append((tag, {}))
                    else:
                        if len(tagstack) == 0:
                            print('"%s", line %d: closer [/%s] with tag stack empty.' % (filename, lineno+1, tag))
                        elif tagstack[-1][0] != tag:
                            print('"%s", line %d: unbalanced [%s] closed with [/%s].' % (filename, lineno+1, tagstack[-1][0], tag))
                        else:
                            if validate:
                                validate_on_pop(tagstack, tag, filename, lineno)
                            tagstack.pop()
                if tagstack:
                    for instance in re.finditer(r'([a-z][a-z_]*[a-z])\s*=(.*)', trimmed):
                        attribute, value = instance.groups()
                        if '#' in value:
                            value = value.split("#")[0]
                        tagstack[-1][1][attribute] = value.strip()
            if "wmllint: validate-on" in comment:
                validate = True
            if "wmllint: validate-off" in comment:
                validate = False
            if "wmllint: unbalanced-on" in comment:
                unbalanced = True
            if "wmllint: unbalanced-off" in comment:
                unbalanced = False
            if "wmllint: match" in comment:
                comment = comment.strip()
                try:
                    fields = comment.split("match ", 1)[1].split(" with ", 1)
                    if len(fields) == 2:
                        notepairs.append((fields[0], fields[1]))
                except IndexError:
                    pass
    # It's an error if the tag stack is nonempty at the end of any file:
    if tagstack:
        print('"%s", line %d: tag stack nonempty (%s) at end of file.' % (filename, lineno, tagstack))
    tagstack = []
    if iswml(filename):
        # Perform checks that are purely local.  This is an
        # optimization hack to reduce parsing overhead.
        for nav in WmllintIterator(newdata, filename):
            try:
                (key, prefix, value, comment) = parse_attribute(nav.text)
                local_sanity_check(filename, nav, key, prefix, value, comment)
            except TypeError:
                key = prefix = value = comment = None
                local_sanity_check(filename, nav, key, prefix, value, comment)
        # Perform file-global semantic sanity checks
        newdata = global_sanity_check(filename, newdata)
        # OK, now perform WML rewrites
        newdata = hack_syntax(filename, newdata)
        # Run everything together
        filetext = "".join(newdata)
        transformed = filetext
    else:
        # Map or mask -- just run everything together
        transformed = "".join(newdata)
    # Simple check for unbalanced macro calls
    linecount = 1
    quotecount = 0
    display_state = False
    singleline = False
    for i in range(len(transformed)):
        if transformed[i] == '\n':
            if singleline:
                singleline = False
                if not display_state and quotecount % 2 and transformed[i:i+2] != "\n\n" and transformed[i-1:i+1] != "\n\n":
                    print('"%s", line %d: nonstandard word-wrap style within message' % (filename, linecount))
            linecount += 1
        elif transformed[i-7:i] == "message" and transformed[i] == '=':
            singleline = True
        elif re.match(" *wmllint: *display +on", transformed[i:]):
            display_state = True
        elif re.match(" *wmllint: *display +off", transformed[i:]):
            display_state = False
        elif transformed[i] == '"' and not display_state:
            quotecount += 1
            if quotecount % 2 == 0:
                singleline = False
    # Return None if the transformation functions made no changes.
    if "".join(unmodified) != transformed:
        return transformed
    else:
        return None

def inner_spellcheck(nav, value, spelldict):
    "Spell-check an attribute value or string."
    # Strip off translation marks
    if value.startswith("_"):
        value = value[1:].strip()
    # Strip off line continuations, they interfere with string-stripping
    value = value.strip()
    if value.endswith("+"):
        value = value[:-1].rstrip()
    # Strip off string quotes
    value = string_strip(value)
    # Discard extraneous stuff
    replacements = (
        ("...", " "),
        ("\"", " "),
        ("\\n", " "),
        ("/", " "),
        ("@", " "),
        (")", " "),
        ("(", " "),
        ("…", " "),  # UTF-8 ellipsis
        ("—", " "),  # UTF-8 em dash
        ("–", " "),  # UTF-8 en dash
        ("―", " "),  # UTF-8 horizontal dash
        ("−", " "),  # UTF-8 minus sign
        ("’", "'"),  # UTF-8 right single quote
        ("‘", "'"),  # UTF-8 left single quote
        ("”", " "),  # UTF-8 right double quote
        ("“", " "),  # UTF-8 left double quote
        ("•", " "),  # UTF-8 bullet
        ("◦", ""),              # Why is this necessary?
        ("''", ""),
        ("female^", " "),
        ("male^", " "),
        ("teamname^", " "),
        ("team_name^", " "),
        ("UI^", " "),
        ("^", " "),
    )

