This changes the way we deal with page use counters so that we can
handle out of process iframes.
Currently, when a parent document is being destroyed, we poke into all
of the sub-documents to merge their use counters into the parent's page
use counters, which we then report via Telemetry. With Fission enabled,
the sub-documents may be out of process. We can't simply turn these
into async IPC calls, since the parent document will be destroyed
shortly, as might the content processes holding the sub-documents.
So instead, each document during its initialization identifies which
ancestor document it will contribute its page use counters to, and
stores its WindowContext id to identify that ancestor. A message is
sent to the parent process to notify it that page use counter data will
be sent at some later point. That later point is when the document
loses its window. It doesn't matter if the ancestor document has
already been destroyed at this point, since all we need is its
WindowContext id to uniquely identify it. Once the parent process has
received all of the use counters it expects to accumulate to a given
WindowContext Id, it reports them via Telemetry.
Reporting of document use counters remains unchanged and is done by each
document in their content process.
While we're here, we also:
* Limit use counters to be reported for a pre-defined set of document
URL schemes, rather than be based on the document principal.
* Add proper MOZ_LOG logging for use counters instead of printfs.
Differential Revision: https://phabricator.services.mozilla.com/D87188
We use it to merge blobs, but that doesn't match the spec. Per:
https://html.spec.whatwg.org/#the-list-of-available-images
We should key off the URI, and thus a revoked image should be able to be
reused if the request is cached. This works to allow printing revoked
blob images, which other browsers allow.
I don't see any way to make that work easily while preserving the blob
image optimization. That being said, it seems other browsers also
re-decode when creating different URLs for the same blob (see the
test-case attached to the bug), so it seems we should be able to live
without it.
Differential Revision: https://phabricator.services.mozilla.com/D90544
This patch was generated by running:
```
perl -p -i \
-e 's/^(\s+)([a-zA-Z0-9.]+) = NS_ConvertUTF8toUTF16\((.*)\);/\1CopyUTF8toUTF16(\3, \2);/;' \
-e 's/^(\s+)([a-zA-Z0-9.]+) = NS_ConvertUTF16toUTF8\((.*)\);/\1CopyUTF16toUTF8(\3, \2);/;' \
$FILE
```
against every .cpp and .h in mozilla-central, and then fixing up the
inevitable errors that happen as a result of matching C++ expressions with
regexes. The errors fell into three categories:
1. Calling the convert functions with `std::string::c_str()`; these were
changed to simply pass the string instead, relying on implicit conversion
to `mozilla::Span`.
2. Calling the convert functions with raw pointers, which is not permitted
with the copy functions; these were changed to invoke `MakeStringSpan` first.
3. Other miscellaneous errors resulting from over-eager regexes and/or the
replacement not being type-aware. These changes were reverted.
Differential Revision: https://phabricator.services.mozilla.com/D88903
The abstract observer base classes are moved to a separate header file
nsRefreshObservers.h and the includes are adjusted accordingly.
Some method implementations are moved to the corresponding implementation files
to avoid the need to include the nsRefreshDriver.h file in the header.
Differential Revision: https://phabricator.services.mozilla.com/D85764
Having two classes in the inheritance chain inherit from SupportsWeakPtr
now won't compile, but you can use WeakPtr<Derived> when any base class
inherits from SupportsWeakPtr.
Differential Revision: https://phabricator.services.mozilla.com/D83674
Make them perform the image load (if needed), instead of copying the
image requests from the original document.
This is needed for CSS for stuff like:
@media print {
#foo::before {
content: url(bar.png);
}
}
And so on. For images, we should do this as well. Nothing prevents you
from doing:
<picture>
<source srcset="print.png" media="print">
<source srcset="screen.png" media="not print">
<img>
</picture>
And that should in theory work. It works after this patch, and I added a
test for that.
This patch is a bit bigger than I'd like, but I didn't find a more
reasonable way to split it up.
Making static docs able to do image loads is most of the patch and is
mostly straight-forward. This allows to remove the hacky "change the
loading document" thing that CSS images do, which is just working around
the CSP of the print document.
I need to enable background colors in printpreview_helper so as to be
able to have a reference page for all the different image types.
Differential Revision: https://phabricator.services.mozilla.com/D81779
Make them perform the image load (if needed), instead of copying the
image requests from the original document.
This is needed for CSS for stuff like:
@media print {
#foo::before {
content: url(bar.png);
}
}
And so on. For images, we should do this as well. Nothing prevents you
from doing:
<picture>
<source srcset="print.png" media="print">
<source srcset="screen.png" media="not print">
<img>
</picture>
And that should in theory work. It works after this patch, and I added a
test for that.
This patch is a bit bigger than I'd like, but I didn't find a more
reasonable way to split it up.
Making static docs able to do image loads is most of the patch and is
mostly straight-forward. This allows to remove the hacky "change the
loading document" thing that CSS images do, which is just working around
the CSP of the print document.
I need to enable background colors in printpreview_helper so as to be
able to have a reference page for all the different image types.
Differential Revision: https://phabricator.services.mozilla.com/D81779
If you do a diff, all of the different shared surfaces get merged together,
because the diffing process removes all PIDs from the paths. This means
you end up hitting an assertion when the diffing process tries to combine
entries with the same path but different kinds.
Differential Revision: https://phabricator.services.mozilla.com/D83585
Also: adjust include paths to be consistent for usages of various SVG headers,
and remove unused SVG includes (mostly for "utils" classes),
and drop stray "ns" from already-renamed SVG classes in various code comments.
