[seed] Created tag SEED_3_8_0

[Tim Horton <hortont src gnome org>]

The unsigned tag 'SEED_3_8_0' was created.

Tagger: Tim Horton <hortont424 gmail com>
Date: Tue Apr 16 00:59:33 2013 -0700

    Tag 3.8

Changes since the last tag 'SEED_3_2_0':

Alan Knowles (25):
      fix #659548 - stop CFLAGS getting unconditionally overwritten
      allow [ GObject.GTYPE_INT, 10 ] syntax to work again for not-null elements
      fix most of #660644 - except function-info.js (my gdb is broken)
      fix #663289 - gtype signal parsing not resetting param count
      fix #663609 - documentation on seed_value_* now matches actual return values
      fix #653077 - fix for webkit 1.0 configure (Dominique Leuenberger)
      Add an xorg extension (experimental) - at present only has one function to get the screen idle time...
      xorg module - close display after using it
      xorg module - libxss check has bad error message
      Very hacky support for <array length="x" zero-teminated="0" c:type="gchar**"><type name="..."/>, used 
by GLib.file_get_contents in latest girs
      tidy up of binary string support, should now solve quite a few of the short string issues with NULL 
chars in
      enum types are not all long any more, we need to check the type and recurse to return the right type..
      rename with_length to _full, since relay of calling for enums sent the wrong data, and got the wrong 
result..
      finally get the code right to handle array length offset information
      free binary strings always, this is a bit risky, as it assumes that when we see an array of c uints and 
its used as a binary string, we should always free it... - Im not sure if it should really check transfer 
settings.. - but this will crash badly if it is not right...
      refactor main call code, remove quite a few redundant ifdefs, keep track of caller allocations so they 
can be freed on failure
      initialize and free jchar in binary string, Im still not sure this is the correct way to do string 
creation, need to run some tests to ensure it really works
      fix ref code for iface_info broken by refactoring
      creating a JSRef to find the length value appears to cause segfaults - replacing with simpler code.. -
      remove unused variables / compiler warnings
      major memory reductions by garbage collecting at end of closure (example was timeout loops) - it might 
be worth exposing this somehow so  user space so it can be called manually...
      check if object is abstract before construction - thanks to Ryan Lortie
      add backtrace stack to the exception string
      part fix to bug #694458 - first step to getting make check working - passes most tests, just need to 
add fix module lookup path
      Adding ondemand for reference -

Alban Crequy (5):
      fix binding GStrv to javascript value
      SeedEngine: add missing pthread_key_create
      seed_init_with_context: reuse existing context group
      implement hashtables conversion between js and gi values
      support signals with annotated types

Alexandre Mazari (1):
      Add a GLib namespace extension providing comodities for conversion from/to GVariant.

Eric Le Bihan (1):
      Added support for uint8 array as return type

Funda Wang (1):
      fix #659910 - fix linkage against libgio - broken by webkit going...

Guillaume Desmottes (2):
      add seed_importer_add_search_path()
      seed_prepare_global_context: don't include Seed.js if not needed

Gustavo Noronha Silva (2):
      Append user data to the new arguments array, not the source one
      Remove duplicated code, call function instead

Sunil Mohan Adapa (1):
      Fix #637248 - Add support for implicitly converting GObject to GValue

Tim Horton (2):
      Postrelease version bump.
      Seed 3.8