    for old, new in replacements:
        value = value.replace(old, new)

    if '<' in value:
        value = re.sub("<ref>.*< ref>", "", value)
        value = re.sub("<[^>]+>text='([^']*)'<[^>]+>", r"\1", value)
        value = re.sub("<[0-9,]+>", "", value)
    # Fold continued lines
    value = re.sub(r'" *\+\s*_? *"', "", value)
    # It would be nice to use pyenchant's tokenizer here, but we can't
    # because it wants to strip the trailing quotes we need to spot
    # the Dwarvish-accent words.
    for token in value.split():
        # Try it with simple lowercasing first
        lowered = token.lower()
        if d.check(lowered):
            continue
        # Strip leading punctuation and grotty Wesnoth highlighters
        # Last char in this regexp is to ignore concatenation signs.
        while lowered and lowered[0] in " \t(`@*'%_+":
            lowered = lowered[1:]
        # Not interested in interpolations or numeric literals
        if not lowered or lowered.startswith("$"):
            continue
        # Suffix handling. Done in two passes because some
        # Dwarvish dialect words end in a single quote
        while lowered and lowered[-1] in "_-*).,:;?!& \t":
            lowered = lowered[:-1]
        if lowered and spelldict.check(lowered):
            continue;
        while lowered and lowered[-1] in "_-*').,:;?!& \t":
            lowered = lowered[:-1]
        # Not interested in interpolations or numeric literals
        if not lowered or lowered.startswith("$") or lowered[0].isdigit():
            continue
       # Nuke balanced string quotes if present
        lowered = string_strip(lowered)
        if lowered and spelldict.check(lowered):
            continue
        # No match? Strip possessive suffixes and try again.
        elif lowered.endswith("'s") and spelldict.check(lowered[:-2]):
            continue
        # Hyphenated compounds need all their parts good
        if "-" in lowered:
            parts = lowered.split("-")
            if [w for w in parts if not w or spelldict.check(w)] == parts:
                continue
        # Modifier literals aren't interesting
        if re.match("[+-][0-9]", lowered):
            continue
        # Match various onomatopoetic exclamations of variable form
        if re.match("hm+", lowered):
            continue
        if re.match("a+[ur]*g+h*", lowered):
            continue
        if re.match("(mu)?ha(ha)*", lowered):
            continue
        if re.match("ah+", lowered):
            continue
        if re.match("no+", lowered):
            continue
        if re.match("um+", lowered):
            continue
        if re.match("aw+", lowered):
            continue
        if re.match("o+h+", lowered):
            continue
        if re.match("s+h+", lowered):
            continue
        nav.printError('possible misspelling "%s"' % token)


def spellcheck(fn, d):
    "Spell-check a file using an Enchant dictionary object."
    local_spellings = []
    # Accept declared spellings for this file
    # and for all directories above it.
    up = fn
    while True:
        if not up or is_root(up):
            break
        else:
            local_spellings += declared_spellings.get(up,[])
            up = os.path.dirname(up)
    local_spellings = [w for w in local_spellings if not d.check(w)]
    #if local_spellings:
    #    print("%s: inherited local spellings: %s" % (fn, local_spellings))
    for word in local_spellings:
        d.add_to_session(word)