Differential Revision: https://phabricator.services.mozilla.com/D83140
Current gecko dispatches tasks of releasing image::Image to main thread at some places. The task was dispatched individually. Then there were cases that the releasing took long time. It increased peak memory usage and caused a problem like Bug 1639280. When main thread is very busy like WebGL, it takes longer time until the Image is released on main thread. If Images are released quickly on main thread, we could reduce peak memory usage.
When SurfaceCache::ReleaseImageOnMainThread() is called, there could be already an ongoing task for releasing Image. It could reduce a duration until release on main thread.
Differential Revision: https://phabricator.services.mozilla.com/D82864
If the scrollable div ends up as the first-encountered scrollframe, it gets a
big displayport that is just over 5000px CSS pixels tall (on GeckoView). When we
check for images being "nearly visible" we expand that displayport further by a
pref-controlled multiplier, which ends up doubling the displayport height to
over 10000 CSS pixels. So for the image to actually get detected as outside the
"nearly visible rect" it needs to be positioned further down in the iframe.
This test was previously passing on GeckoView because the scrollable div was
not getting detected as the first-encountered scrollframe; the root scrollframe
for the top-level document was the one getting the displayport even though it
didn't really need it. A later patch in this bug changes the first-encountered
scrollframe, causing this test to fail without this patch.
Differential Revision: https://phabricator.services.mozilla.com/D82777
It's possible only `OnStartRequest()` and `OnStopRequest()` was called, but no `OnDataAvailable()`. If so, when `OnStopRequest()` is being called, we would have no image but the channel status is a success.
Therefore, before setting the image container, we should also ensure we get the image already.
Differential Revision: https://phabricator.services.mozilla.com/D80744
We did not pass FLAG_HIGH_QUALITY_SCALING down to the document used to
rasterize the SVG to a surface, resulting in embedded raster images to
not use high quality downscaling.
Differential Revision: https://phabricator.services.mozilla.com/D56422
This test fails already on central if you change this line:
https://searchfox.org/mozilla-central/rev/027893497316897b8f292bde48dbb6da2391a331/image/test/unit/test_private_channel.js#93
by:
Ci.nsIReferrerInfo.EMPTY
Because the previous part of the test populates the image cache. The
test wants to check that the channel for the image load is properly
flagged as private and thus that the http cache is partitioned
appropriately, thus clearing the image caches seems sane.
While at it, also fix it so that we send a valid image instead of base64
text, though that change is not technically needed so feel free to ask
me to remove it.
Differential Revision: https://phabricator.services.mozilla.com/D79848
Make the stylesheet cache respect the same headers as the image cache
does. This makes no-cache stylesheets work as they do now, which is
useful for developers that want to develop sites locally, and for
shift-reloads, etc.
Differential Revision: https://phabricator.services.mozilla.com/D78659
We really should be color managing them. It was a mistake and
misunderstanding of how Chrome worked because of differences in how we
color manage canvas.
Differential Revision: https://phabricator.services.mozilla.com/D79295
We can perform the same function as RecyclingSourceSurface by checking
the ref count of the underlying surface directly. We need to ensure
WebRender is explicitly aware that it is a recycled surface, but that is
easily achieved by changing the type of the surface. This avoids
unnecessary heap allocations, particularly in the case where many
elements on the same page refer to the same animation (and thus
duplicating RecyclingSourceSurface objects).
Differential Revision: https://phabricator.services.mozilla.com/D77513
Make the stylesheet cache respect the same headers as the image cache
does. This makes no-cache stylesheets work as they do now, which is
useful for developers that want to develop sites locally, and for
shift-reloads, etc.
Differential Revision: https://phabricator.services.mozilla.com/D78659
If we hit an OOM or similar when creating a SourceBuffer for an image,
we will hit COMPLETE immediately as the next read that goes to the
SourceBuffer from a SourceBufferIterator. If the iterator has yet to do
a read, that means it will hit COMPLETE right away without giving
anything. This patch fixes the WebP decoder to handle this case, as well
as adds test cases for all of the current image decoders.
Differential Revision: https://phabricator.services.mozilla.com/D77131
Dynamically toggle support for displaying image/webp and image/avif according
to the values of the image.webp.enabled and image.avif.enabled prefs.
Differential Revision: https://phabricator.services.mozilla.com/D76977
There's no use case for stateful comparators, so they can be just plain
function pointers.
This is used in some hot places like CSS selector matching.
Differential Revision: https://phabricator.services.mozilla.com/D77084
This drops `followLinks` as 1) it has been no-op on UNIX for a long time and 2) its Windows implementation never had a proper symlink support.
Differential Revision: https://phabricator.services.mozilla.com/D75569
This drops `followLinks` as 1) it has been no-op on UNIX for a long time and 2) its Windows implementation never had a proper symlink support.
Differential Revision: https://phabricator.services.mozilla.com/D75569
Also, add `clang-format off` directives to files which are ignored by
.clang-format-ignore so that the editor isn't trying to reformat them
Differential Revision: https://phabricator.services.mozilla.com/D76217
Since AnimationState::UpdateState now depends on a specific frame of the animation being in the surface cache we need to call UpdateState whenever we add a new frame.
Depends on D73583
Differential Revision: https://phabricator.services.mozilla.com/D73584
It takes a frame count argument but probably never gets a value we haven't already seen in a NotifyProgress call, so this doesn't fix anything as far as I know.
