As Behdad suggested, we can dramatically speed up the test suite by
short-circuiting the write to a png file, only to then immediately read it
back in. So for the raster based surfaces, we avoid the round-trip through
libpng by implementing a new boilerplate method to directly extract the image
buffer from the test result. A secondary speedup is achieved by caching the
most recent reference image.
By keeping a static reference to the user font face, it is erroneously kept
alive during a call to cairo_debug_reset_static_data(). (A violation of
the caller's contract to ensure that no active reference to a cairo object
is held by the caller.)
When cleaning files, limit the glob to only match *-out.{pdf,ps,svg} in
order to distinguish the current practice of naming vector output files
from possible future vector target/reference files.
Construct the test name to pass to the boilerplate creation routines such
that it uniquely identifies the test in terms of test, target, content and
pass (similar, offset, thread). This allows the vector targets to create
output different output files for each test, whereas before, later tests
would overwrite existing files making debugging more difficult.
Currently fallback-resolution is included in the test suite if we have all
of the vector surfaces available. This commit enables individual support
for the vector surfaces, so that the test can be run even if one or more
of the surfaces are not available.
In order to run under memfault, the framework is first extended to handle
running concurrent tests - i.e. multi-threading. (Not that this is a
requirement for memfault, instead it shares a common goal of storing
per-test data). To that end all the global data is moved into a per-test
context and the targets are adjusted to avoid overlap on shared, global
resources (such as output files and frame buffers). In order to preserve
the simplicity of the standard draw routines, the context is not passed
explicitly as a parameter to the routines, but is instead attached to the
cairo_t via the user_data.
For the masochist, to enable the tests to be run across multiple threads
simply set the environment variable CAIRO_TEST_NUM_THREADS to the desired
number.
In the long run, we can hope the need for memfault (runtime testing of
error paths) will be mitigated by static analysis. A promising candidate
for this task would appear to be http://hal.cs.berkeley.edu/cil/.
And update user-font text_to_glyphs() method to match.
Currently disable the win32-font text_to_glyphs(), until that one
is updated. Or better yet, remove it and implement ucs4_to_index().
It's the toy font API afterall.
The init func does not actually need to draw anything, but having a cairo_t
similar to that passed to render_glyph is handy for computing font extents.
This is because cairo makes doing some things really hard (if not impossible)
without a cairo_t.
The user-font-proxy test case is a great example of how the added cairo_t
makes life much easier.
The PDF output no longer requires a PDF specific reference image. The
lastest poppler from git is required to get the fix for a bug in Type
3 font rendering.
This way the same callback code can be used to render multiple different
glyph arrays. Change done for education purposes, otherwise doesn't
make any difference in the test.
10000 that is. xlib fails now again. Not because of glyph size issues.
Because we skip rendering any glyphs with positions not in range -1024..15359.
Working on a fix.