    # Process this individual file
    for nav in WmllintIterator(filename=fn):
        #print("element=%s, text=%s" % (nav.element, repr(nav.text)))
        # Recognize local spelling exceptions
        if not nav.element and "#" in nav.text:
            comment = nav.text[nav.text.index("#"):]
            words = re.search("wmllint: local spellings? (.*)", comment)
            if words:
                for word in words.group(1).split():
                    word = word.lower()
                    if not d.check(word):
                        d.add_to_session(word)
                        local_spellings.append(word)
                    else:
                        nav.printError("spelling '%s' already declared" % word)
    #if local_spellings:
    #    print("%s: with this file's local spellings: %s" % (fn,local_spellings))

    for nav in WmllintIterator(filename=fn):
        # Spell-check message and story parts
        if nav.element in spellcheck_these:
            # Special case, beyond us until we can do better filtering..
            # There is lots of strange stuff in text- attributes in the
            # helpfile(s).
            if nav.element == 'text=' and '[help]' in nav.ancestors():
                continue
            # Remove pango markup
            if "<" in nav.text or ">" in nav.text or '&' in nav.text:
                nav.text = pangostrip(nav.text)
            # Spell-check the attribute value
            (key, prefix, value, comment) = parse_attribute(nav.text)
            if "no spellcheck" in comment:
                continue
            inner_spellcheck(nav, value, d)
        # Take exceptions from the id fields
        if nav.element == "id=":
            (key, prefix, value, comment) = parse_attribute(nav.text)
            value = string_strip(value).lower()
            if value and not d.check(value):
                d.add_to_session(value)
                local_spellings.append(value)
    #if local_spellings:
    #    print("%s: slated for removal: %s" % (fn, local_spellings))
    for word in local_spellings:
        try:
            d.remove_from_session(word)
        except AttributeError:
            print("Caught AttributeError when trying to remove %s from dict" % word)

vctypes = (".svn", ".git", ".hg")

def interesting(fn):
    "Is a file interesting for conversion purposes?"
    return fn.endswith(".cfg") or is_map(fn) or issave(fn)

def allcfgfiles(directory):
    "Get the names of all interesting files under directory."
    datafiles = []
    if not os.path.isdir(directory):
        if interesting(directory):
            if not os.path.exists(directory):
                print("wmllint: %s does not exist" % directory, file=sys.stderr)
            else:
                datafiles.append(directory)
    else:
        for root, dirs, files in os.walk(directory):
            for vcsubdir in vctypes:
                if vcsubdir in dirs:
                    dirs.remove(vcsubdir)
            for name in files:
                if interesting(os.path.join(root, name)):
                    datafiles.append(os.path.join(root, name))
    datafiles.sort() # So diffs for same campaigns will cluster in reports
    return map(os.path.normpath, datafiles)

if __name__ == '__main__':
    parser = argparse.ArgumentParser(
        description = '''Convert Battle of Wesnoth WML from older versions to newer ones.
Also validates WML to check for errors.''',
        epilog = '''For more about wmllint, including how to prevent unwanted conversions and false
positive warnings with magic comments, read the introduction in the wmllint
file itself.
See also: http://wiki.wesnoth.org/Maintenance_tools.''',
        formatter_class = argparse.RawTextHelpFormatter
    )

    mode = parser.add_mutually_exclusive_group()
    mode.add_argument("-c", "--clean", action="store_true",
                      help="Clean up -bak files.")
    mode.add_argument("-D", "--diffs", action="store_true",
                      help="Display diffs between converted and unconverted files.")
    mode.add_argument("-d", "--dryrun", action="store_true",
                      help="List changes (-v) but don't perform them.")
    mode.add_argument("-r", "--revert", action="store_true",
                      help="Revert the conversion from the -bak files.")
    parser.add_argument("-m", "--missing", action="store_true",
                        help="""Don't warn about tags without side= keys now applying
to all sides.""")
    parser.add_argument("-s", "--stripcr", action="store_true",
                        help="Convert DOS-style CR/LF to Unix-style LF.")
    parser.add_argument("-v", "--verbose", action="count", default=0,
                        help="""-v        lists changes.
-v -v     names each file before it's processed.
-v -v -v  shows verbose parse details.""")
    parser.add_argument("-K", "--known", action="store_true",
                        help="""Suppress check for unknown unit types, recruits, races,
scenarios, etc.""")
    parser.add_argument("-S", "--nospellcheck", action="store_false",
                        help="Suppress spellchecking")
    parser.add_argument("-Z", "--stringfreeze", action="store_true",
                        help="Suppress repair attempts of newlines in messages")
    # -f --future has been removed; there have been no experimental conversions since 1.4
    # -p --progress has been removed; similar to existing -v -v
    parser.add_argument("directories", action="store", nargs="*",
                        help="""Any number of directories. Each directory is converted.
If no directories are specified, acts on the current
directory.""")