Depends on D73564
Differential Revision: https://phabricator.services.mozilla.com/D73583
The img.decode promise for animated images that aren't in the document might not resolve in common cases. Specifically for animated images that have been decoded and then discarded, so that mCompositedFrameInvalid is true.
mCompositedFrameInvalid is cleared by either:
1) the image is fully decoded and NotifyDecodeComplete is called.
2) RequestRefresh is able to advance to the current frame.
We don't call RequestRefresh on animated images that aren't in the document and we only decode a certain number of frames ahead of the current frame (even for fully retained animation buffers), so neither of these will happen with long enough animated images.
So we need to make sure that mCompositedFrameInvalid eventually gets cleared any time a decode is requested (which the FrameAnimator gets notified about via GetCompositedFrame). We do that be calling UpdateState in GetCompositedFrame whenever mCompositedFrameInvalid is true and the image is in the surface cache.
This runs into a second problem: UpdateState uses |mIsCurrentlyDecoded = aResult.Surface().IsFullyDecoded()| to determine when to clear mCompositedFrameInvalid. As mentioned above, we don't fully decode animated images. Further, for animated images that are large enough we use a discarding animation buffer that never keeps around all the frames and so IsFullyDecoded always returns false. So we replace that with a check that we can seek to the current animated frame index.
Differential Revision: https://phabricator.services.mozilla.com/D73564
Raw Cr.ERROR don't get stack information, same as throwing JS literals instead
of `new Error()`s.
This was done automatically with a new eslint rule that will be introduced in
the next commit. One instance of a raw Cr.ERROR was not replaced since it is
used in a test that specifically checks the preservation of raw Cr values in
XPCJS. The rule will be disabled for that instance.
Differential Revision: https://phabricator.services.mozilla.com/D28073
It appears some websites assume we will not color manage untagged images
and encode metadata in the image's surface data. Chrome matches this
behaviour. We should probably do the same for webcompat.
Differential Revision: https://phabricator.services.mozilla.com/D73737
There are many limitations currently, but this prototype should successfully
render most basic AVIF images. Known limitations:
- No support for any derived image items (crop, rotate, etc.)
- No support for alpha planes
- No support for ICC profiles (bug 1634741)
- The primary image item must be an av01 (no grid support)
- HDR images aren't tone-mapped
Differential Revision: https://phabricator.services.mozilla.com/D68498
There are many limitations currently, but this prototype should successfully
render most basic AVIF images. Known limitations:
- No support for any derived image items (crop, rotate, etc.)
- No support for alpha planes
- The primary image item must be an av01 (no grid support)
- HDR images aren't tone-mapped
Differential Revision: https://phabricator.services.mozilla.com/D68498
There are many limitations currently, but this prototype should successfully
render most basic AVIF images. Known limitations:
- No support for any derived image items (crop, rotate, etc.)
- No support for alpha planes
- The primary image item must be an av01 (no grid support)
- HDR images aren't tone-mapped
Differential Revision: https://phabricator.services.mozilla.com/D68498
Since we don't support downscaling animated images we'll need something like this even we were to try to request a smaller sized decode.
Differential Revision: https://phabricator.services.mozilla.com/D71523
We invalidate for more things than just frames advancing (if we are redecoding and still showing the same frame as before discarding for example). So there is no reason that the dirty rect and a frame being advanced should be linked.
Differential Revision: https://phabricator.services.mozilla.com/D71483
This is actually a one line change. The rest is just removing unneeded parameters.
The composited frame used to be valid for animations that had finished because we saved a separate composited frame, and since the animation had finished it was the final frame, so it had to be valid to display.
Now we don't have a separate composited frame so we only have a valid frame to display if we've decoded up to the current frame. For non-finished animations RequestRefresh will clear mCompositedFrameInvalid, but for finished animations we have to clear it, the easiest place is when we know we are fully decoded.
This problem caused img.decode to never fulfill or reject because we never sent out any frame update notifications because mCompositedFrameInvalid was always false because mAnimationFinished was always true. So we didn't get the invalidation that flipping mCompositedFrameInvalid to false when a full decode finished would cause.
Differential Revision: https://phabricator.services.mozilla.com/D70838
We invalidate for more things than just frames advancing (if we are redecoding and still showing the same frame as before discarding for example). So there is no reason that the dirty rect and a frame being advanced should be linked.
Differential Revision: https://phabricator.services.mozilla.com/D71483
This is actually a one line change. The rest is just removing unneeded parameters.
The composited frame used to be valid for animations that had finished because we saved a separate composited frame, and since the animation had finished it was the final frame, so it had to be valid to display.
Now we don't have a separate composited frame so we only have a valid frame to display if we've decoded up to the current frame. For non-finished animations RequestRefresh will clear mCompositedFrameInvalid, but for finished animations we have to clear it, the easiest place is when we know we are fully decoded.
This problem caused img.decode to never fulfill or reject because we never sent out any frame update notifications because mCompositedFrameInvalid was always false because mAnimationFinished was always true. So we didn't get the invalidation that flipping mCompositedFrameInvalid to false when a full decode finished would cause.
Differential Revision: https://phabricator.services.mozilla.com/D70838
We need this since nsLayoutUtils::SurfaceFromElement expects the
returned frame size to be correct, and we are now wrapping a source
element's image with an OrientedImage.
Differential Revision: https://phabricator.services.mozilla.com/D71243
This makes EXIF orientation metadata honored by default.
