ADG Git Change Log

Age Message
6 months 22 hours build: use INTROSPECTION_{GIRDIR,TYPELIBDIR}

Rely on gobject-introspection provided variables instead of handcraft a
similar solution in configure. They live in introspection.m4 at least
since gobject-introspection-0.7.0.
6 months 2 days build: GTK+3 is called libgtk-3-dev in TravisCI
6 months 2 days build: enable GTK+2 and GTK+3 build testing in TravisCI
6 months 2 days cpml: cast to int in primitive to avoid clang warning
6 months 2 days build: include Lua support in the NSIS installer

Install typelibs, girfiles, Lua and everything needed to run the ADG
canvas via gobject-introspection and LGI. Still to be tested though.
6 months 2 days build: rename USRDIR to HOSTPREFIX

HOSTPREFIX is much more guessable because similar to the PREFIX variable
commonly used by the installation process.
6 months 2 days build: initial import of build/build-win{32,64}

Provide better MinGW cross-compilation scripts that cross-build with
gobject-introspection enabled.
6 months 2 days build: share dist flags

The flags for configuring dist releases [1] are now shared. Avoid
guessing the proper number of jobs to pass and use "make -j" without
value instead.

[1] "make distcheck" and "make installer"
6 months 26 days doc: correct typo in NEWS
6 months 27 days doc: update DOAP metadata
6 months 27 days Merge branch 'master' of
6 months 28 days i18n: make update-po
6 months 28 days build: include cairo-gobject in docs/adg

The gtk-doc scanner needs to linked against the cairo-gobject library.
This was working anyway on Archlinux but Ubuntu seems to be more picky
on the library order, so without this commit the doc generation was
6 months 28 days build: bump version to 0.8.0
6 months 28 days doc: update NEWS
6 months 28 days doc: update TODO
7 months 5 days cpml: drop CPML_GNUC_CONST macro

Actually the main function to get the primitive class from the type
number relies on a hash table, so further optimization are bogus.

Dropped the macro equivalent to G_GNUC_CONST because it is no more
needed (and because I suspect that macro was useless anyway).
Avoid relying on th
7 months 5 days cpml: use hash table to get the primitive class

Instead of a serie of if/else, use an hash table to get the class (a
struct filled with function pointers) from the primitive type (an
integer number).

Lowered the CPML_ARC value to keep that table small.
7 months 6 days tests: check fillets and chamfers in AdgPath
7 months 7 days tests: include GTK+ tests for AdgCanvas

The inclusion of config.h was missing, hence the GTK2_ENABLED or
GTK3_ENABLED defines were always undefined.

Added missing piece and corrected errors in tests.
7 months 7 days build: install GTK+2 in travis

By installing GTK+2, all GTK+ dependent tests are enabled also on
7 months 7 days tests: check adg_canvas_export()
7 months 7 days adg: correct size in adg_canvas_export()

Previously the above function was getting the canvas size through
adg_canvas_get_size(). This is the wrong API because it is used to get the
explicit size (that could possibly be 0), not the real size.

Use the proper API instead, i.e. adg_entity_get_extents().
7 months 8 days tests: randomly enable extents rendering
7 months 8 days tests: ensure AdgEntity has a parent when mixing locals