    namespace = parser.parse_args()
    clean = namespace.clean
    diffs = namespace.diffs
    dryrun = namespace.dryrun
    missingside = namespace.missing
    revert = namespace.revert
    stringfreeze = namespace.stringfreeze
    stripcr = namespace.stripcr
    verbose = namespace.verbose
    dospellcheck = namespace.nospellcheck # WARNING! We store the opposite of the value needed!
    inconsistency = namespace.known
    arguments = namespace.directories # a remnant of getopt...

    if dryrun:
        verbose = max(1, verbose)

    post15 = False

    def hasdigit(str):
        for c in str:
            if c in "0123456789":
                return True
        return False

    def texttransform(filename, lineno, line):
        "Resource-name transformation on text lines."
        original = line
        # Perform line changes
        if "wmllint: noconvert" not in original:
            for (old, new) in linechanges + mapchanges:
                line = line.replace(old, new)
            # Perform any base terrain string conversions needed in
            # [terrain_type] aliasof=, mvt_alias=, and def_alias= attributes.
            if under("terrain_type"):
                match = re.search(r"\b(?:aliasof|mvt_alias|def_alias)\s*=(.*)$", line)
                if match:
                    aliases = match.group()
                    for (old, new) in aliaschanges:
                        aliases = aliases.replace(old, new)
                    line = line.replace(match.group(), aliases)
        # Perform tag renaming for 1.5.  Note: this has to happen before
        # the sanity check, which assumes [unit] has already been
        # mapped to [unit_type].  Also, beware that this test will fail to
        # convert any unit definitions not in conventionally-named
        # directories -- this is necessary in order to avoid stepping
        # on SingleUnitWML in macro files.  The post15 flag expresses whether
        # we've seen a [unit_type] and can therefore assume the files have
        # undergone 1.4 -> 1.5 conversion.
        global post15
        if "units" in filename and not post15:
            if '[unit_type]' in line:
                post15 = True
            else:
                line = line.replace("[unit]", "[unit_type]")
                line = line.replace("[+unit]", "[+unit_type]")
                line = line.replace("[/unit]", "[/unit_type]")
        # Handle SingleUnitWML or Standard Unit Filter or SideWML
        # Also, when macro calls have description= in them, the arg is
        # a SUF being passed in.
        if tagstack and ((under("unit") and "units" not in filename) or \
               standard_unit_filter() or \
               under("side") or \
               re.search("{[A-Z]+.*description=.*}", line)):
            if "id" not in tagstack[-1][1] and "_" not in line:
                line = re.sub(r"\bdescription\s*=", "id=", line)
            if "name" not in tagstack[-1][1]:
                line = re.sub(r"user_description\s*=", "name=", line)
            if "generate_name" not in tagstack[-1][1]:
                line = re.sub(r"generate_description\s*=", "generate_name=", line)
        # Now, inside objects...
        if under("object") and "description" not in tagstack[-1][1]:
            line = re.sub(r"user_description\s*=", "description=", line)
        # Alas, WML variable references cannot be converted so
        # automatically.
        if ".description" in line:
            print('"%s", line %d: .description may need hand fixup' % \
                  (filename, lineno))
        if ".user_description" in line:
            print('"%s", line %d: .user_description may need hand fixup' % \
                  (filename, lineno))
        # In unit type definitions
        if under("unit_type") or under("female") or under("unit"):
            line = line.replace("unit_description=", "description=")
            line = line.replace("advanceto=", "advances_to=")
        # Inside themes
        if within("theme"):
            line = line.replace("[unit_description]", "[unit_name]")
        # Report the changes
        if verbose > 0 and line != original:
            msg = "%s, line %d: %s -> %s" % \
                  (filename, lineno, original.strip(), line.strip())
            print(msg)
        return line