Introduce OrientedPixel and UnorientedPixel typed rects and sizes and
use them throughout RasterImage so that we don't confuse which we want.
The reason for doing this rather than having the imgLoader wrap every
RasterImage it creates with an OrientedImage is that returning the
wrapper messes with various notifications, as OrientedImage is not an
ImageResource.
(It would be even better if the JPEG decoder could decode to imgFrames
handling the EXIF orientation itself, but that's a more complicated
change.)
Differential Revision: https://phabricator.services.mozilla.com/D70273
RasterImage will make use of them.
Note that there is one bug fix in this patch, which is that
OrientedImage::OrientSurface now creates a surface of the correct size.
(Previously this code was creating a surface with the underlying
image's size, rather than the correctly oriented size. But we must
not have been calling into that code with our current uses of
OrientedImage.)
Differential Revision: https://phabricator.services.mozilla.com/D70271
This makes EXIF orientation metadata honored by default.
Introduce OrientedPixel and UnorientedPixel typed rects and sizes and
use them throughout RasterImage so that we don't confuse which we want.
The reason for doing this rather than having the imgLoader wrap every
RasterImage it creates with an OrientedImage is that returning the
wrapper messes with various notifications, as OrientedImage is not an
ImageResource.
(It would be even better if the JPEG decoder could decode to imgFrames
handling the EXIF orientation itself, but that's a more complicated
change.)
Differential Revision: https://phabricator.services.mozilla.com/D70273
--HG--
extra : moz-landing-system : lando
RasterImage will make use of them.
Note that there is one bug fix in this patch, which is that
OrientedImage::OrientSurface now creates a surface of the correct size.
(Previously this code was creating a surface with the underlying
image's size, rather than the correctly oriented size. But we must
not have been calling into that code with our current uses of
OrientedImage.)
Differential Revision: https://phabricator.services.mozilla.com/D70271
--HG--
extra : moz-landing-system : lando
This fixes the web-observable bits, but still not the context menu. Patch
incoming for that.
Differential Revision: https://phabricator.services.mozilla.com/D70936
--HG--
extra : moz-landing-system : lando
Currently, `NotificationObserverWrapper` would leak because no one delete it. When clearing the weak pointer in `ImageDecoderListener`, that won't destroy `NotificationObserverWrapper` itself, but only destroy the `WeakReference<NotificationObserverWrapper>`.
Therefore, we should remove `NotificationObserverWrapper` and use a strong reference to hold `imgINotificationObserver` directly.
Differential Revision: https://phabricator.services.mozilla.com/D70569
--HG--
extra : moz-landing-system : lando
We use FRAME_UPDATE and FRAME_COMPLETE notifications to check if we should resolve/reject img.decode promises. But all of the FRAME_UPDATE/FRAME_COMPLETE notifications for image/test/crashtests/1443232-1.gif come before we change the surface cache entry from a placeholder to available. So RequestDecodeForSize returns false and we don't resolve and we don't get anymore notifications.
In this case the DECODE_COMPLETE notifications comes after we mark the surface cache entry as available so checking if we should resolve for DECODE_COMPLETEs would also work, but it seems reasonable for consumers to expect that they can get a frame after getting a FRAME_COMPLETE notification for it.
Differential Revision: https://phabricator.services.mozilla.com/D70312
--HG--
extra : moz-landing-system : lando
Currently we can only use the gfx.color_management.force_srgb pref to
force all images to sRGB, or just accept device space. It would be nice
to be able to test device space in our tests, as well as sRGB. This
patch adds a surface flag which allows us to selectively output sRGB.
This will also be useful for clipboard and re-encoding purposes, since
they want a neutral output. In an ideal world we would just output the
color profile and the pixel data in the original color space, but for
now this is a relatively simple approach that works on all platforms and
interops well with all applications.
Differential Revision: https://phabricator.services.mozilla.com/D65734
--HG--
extra : moz-landing-system : lando
WillDrawOpaqueNow tries to answer the question "if Draw is called on this image 'pretty soon' from now will it draw something opaque (so I know if I can cull stuff behind it for painting)".
But OrientedImage wants "is the frame I have right now opaque (so I can choose the right format for a surface I'm trying to draw the flipped image into)".
Bug 1260324 replaces imgIContainer::IsOpaque with WillDrawOpaquaNow, but it should have considered this case a bit more.
RasterImage::GetFrame already checks if the surface is finished before returning, and we have the format on the returned frame, so we don't need anything complicated here.
Differential Revision: https://phabricator.services.mozilla.com/D70303
--HG--
extra : moz-landing-system : lando
Just waiting for the load event of the image doesn't even mean the image has been painted at all yet, but by the time MozReftestInvalidate fires the first frame should be guaranteed to have been painted.
Also, switch to using new img.decode api. img.decode will decode an animated version of the img, whereas drawing to a canvas will kick off a first frame only decode which the page can't use.
Differential Revision: https://phabricator.services.mozilla.com/D70104
--HG--
extra : moz-landing-system : lando
Loading just the image results in "theatre mode" display of the image (centered, dark background). But the test needs to be in an html document. We either changed how top level images were displayed or stopped using a style sheet in reftests that displayed images like this since the tests were last enabled.
Differential Revision: https://phabricator.services.mozilla.com/D70103
--HG--
extra : moz-landing-system : lando
Just waiting for the load event of the image doesn't even mean the image has been painted at all yet, but by the time MozReftestInvalidate fires the first frame should be guaranteed to have been painted.
