Bug 1365460 introduced code paths behind MOZ_DEBUG #ifdefs, but
MOZ_DEBUG is never defined, while it is available in CONFIG in
moz.builds. This is kind of a confusing situation, but the fact that
we've been able to avoid those problems for so long would tend to
put the blame on mozjemalloc, and fixes should go there.
Except that bug 1261161 explains that the only existing alternative
(the DEBUG #define), as used in MFBT, is not working for spidermonkey,
so it actually makes sense to converge to MOZ_DEBUG rather than DEBUG.
So start defining MOZ_DEBUG globally, fixing the mozjemalloc issues of
not having the debug code enabled. Bug 1261161 can then take care of
changing the DEBUG #ifdefs.
--HG--
extra : rebase_source : 37e3d03ac8350c62c8059d4ca01d1ecfdf5f421a
The way inlining is disabled in mozjemalloc is via a #define of "inline"
to nothing, which is a dubious way to do that. This makes the compiler
trigger warnings we -Werror on for some static functions. While there
are such functions in mozjemalloc.cpp that could be fixed by wrapping
them in the right #ifdefs, there are also others coming from headers,
and it's not something that can be fixed in a satisfactory way.
The right way to disable inlining is to pass the right compiler flags
for that. But inlining is the least of the problems to debug optimized
C++ code, so it feels like if debugging requires some optimization
tweaking, it should be done manually with compile flags when needed,
instead of fiddling with #defines to remove keywords.
--HG--
extra : rebase_source : 962c3409f86060c4d5ddf966778b58b64f89c31d
Instead of unconditionally pushing and popping clips per display item,
this patch changes things so that for each recursive display list, we
create an ItemClips struct. We push this onto the stack when we enter
the display list, and pop it off at the end. For each display item, we
check to see if the clips would actually change compared to the previous
display item, and only do the pop/repush in that case.
MozReview-Commit-ID: GadIp2J8TrA
--HG--
extra : rebase_source : ba64c6b4659b8e51cab19b807088b9a50d71b85a
This makes ScrollingLayersHelper a non-RAII type class, and instead adds
methods to notify it of when we start processing a new transaction or a
new display item within the transaction. This patch has no functional
changes, it's non-obvious refactoring.
MozReview-Commit-ID: 3yq9sPiHMge
--HG--
extra : rebase_source : 286423f56de59211e320f015cb1004a1e98332b8
Storing the per-item clip state in a struct like this will allow us to
easily compare the desired clip state across items, so we can avoid
doing unnecessary work when going from one item to the next. This patch
has no functional changes, it's just refactoring.
MozReview-Commit-ID: 49B6hmsWZ4V
--HG--
extra : rebase_source : 8ac4bbf039f81bc2d26e92924ed041fa3d18e5ba
Instead just keep a ref to it as a member variable. No functional
change.
MozReview-Commit-ID: 5fccUlSifsA
--HG--
extra : rebase_source : a14672f926c383354db76e553ae23613fe4a432a
Most of this change is just fiddling with function signatures so that they take
a LayerManager* instead of a Layer* (or in some cases, both). This allows
the WebRender codepaths to pass a WebRenderLayerManager* instead of having to
produce a Layer* which it doesn't have.
MozReview-Commit-ID: Fb0C8OUVDin
--HG--
extra : rebase_source : e4c3324cfa20c295db85d5c09df8d8d77865bb6a
When server responses a field value with null, we have to ignore it, or we fail to SetRequestMethod for the http channel.
MozReview-Commit-ID: EW9NHMy22Fc