    try:
        # If a backslash comes before a quote, it will be interpreted as an
        # escape to a literal quote rather than a Windows directory delimiter,
        # causing Windows to find "no such file or directory". This block deals
        # with this issue, but it is impossible to handle all cases if multiple
        # (intended) arguments are involved. We also activate globbing on
        # Windows, if there is a wildcard.
        if arguments and sys.platform == 'win32':
            wildcard = False
            ugly = False
            newargs = []
            for i, arg in enumerate(arguments):
                if not wildcard and '*' in arg:
                    wildcard = True
                    from glob import glob
                if '"' in arg:
                    if len(arguments) - i > 1:
                        ugly = True
                        break
                    test = arg.split('"', 1)
                    if " " in test[1].lstrip():
                        ugly = True
                        break
                    print('\n\nWARNING!! A backslash followed by a quote (\\") is \
interpreted not as a directory separator and an argument delimiter, but as an escape \
to a literal quote. Two quotes together ("") will also be interpreted as including a \
literal quote. Your system sees the file/directory you are targeting as:\n\n%s\n\nAlthough \
wmllint believes it can resolve this particular instance, please do not repeat this \
in the future. (If you are using a final backslash at the end of your argument, it \
is not necessary; also, frontslashes will also be recognized as directory separators.)\n\n'%arg, file=sys.stderr)
                    arguments.remove(arg)
                    arg = re.sub(r'([^ ])"([^ ])', r'\1\\\2', arg)
                    for new in arg.rstrip('"').split('"'):
                        arguments.insert(i, new.strip())
                        i += 1
                        print('wmllint: resolving address as: %s' % new.strip(), file=sys.stderr)
            if ugly:
                print("""
    WARNING!! A backslash followed by a quote (\\"), or two quotes (""), is misinterpreted by your system to mean you want a literal quote character, not a path encloser:

    %s

    After exiting this message: hit the up arrow key, edit your command, and press Enter.
""" % re.sub(r'"', '-->>"<<--', arg), file=sys.stderr)

                moreugly = input('Press "H" if you need more help, or Enter to exit: ')
                if moreugly.lower().startswith('h'):
                    print("""
        Explanation:

Windows' use of the backslash as a directory separator is clashing with the use of the backslash as an escape. As an escape, the backslash tells your system that you want a normally special character to be its literal self (or sometimes, that you want a normally ordinary character to have a special meaning). Your system interprets '\\"' as an escape for a literal quote character. Two quotes together are also interpreted as a literal quote.

'"Campaign\\"' is interpreted as 'Campaign"' instead of 'Campaign\\'.
'"My Folder\\"Campaign' is interpreted as 'My Folder"Campaign' (not 'My Folder\Campaign').
'"My Folder\Campaign\\" "My Folder\Another_Campaign"' is interpreted as 'My Folder\Campaign" My' and 'Folder\Another_Campaign'.

In your case, your system interprets your arguments as:

    %s

        Solutions:

(1) If your problem is adjoining quotes, delete (or move) one of them
(  "folder\Campaign"" -> "folder\Campaign"  )

(2) If you are using a final backslash at the end of a directory address, it is not necessary
(  "folder\Campaign\\" -> "folder\Campaign" -- but NOT "folder\Campaign\\"file.cfg -> "folder\Campaign"file.cfg  )

(3) If you add a second backslash, your system will then escape the backslash instead of the quote
(  "folder\Campaign\\" -> "folder\Campaign\\\\"  )

(4) Frontslashes will also be recognized as directory separators
(  "folder\Campaign\\" -> "folder/Campaign"  )