Also, switch to using new img.decode api. img.decode will decode an animated version of the img, whereas drawing to a canvas will kick off a first frame only decode which the page can't use.
Differential Revision: https://phabricator.services.mozilla.com/D70104
--HG--
extra : moz-landing-system : lando
Loading just the image results in "theatre mode" display of the image (centered, dark background). But the test needs to be in an html document. We either changed how top level images were displayed or stopped using a style sheet in reftests that displayed images like this since the tests were last enabled.
Differential Revision: https://phabricator.services.mozilla.com/D70103
--HG--
extra : moz-landing-system : lando
TabGroup never really made any difference in which thread something go
dispatched to. This was the intended use, but development of TabGroups
with abstract main threads never made it that far. The good thing is
that thish makes it safe to also remove to the SystemGroup and instead
switch all SystemGroup dispatches to dispatches to main thread.
Timers for setTimeout and workers were the sole users of wrapped and
throttled event targets, that those throttled queues have been moved
to the BrowsingContextGroup and are now accessed explicitly.
The SchedulerEventTarget has been removed, since there are no longer a
separate event target for every TaskCategory. Instead a
LabellingEventTarget has been added to DocGroup to handle the case
where an event is dispatched do DocGroup or when an AbstractThread is
created using a DocGroup. This means that we'll actually label more
events correctly with the DocGroup that they belong to.
DocGroups have also been moved to BrowsingContextGroup.
Depends on D67636
Differential Revision: https://phabricator.services.mozilla.com/D65936
--HG--
extra : moz-landing-system : lando
This patch also tries to remove the event target entirely if it would
default to the main thread on a null event target.
Depends on D67634
Differential Revision: https://phabricator.services.mozilla.com/D67635
--HG--
extra : moz-landing-system : lando
To be able to remove SystemGroup, NS_ReleaseOnMainThreadSystemGroup
needs to have its dependency on SystemGroup removed. Since all
releases using SystemGroup would've released on the main thread anyway
we can safely replace NS_ReleaseOnMainThreadSystemGroup with
NS_ReleaseOnMainThread.
Depends on D64390
Differential Revision: https://phabricator.services.mozilla.com/D67631
--HG--
extra : moz-landing-system : lando
Mostly a matter of:
rg -l '\->LoadingPrincipal' | xargs sed -i 's/->LoadingPrincipal/->GetLoadingPrincipal/g'
And then clang-format. But I tweaked manually nsHttpChannelAuthProvider (move
the variable where it's used, don't take a useless strong ref),
AddonContentPolicy (move the declaration of the variable to the if condition),
and BackgroundUtils (same).
Differential Revision: https://phabricator.services.mozilla.com/D69828
--HG--
extra : moz-landing-system : lando
Mostly a matter of:
rg -l '\->LoadingPrincipal' | xargs sed -i 's/->LoadingPrincipal/->GetLoadingPrincipal/g'
And then clang-format. But I tweaked manually nsHttpChannelAuthProvider (move
the variable where it's used, don't take a useless strong ref),
AddonContentPolicy (move the declaration of the variable to the if condition),
and BackgroundUtils (same).
Differential Revision: https://phabricator.services.mozilla.com/D69828
--HG--
extra : moz-landing-system : lando
Currently we can only use the gfx.color_management.force_srgb pref to
force all images to sRGB, or just accept device space. It would be nice
to be able to test device space in our tests, as well as sRGB. This
patch adds a surface flag which allows us to selectively output sRGB.
This will also be useful for clipboard and re-encoding purposes, since
they want a neutral output. In an ideal world we would just output the
color profile and the pixel data in the original color space, but for
now this is a relatively simple approach that works on all platforms and
interops well with all applications.
Differential Revision: https://phabricator.services.mozilla.com/D65734
--HG--
extra : moz-landing-system : lando
gfx::Color is currently misused in many places. The DrawTargets expect
the color space to be in device space, e.g. what we are actually going
to draw using. Everything sitting above generally deals with sRGB, as
specified in CSS. Sometimes we missed the conversion from sRGB to device
space when issuing draw calls, and similarly sometimes we converted the
color to device space twice.
This patch splits the type in two. sRGBColor and DeviceColor now
represent sRGB and device color spaces respectively. DrawTarget only
accepts DeviceColor, and one can get a DeviceColor from an sRGBColor via
the ToDeviceColor helper API. The reftests now pass with color
management enabled for everything (e.g. CSS) instead of just tagged
raster images.
There will be a follow up patch to enable color management everywhere by
default on all supported platforms.
Differential Revision: https://phabricator.services.mozilla.com/D64771
--HG--
extra : moz-landing-system : lando
In the PNG decoder, when an image is tagged as sRGB, we create a new
sRGB profile, and a new transform to map that to device space. Creating
a transform is fairly expensive, and we already cache sRGB to device
space transforms in gfxPlatform. This patch ensures that we use the
cached transforms instead. It should be noted that we ignore the intent
from the sRGB tag; nominally we could transform differently as a result
of that, but qcms does not actually implement any of that.
Differential Revision: https://phabricator.services.mozilla.com/D65320
--HG--
extra : moz-landing-system : lando
Given that we are going to add ContentBlockingAllowList in
CookieSettings, so CookieSettings will be responsible for more stuff than the
cookie behavior and cookie permission. We should use a proper name to
reflect the purpose of it. The name 'CookieSettings' is misleading that
this is only for cookie related stuff. So, we decide to rename
'CookieSettins' to 'CookieJarSettings' which serves better meaning here.
