rlutz / geda-gaf

Scripts executed by the exec action now receive the current context.
Also reverted the memory leak fix in s_slot.c to match the master
branch as a better version of the fix lives on the mem_leak branch
and having different versions in two places is bad.

Found several bugs and some untested error cases so fixed/added those
as well. Don't really recall them all.

- Most things in gEDA/gaf are case-sensitive, so it makes sense to treat
  SAB arguments as case-sensitive, as well.

Done with a caveat. Becasue of the special meaning of the 'none' context
it was left as case-insensitive. All others are now case sensitive.

- You don't need "nobase_" on your "example_DATA" target (you don't have
  files in subdirectories), but it doesn't hurt, either.

Fixed it anyway

- Header files (especially public ones) should only include those other
  header files which they minimally require in order to work.  AFAICT,
  for "sab.h", this would be just "struct.h".

Fixed.

- Don't assign variables inside function calls unless you have a good
  reason to do so.  Assigning from "strtok" inside "while" in sab.c is one
  of the few cases where this is ok (as the assignment can't be put on a
  separate line without hurting readability); most other cases are not.

<sigh> Fixed as far as I know. The one in s_slot.c (part of my memory leak
fix) will be fixed as part of the broader memory leak fixes on the master
branch.

- "TOPLEVEL *" variables are called "toplevel" (all lower-case) in the
  existing code, so it's a good idea to follow this convention for new
  code, as well.

Fixed

- gEDA/gaf Python code should reside within the "gaf" top-level package.
  You basically have two choices for organizing the SAB code: either
  (1) have individual sub-modules of gaf by placing the code files in
      "xorn/src/python/", then import "gaf.sab" and call
      "gaf.sab.process", or
  (2) group the SAB source as a sub-package by placing the code in
      "xorn/src/python/sab/", then import "gaf.sab.sab" and call
      "gaf.sab.sab.process"
  I'd go with the first option, but either way is okay.

Went with option #1

- You are processing the "--sab-context" command-line argument in
  "netlist.py", so it is not necessary to process it in "gnetlist.in",
  as well.  You can just pass it through unchanged.

Fixed

- User-visible messages should be wrapped in _("...") in order to be
  translatable.

Fixed in python...

- Since you use the tuple (order, action, parms) in multiple places, it
  may be worth defining a collections.namedtuple for that purpose.

Done

- Don't use something like 1000000000 as a special value.  If you have
  to store the absence of a value in-band, use something that cannot
  ever occur as a valid value, like -1 or None.  (And if you *really*
  have to use a magic value somewhere, please define it as a constant.)

Fixed

- In modify_refdes.py, the variable "refdes" is either a list or a
  string.  Since the expression "'C' in ..." works in both cases, I
  suggest replacing "list(levels.upper())" with "levels.upper()".

Done

- You don't need types.StringType and types.ListType;
  "isinstance(..., str)" and "isinstance(..., list)" work just fine.

Fixed as far as I know

- Using "dir(m)" to determine whether a variable is defined in a module
  feels a bit too hacky.  I'm not sure yet how I'd prefer this to be
  solved; if you absolutely have to make the code depend on whether or
  not there is such a variable, use it and catch an AttributeError.

<sigh> Fixed...still makes my skin crawl

- Instead of:      ''.zfill(4 * indent).replace('0', ' ')
  you can just do: '    ' * indent

Done

sab would crash if there was nothing to do for the specified
context.

Fixed a bunch (most? [maybe all? nah..]) of the formatting issues.

Changed SAB to non-verbose output, warnings are still issued.
Fixed a potential problem by checking to make sure contexts are strings

        modified:   xorn/src/sab/sab.py

This patch shows how to convert an existing backend to enable SAB
processing.

