Skip to content
Commits on Mar 23, 2016
  1. @phooji

    [pfff] unbreak parsing `$$variable` with a quick lexer hack

    phooji committed Mar 23, 2016
    Summary:
    Managed to break this in D3005435. The fix is to explicitly
    look for `$$name` cases in the lexer, and push the dollar
    token separate from the `VARIABLE`. This way, the parser
    gets the tokens it would see prior to D3005435.
    
    Test Plan:
      ~/pfff $ make test
      ~/pfff $ ./pfff -dump_php ./tests/php/parsing/dynamic_var.php
    
    **Output**
    
      [StmtList(
       [ExprStmt(
             Assign(IdVar(DName(("o", i_1)), Ref(NoScope)), i_2,
                     Sc(C(String(("foo", i_3))))), i_4);
                         ExprStmt(
                               Assign(IdVar(DName(("foo", i_5)), Ref(NoScope)), i_6,
                                       Sc(C(Int(("42", i_7))))), i_8);
                                           Echo(i_9, [Left(Deref(i_10, IdVar(DName(("o", i_11)), Ref(NoScope))))],
                                                 i_12)]); FinalDef(i_13)]
    
    (Previous output after D3005435 was to barf :()
    
    Reviewers: logozzo, dreeves
    
    Differential Revision: https://phabricator.fb.com/D3089181
  2. @phooji

    [pfff] add parser support for $$ dollardollar support for |> pipe in …

    phooji committed Mar 2, 2016
    …Hack/PHP
    
    Summary:
    Added to the lexer and wrapped as a regular identifier
    using `$` as the name.
    
    Test Plan:
        $ make clean
        $ ./configure -novisual
        $ make depend
        $ make
        $ make test
    
    Reviewers: logozzo
    
    Subscribers: dominik, hansonw
    
    Differential Revision: https://phabricator.fb.com/D3005435
  3. @phooji

    [pfff] add support for the |> pipe operator for Hack/PHP

    phooji committed Mar 2, 2016
    Summary: Pretty much like title. Adding `$$` in a separate diff.
    
    Test Plan:
      $ make clean
      $ ./configure -novisual
      $ make depend
      $ make
      $ make test
    
    Reviewers: logozzo, zacharyl
    
    Subscribers: dominik, hansonw
    
    Differential Revision: https://phabricator.fb.com/D3005436
Commits on Mar 4, 2016
  1. @aryx

    Merge pull request #139 from damiendoligez/patch-1

    aryx committed Mar 4, 2016
    Remove `-warn-error +a` in released builds.
Commits on Mar 2, 2016
  1. @damiendoligez

    Remove `-warn-error +a` in released builds.

    damiendoligez committed Mar 2, 2016
    Don't use `-warn-error +a` in released source. If you want to be warning-clean, you should activate the flag only in debug/development mode. Here I'm proposing a simple solution that relies on GNUmake features. You might want to do it in the configure script instead.
    
    As it is now, pfff 0.29 breaks on OCaml 4.03.0+beta1 because of this line (and the use of `String.capitalize`, which is now deprecated).
Commits on Jan 21, 2016
  1. @aryx
  2. @aryx
  3. @aryx

    * changes.txt: misc

    aryx committed Jan 21, 2016
  4. @aryx
Commits on Jan 8, 2016
  1. @aryx

    Merge pull request #138 from qoelet/master

    aryx committed Jan 8, 2016
    Improve codegraph path handling
  2. @qoelet

    Improve codegraph path handling

    qoelet committed Jan 9, 2016
Commits on Dec 18, 2015
  1. @phooji

    [pfff] add support for parsing `??`

    phooji committed Dec 17, 2015
    Summary:
    Similar to the existing `?:` support; produces similar AST nodes.
    
    This is a shortcut and actually has the wrong associativity, since `??` is
    right-associative and `?:` is left-associative. We can fix that if we run into
    issues; for now we really just need to not throw parse errors on PHP 7
    features like this.
    
    Test Plan:
    This is a quick fix to ensure that we don't break things
    as people start using `??` around the codebase.
    
    Reviewers: mqian, hansonw, leegao
    
    Reviewed By: leegao
    
    Subscribers: dominik, hansonw, alexeyp
    
    Differential Revision: https://phabricator.fb.com/D2770864
    
    Signature: t1:2770864:1450414537:44ccdccf5eda9ade67b08d99686c424d891f3c32
Commits on Nov 4, 2015
  1. @aryx

    Merge pull request #134 from mmcco/master

    aryx committed Nov 4, 2015
    Remove GNU make version check
  2. @mmcco

    Build systems very often (usually?) don't check the make version.

    mmcco committed Nov 4, 2015
    Regardless, the existing check is really brittle. For example, only GNU
    grep has the --colour flag, and it's available on the BSDs only as
    ggrep. Even when fixing this, the make version check doesn't work for me
    on OpenBSD. It's probably easiest just to remove it.
    
    This patch fixes configuration on OpenBSD for me.
    
    Also corrects a typo.
Commits on Sep 12, 2015
  1. @aryx
  2. @aryx

    * code_map/Visual.tex.nw: misc

    aryx committed Sep 12, 2015
  3. @aryx
Commits on Aug 31, 2015
  1. @aryx

    * code_map/view2.ml: +> to |>

    aryx committed Aug 31, 2015
  2. @aryx

    CEs on Typedef constructor removal

    aryx committed Aug 31, 2015
  3. @aryx
  4. @aryx

    * h_program-lang/highlight_code.ml: get rid of Typedef,

    aryx committed Aug 31, 2015
    redudant with entity_code.ml
Commits on Aug 24, 2015
  1. @aryx

    Merge pull request #132 from juergenhoetzel/master

    aryx committed Aug 24, 2015
    * lang_js/analyze/module_js.ml: Fix invalid documentation comments.
  2. @juergenhoetzel
Commits on Aug 20, 2015
  1. @aryx

    * travis-ci.sh: yes

    aryx committed Aug 20, 2015
  2. @aryx

    * travis-ci.sh:

    aryx committed Aug 20, 2015
  3. @aryx

    * travis-ci.sh: install camlp4

    aryx committed Aug 20, 2015
  4. @aryx
  5. @aryx

    * lang_js/analyze/utils_js.mli:

    aryx committed Aug 20, 2015
    fix issue #130
  6. @aryx

    * changes.txt:

    aryx committed Aug 20, 2015
Commits on Aug 5, 2015
  1. @aryx

    * lang_nw/parsing/parse_nw.ml: cleanup, remove program2, and give

    aryx committed Aug 5, 2015
    priority to chapter/section over [[]] in highlighter
  2. @aryx
  3. @aryx
  4. @aryx
  5. @aryx
  6. @aryx
Something went wrong with that request. Please try again.