Differential Revision: https://phabricator.services.mozilla.com/D63935
--HG--
rename : netwerk/cookie/CookieSettings.cpp => netwerk/cookie/CookieJarSettings.cpp
rename : netwerk/cookie/nsICookieSettings.idl => netwerk/cookie/nsICookieJarSettings.idl
extra : moz-landing-system : lando
This patch adds support for color management in BMPs added in
BITMAPV4HEADER and BITMAPV5HEADER. While display of BMPs is relatively
rare this comes in handy when interacting with the Windows clipboard.
Differential Revision: https://phabricator.services.mozilla.com/D64880
--HG--
extra : moz-landing-system : lando
This patch adds support for color management in BMPs added in
BITMAPV4HEADER and BITMAPV5HEADER. While display of BMPs is relatively
rare this comes in handy when interacting with the Windows clipboard.
Differential Revision: https://phabricator.services.mozilla.com/D64880
--HG--
extra : moz-landing-system : lando
On some RDL situations we may create the new item, then destroy the old one
afterwards.
When this is the order of operations, the image would end up unregistered, and
thus not invalidating the canvas frame.
Differential Revision: https://phabricator.services.mozilla.com/D64995
--HG--
extra : moz-landing-system : lando
This patch converts the BMP decoder to use SurfacePipe instead of using
AllocateFrame and Downscaler directly. As a result, it now uses the
accelerated premultiplication path, honours the
SurfaceFlags::NO_PREMULTIPLY_ALPHA flag, and allows for a path forward
to support color management and clipboard better.
Differential Revision: https://phabricator.services.mozilla.com/D64866
--HG--
extra : moz-landing-system : lando
We are seeing some crash reports with RecyclingSourceSurface containing
null pointers in the signature. The proto signature doesn't make any
sense however. Let's use NotNull to reassure ourselves that there isn't
something going wrong when we create RecyclingSourceSurface.
Differential Revision: https://phabricator.services.mozilla.com/D64066
--HG--
extra : moz-landing-system : lando
Currently, the GetCMSOutputProfile() and related methods pass their output
using the old C-style "ptr, len" parameters. This makes them more difficult
to deal with later in this change when they need to be safely passed over IPC.
This refactors them to return nsTArray<uint8_t> results instead. I also
removed some old cruft and refactored the existing code.
Differential Revision: https://phabricator.services.mozilla.com/D63583
--HG--
extra : moz-landing-system : lando
With the new 'defaults' key being added, 'default-preferences' is now
redundant. This commit converts all existing uses of 'default-preferences' to
use 'defaults' instead.
Differential Revision: https://phabricator.services.mozilla.com/D63247
--HG--
extra : moz-landing-system : lando
This removes the need for explicit #ifdef NS_BUILD_REFCNT_LOGGING without
introducing user-defined destructors when it is not defined.
Also, some uses of virtual for declaring destructors are replaced by the
appropriate override declaration through these changes.
Differential Revision: https://phabricator.services.mozilla.com/D62604
--HG--
extra : moz-landing-system : lando
This removes the need for explicit #ifdef NS_BUILD_REFCNT_LOGGING without
introducing user-defined destructors when it is not defined.
Also, some uses of virtual for declaring destructors are replaced by the
appropriate override declaration through these changes.
Differential Revision: https://phabricator.services.mozilla.com/D62604
--HG--
extra : moz-landing-system : lando
Gets rid of `NS_NewThread`. Where it was used in testing, I gave the new named threads names relevant to their tests.
Differential Revision: https://phabricator.services.mozilla.com/D62475
--HG--
extra : source : 541b98270c9985c5bd3569ff3ff8bc6c3d3c650a
Gets rid of `NS_NewThread`. Where it was used in testing, I gave the new named threads names relevant to their tests.
Differential Revision: https://phabricator.services.mozilla.com/D62475
--HG--
extra : moz-landing-system : lando
The trickier part is that we represent -moz-image-rect as a Rect() type instead
of image with non-null clip-rect. So we need to add a bit of code to
distinguish "image request types" from other types of images.
But it's not too annoying, and we need to do the same for fancier images like
image-set and such whenever we implement it, so seems nice to get rid of
most explicit usages of nsStyleImage::GetType().
Differential Revision: https://phabricator.services.mozilla.com/D62164
--HG--
extra : moz-landing-system : lando
We include it everywhere because it's included from gfxTypes.h.
This should avoid including all the generated bindings _everywhere_.
Differential Revision: https://phabricator.services.mozilla.com/D62174
--HG--
extra : moz-landing-system : lando
This patch adds reporting the surface types used by the image frame in a
bit mask (such if it is a CAPTURE including a DATA_SHARED, the mask will
be 1 << CAPTURE | 1 << DATA_SHARED), as well as an estimated size
included in the report as decoded-unknown for when we do not know if the
surface is on the heap or the non-heap specifically. This is the default
implementation for a SourceSurface as well, so we should no longer have
the case where surfaces appear empty despite being in the cache. It also
makes requests being validated as always notable for reporting purposes.
Differential Revision: https://phabricator.services.mozilla.com/D61458
--HG--
extra : moz-landing-system : lando
On Windows, optimized surfaces can become invalid due to a device reset
or GPU process crash when D2D draw targets were being used. This is
because SourceSurfaceD2D1 checks if the D2D context has changed from the
one it was optimized for.