(5) If there are no spaces in your address, it is not necessary to use quotes
(  "folder\Campaign\\" -> folder\Campaign -- but NOT "My Folder\Campaign\\" -> My Folder\Campaign  )

(6) You can move the affected quote
(  M"y Folder\\"Campaign -> M"y Fold"er\Campaign -- but NOT Fil"es and Folders\My Folder\\"Campaign -> Fil"es and Folders\M"y Folder\Campaign  )

        Reminder:

(a) Hit the up arrow key. The up and down arrows move through your command history; one press of the up arrow will take you to your last command.
(b) Edit your command. Use any of the solutions described above.
(c) Press Enter""" % repr(arguments)[1:-1], file=sys.stderr)
                sys.exit(2)
            if wildcard:
                for arg in arguments:
                    for wild in glob(arg):
                        newargs.append(wild)
                if newargs:
                    arguments = newargs
                else:
                    print('wmllint: wildcard did not match any files or directories (%s)'%arguments, file=sys.stderr)
                    sys.exit(1)

        if not arguments:
            arguments = ["."]

        for directory in arguments:
            ofp = None
            for fn in allcfgfiles(directory):
                if verbose >= 2:
                    print(fn + ":")
                backup = fn + "-bak"
                if clean or revert:
                    # Do housekeeping
                    if os.path.exists(backup):
                        if clean:
                            print("wmllint: removing %s" % backup)
                            if not dryrun:
                                os.remove(backup)
                        elif revert:
                            print("wmllint: reverting %s" % backup)
                            if not dryrun:
                                if sys.platform == 'win32':
                                    os.remove(fn)
                                os.rename(backup, fn)
                elif diffs:
                    # Display diffs
                    if os.path.exists(backup):
                        fromdate = time.ctime(os.stat(backup).st_mtime)
                        todate = time.ctime(os.stat(fn).st_mtime)
                        with codecs.open(backup, "r", "utf8") as fromlines, \
                             codecs.open(fn, "r", "utf8") as tolines:
                            diff = difflib.unified_diff(fromlines.readlines(),
                                                        tolines.readlines(),
                                                        backup, fn, fromdate, todate, n=3)
                        sys.stdout.writelines(diff)
                else:
                    if "~" in fn:
                        print("wmllint: ignoring %s, the campaign server won't accept it." % fn)
                        continue
                    # Do file conversions
                    try:
                        changed = translator(fn, [maptransform], texttransform)
                        if changed:
                            print("wmllint: converting", fn)
                            if not dryrun:
                                if sys.platform == 'win32' and os.path.exists(backup):
                                    os.remove(backup)
                                os.rename(fn, backup)
                                if fn.endswith(".gz"):
                                    with gzip.open(fn, "w") as ofp:
                                        ofp.write(changed)
                                else:
                                    with codecs.open(fn, "w", "utf8") as ofp:
                                        ofp.write(changed)
                    #except maptransform_error, e:
                    #    print("wmllint: " + repr(e), file=sys.stderr)
                    except:
                        print("wmllint: internal error on %s" % fn, file=sys.stderr)
                        (exc_type, exc_value, exc_traceback) = sys.exc_info()
                        raise exc_type(exc_value).with_traceback(exc_traceback)
        if not clean and not diffs and not revert:
            # Consistency-check everything we got from the file scans
            if not inconsistency:
                consistency_check()
            # Attempt a spell-check
            if dospellcheck:
                try:
                    import enchant
                    d = enchant.Dict("en_US")
                    checker = d.provider.desc
                    if checker.endswith(" Provider"):
                        checker = checker[:-9]
                    print("# Spell-checking with", checker)
                    for word in declared_spellings["GLOBAL"]:
                        d.add_to_session(word.lower())
                    for directory in arguments:
                        ofp = None
                        for fn in allcfgfiles(directory):
                            if verbose >= 2:
                                print(fn + ":")
                            spellcheck(fn, d)
                except ImportError:
                    print("wmllint: spell check unavailable, install python-enchant to enable", file=sys.stderr)
    except KeyboardInterrupt:
        print("Aborted")

# wmllint ends here