The requirments:
The identifier must be 'SAB_CONTEXT'
The identifier must refer to a list of strings

Generally a single line is all that is needed.

	modified:   xorn/src/backend/gnet_partslist1.py

I forgot to update the README in the examples/SAB directory which
had a TODO listed for the slotting issue addressed by the last commit.
Also added the printout TODO.

	modified:   examples/SAB/README

1. Added isolated SAB code. Given a component object the code extracts
all the sab-param attributes and parses them (ignoring any that are
malformed) producing a data structure. The contents of that data structure
can then be modified. After modification the structure can be converted
back into sab-param attributes which replace the source attributes. Any
attributes that were ignored for being malformed are left untouched.
Currently the only supported modification is to replace bypass pin #s.

	modified:   libgeda/include/Makefile.am
	new file:   libgeda/include/libgeda/sab.h
	modified:   libgeda/src/Makefile.am
	new file:   libgeda/src/sab.c

2. Add calls to the SAB system to s_slot_update_object so the bypass
pin numbers are updated when the slot is changed. Note when manually
changing the slot number the bypass pin changes are reflected in the
sab-param shown in the schematic but not the one shown in the
'Attributes' panel. That one is updated if the component is deselected
and then reselected causeing the panel to reload the attributes. This
should certainly be fixed but right now I don't have the time to figure
out where the change needs to be made.
Also identified a memory leak so fixed that as well.

	modified:   libgeda/src/s_slot.c

TODO: Need to figure out how to have the sab-param attributes affect
printout. I didn't do it because I really don't need it but mostly
because I couldn't come up with good solutions to:
1) How to specify which context you are producing hardcopy for.
2) How to draw bypass connections on a device with more than two
   pins and/or where the pins are not directly across from each other.
3) How to handle exec actions in hardcopy.

Maybe somebody smarter/more motivated than me can figure out a solution.
Adding the SAB functionality to gschem was really only a convience item anyway.

…o do it.

1. Added optional field to the sab-param attribute
	modified:   xorn/src/sab/sab.py
2. Added test cases
	modified:   examples/SAB/sab.sch
3. Updated README accordigly
	modified:   examples/SAB/README

1. Added exec test cases, tweaked README, and yet more revision of test cases
	modified:   examples/SAB/README
	modified:   examples/SAB/sab.sch
2. Added test/example script
	modified:   examples/SAB/Makefile.am
	new file:   examples/SAB/modify_refdes.py
3. Since processing can now occur prior to netlist generation it may be useful to have a do nothing backend
	modified:   xorn/src/backend/Makefile.am
	new file:   xorn/src/backend/gnet_null.py
4. Implement exec action
	modified:   xorn/src/sab/sab.py
	modified:   xorn/src/sab/sab_utils.py

1. Broke out the utility functions so they can be more cleanly used by external scripts (exec to be added)
	modified:   xorn/src/sab/Makefile.am
	modified:   xorn/src/sab/sab.py
	new file:   xorn/src/sab/sab_utils.py
2. Added/modified various test cases to sab.sch
	modified:   examples/SAB/sab.sch
3. Implement the bypass action (modifications to sab_util.py)

Note: bypass makes use of the gaf.netlist.net.merge_into function which appears to be undocumented, but worked very nicely.

1. Extend description of SAB markup
	modified:   README
2. Add test cases to schematic
	modified:   sab.sch
3. Fix a bug that caused gnetlist to pass an empty string as the sab-context
	modified:   ../../xorn/src/command/gnetlist.in
4. Implement discard action
	modified:   ../../xorn/src/sab/sab.py

1. add a new attribute 'sab-param' to the list of attributes in gschem
	modified:   gschem/lib/system-gschemrc.scm
2. set 'sab-param=ab:discard' as the default for spice-title-1.sym
	modified:   symbols/spice/spice-title-1.sym
3. add new parameter 'sab-context' to gnetlist w/ basic checks & pass it to xorn
	modified:   xorn/src/command/gnetlist.in
4. add new package 'sab' to hold the SAB processing code
	modified:   xorn/Makefile.am
	modified:   xorn/configure.ac
	new file:   xorn/src/sab/Makefile.am
	new file:   xorn/src/sab/__init__.py
	new file:   xorn/src/sab/sab.py
5. cause xorn-netlist to use --sab-context if provided, look for a context in the backend
   and call the SAB code if a context is avalable, passing the netlist and context list
	modified:   xorn/src/command/netlist.py
6. add a description of SAB and a test-bed/example
	modified:   configure.ac
	modified:   examples/Makefile.am
	new file:   examples/SAB/Makefile.am
	new file:   examples/SAB/README
	new file:   examples/SAB/sab.sch

gtk_text_view_scroll_to_iter doesn't check if the text view has a
layout; if it hasn't, an assertion in gtk_text_layout_get_iter_location
is violated, and the program aborts.  In order to avoid that, check if
the text view is realized before trying to scroll it.