We assumed when creating a DrawableFrameRef that if we don't have
imgFrame::mRawSurface, then we must have an imgFrame::mOptSurface. This
is incorrect and causes SurfaceCache to believe it has a valid
DrawableSurface, even if the underlying optimized surface cannot be used
for drawing (or has already been freed).
If we can't draw the image, it goes missing from the screen. But since
the cache thinks it has a valid surface, we never actually redecode it.
With this patch, we now check if the imgFrame::mOptSurface has been
freed already, and if not, that it remains valid for drawing purposes.
If either condition is true, the SurfaceCache entry will be removed,
thus permitting a redecode attempt.
Differential Revision: https://phabricator.services.mozilla.com/D60834
--HG--
extra : moz-landing-system : lando
This patch adds an annotation to an image cache entry if it is in the
process of being validated. This is very useful to know when debugging
missing images because state notifications are suppressed when we are
validating and may prevent the image from being displayed.
Differential Revision: https://phabricator.services.mozilla.com/D60526
--HG--
extra : moz-landing-system : lando
This patch adds error and progress tracker states to the memory dump. It
also now will include requests that have yet to create an image, and
fills in what information it is able to without it (such as URI, error
and progress tracker states).
Additionally the notability of an image has changed. If there are any
errors, missing progress trackers, or incomplete surfaces, they will
also now be notable. This is a departure from just the memory footprint.
Differential Revision: https://phabricator.services.mozilla.com/D60495
--HG--
extra : moz-landing-system : lando
This dumps more information about the surface cache. Each surface has a
flag indicating whether or not it has finished decoding; incomplete
surfaces are now marked as such in the cache reports. We now also have
counters for various insertion errors, as well as the composition of the
cache (total vs locked images, total vs locked surfaces).
Differential Revision: https://phabricator.services.mozilla.com/D60450
--HG--
extra : moz-landing-system : lando
This was done by:
This was done by applying:
```
diff --git a/python/mozbuild/mozbuild/code-analysis/mach_commands.py b/python/mozbuild/mozbuild/code-analysis/mach_commands.py
index 789affde7bbf..fe33c4c7d4d1 100644
--- a/python/mozbuild/mozbuild/code-analysis/mach_commands.py
+++ b/python/mozbuild/mozbuild/code-analysis/mach_commands.py
@@ -2007,7 +2007,7 @@ class StaticAnalysis(MachCommandBase):
from subprocess import Popen, PIPE, check_output, CalledProcessError
diff_process = Popen(self._get_clang_format_diff_command(commit), stdout=PIPE)
- args = [sys.executable, clang_format_diff, "-p1", "-binary=%s" % clang_format]
+ args = [sys.executable, clang_format_diff, "-p1", "-binary=%s" % clang_format, '-sort-includes']
if not output_file:
args.append("-i")
```
Then running `./mach clang-format -c <commit-hash>`
Then undoing that patch.
Then running check_spidermonkey_style.py --fixup
Then running `./mach clang-format`
I had to fix four things:
* I needed to move <utility> back down in GuardObjects.h because I was hitting
obscure problems with our system include wrappers like this:
0:03.94 /usr/include/stdlib.h:550:14: error: exception specification in declaration does not match previous declaration
0:03.94 extern void *realloc (void *__ptr, size_t __size)
0:03.94 ^
0:03.94 /home/emilio/src/moz/gecko-2/obj-debug/dist/include/malloc_decls.h:53:1: note: previous declaration is here
0:03.94 MALLOC_DECL(realloc, void*, void*, size_t)
0:03.94 ^
0:03.94 /home/emilio/src/moz/gecko-2/obj-debug/dist/include/mozilla/mozalloc.h:22:32: note: expanded from macro 'MALLOC_DECL'
0:03.94 MOZ_MEMORY_API return_type name##_impl(__VA_ARGS__);
0:03.94 ^
0:03.94 <scratch space>:178:1: note: expanded from here
0:03.94 realloc_impl
0:03.94 ^
0:03.94 /home/emilio/src/moz/gecko-2/obj-debug/dist/include/mozmemory_wrap.h:142:41: note: expanded from macro 'realloc_impl'
0:03.94 #define realloc_impl mozmem_malloc_impl(realloc)
Which I really didn't feel like digging into.
* I had to restore the order of TrustOverrideUtils.h and related files in nss
because the .inc files depend on TrustOverrideUtils.h being included earlier.
* I had to add a missing include to RollingNumber.h
* Also had to partially restore include order in JsepSessionImpl.cpp to avoid
some -WError issues due to some static inline functions being defined in a
header but not used in the rest of the compilation unit.
Differential Revision: https://phabricator.services.mozilla.com/D60327
--HG--
extra : moz-landing-system : lando
rg -l 'mozilla/Move.h' | xargs sed -i 's/#include "mozilla\/Move.h"/#include <utility>/g'
Further manual fixups and cleanups to the include order incoming.
Differential Revision: https://phabricator.services.mozilla.com/D60323
--HG--
extra : moz-landing-system : lando
Now mfbt/Move.h is empty except for that excellent comment about move
semantics... Should we put it somewhere else and delete the header as a
follow-up? Or just delete the header and carry on?
Differential Revision: https://phabricator.services.mozilla.com/D60297
--HG--
extra : moz-landing-system : lando
I was poking at the missing images bug by dropping cache sizes and such and
found that we always ended up repainting some SVGs over and over. This seems to
be the cause. When we fail to insert into the cache we still end up notifying
over and over (because we forget to say aWillCache=false), so we'd end up
repainting, and failing again, etc.
This is easy to see adding some logging to
nsDisplayBackgroundImage::PaintInternal.
Any ideas as to how to add a test for it?
Differential Revision: https://phabricator.services.mozilla.com/D60372
--HG--
extra : moz-landing-system : lando
Changes:
Tighten reftest pixel differences now that reftest has been migrated fully over to ubuntu1804.
Differential Revision: https://phabricator.services.mozilla.com/D59598
--HG--
extra : moz-landing-system : lando
* Makes it possible to selectively enable TRR for pbmode/container/window/etc
Differential Revision: https://phabricator.services.mozilla.com/D48363
--HG--
extra : moz-landing-system : lando
We should not attempt to shut down threads which we did not explicitly
remove from the image decode thread pool, when we want to shut down an
idle thread or otherwise. This behaviour would put us at risk of
shutting down a thread twice, which may trip assertions and/or cause
other issues. Now we assuming that any thread that is already removed
from the pool array has already been gracefully shutdown.
Differential Revision: https://phabricator.services.mozilla.com/D58827
--HG--
extra : moz-landing-system : lando
When one uses SourceSurfaceRawData to wrap a data pointer, it will
perform a copy of said data if GuaranteePersistence is called. This is
done for DrawTargetCapture, which is used with OMTP. Prior to this
patch, image surfaces would be wrapped by a SourceSurfaceRawData when
using the basic compositor on any non-Linux platform (since Linux does
not support volatile memory). This means every time imgFrame::Draw is
called with OMTP, a copy of the image will be made. Images don't need
this property since the data is already going to persist, so this patch
adds a new class SourceSurfaceMappedData, which takes a ScopedMap
keeping the underlying surface open and readable.
Differential Revision: https://phabricator.services.mozilla.com/D58199
--HG--
extra : moz-landing-system : lando
This changeset is a simple find and replace of `MOZ_FALLTHROUGH` and `[[fallthrough]]`.
Unfortunately, the MOZ_FALLTHROUGH_ASSERT macro (to assert on case fallthrough in debug builds) is still necessary after switching from [[clang::fallthrough]] to [[fallthrough]] because:
* MOZ_ASSERT(false) followed by [[fallthrough]] triggers a -Wunreachable-code warning in DEBUG builds
* but MOZ_ASSERT(false) without [[fallthrough]] triggers a -Wimplicit-fallthrough warning in NDEBUG builds.
Differential Revision: https://phabricator.services.mozilla.com/D56440
--HG--
extra : moz-landing-system : lando
Previously, GetCancelled() would have return true should the channel's status was an error.
Doing Cancel(NS_OK) for example, would have made a follow-up call to GetCancelled() return false. However, we can assert that such a call would have been a bug.
Following this change GetCancelled() will only return true if Cancel() was explicitly called.
Differential Revision: https://phabricator.services.mozilla.com/D55401
--HG--
extra : moz-landing-system : lando
There is no functional change with this commit. The default implementation for GetCanceled() is still to check if the status code is a failure.
However, it can be argued that as you had to call Cancel() on the nsIChannel, having to check the nsIHttpChannelInternal interface to determine if you had been canceled in the past was rather a non obvious path.
It makes more sense to check the nsIChannel interface to determine if it's been canceled already and this allows for finer granularity if needed in the future.
Differential Revision: https://phabricator.services.mozilla.com/D55268
--HG--
extra : moz-landing-system : lando
Previously, GetCancelled() would have return true should the channel's status was an error.
Doing Cancel(NS_OK) for example, would have made a follow-up call to GetCancelled() return false. However, we can assert that such a call would have been a bug.
Following this change GetCancelled() will only return true if Cancel() was explicitly called.
Differential Revision: https://phabricator.services.mozilla.com/D55401
--HG--
extra : moz-landing-system : lando
There is no functional change with this commit. The default implementation for GetCanceled() is still to check if the status code is a failure.
However, it can be argued that as you had to call Cancel() on the nsIChannel, having to check the nsIHttpChannelInternal interface to determine if you had been canceled in the past was rather a non obvious path.
It makes more sense to check the nsIChannel interface to determine if it's been canceled already and this allows for finer granularity if needed in the future.
Differential Revision: https://phabricator.services.mozilla.com/D55268
--HG--
extra : moz-landing-system : lando
Now the test runs following steps;
1) Start the test
2) Set iframe's src attribute to load the iframe document
3) Waits for a `loadComplete` message from the iframe
4) Waits 1 second to make sure `decodeComplete` doesn't happen in the iframe
Differential Revision: https://phabricator.services.mozilla.com/D57178
--HG--
extra : moz-landing-system : lando
This is generally pretty straightforward, and rewrites nearly all calls. It
skips the ones that it can detect using frame script globals like
`sendAsyncMessage`, though.
Differential Revision: https://phabricator.services.mozilla.com/D53740
--HG--
extra : moz-landing-system : lando
This is generally pretty straightforward, and rewrites nearly all calls. It
skips the ones that it can detect using frame script globals like
`sendAsyncMessage`, though.
Differential Revision: https://phabricator.services.mozilla.com/D53740
--HG--
extra : moz-landing-system : lando
This is generally pretty straightforward, and rewrites nearly all calls. It
skips the ones that it can detect using frame script globals like
`sendAsyncMessage`, though.
Differential Revision: https://phabricator.services.mozilla.com/D53740
--HG--
extra : moz-landing-system : lando