The main issue was resources_dir_path. Every time it was called it would start from the executable's path and walk up the hierarchy to find a directory named "resources". The sandbox was granted permission to read from the found resources dir, but after the sandbox had been activated resources_dir_path would again start from the executable's path and try to find the resources dir. It would then fail with "Operation not permitted" when trying to canonicalize the path because it didn't have permissions to read metadata under ./target.
To fix this the resources dir path is now cached between resources_dir_path calls.
Correct x and y input for origin_rect
Fix error in construction of the `origin_rect` for `UnioningFragmentBorderBoxIterator`.
r? @asajeffrey
<!-- Reviewable:start -->
---
This change is [<img src="https://reviewable.io/review_button.svg" height="35" align="absmiddle" alt="Reviewable"/>](https://reviewable.io/reviews/servo/servo/10444)
<!-- Reviewable:end -->
This adds in preferences for all the SM 39 available options (as
retrieved from Gecko), and uses the same defaults as Gecko. A few
properties are not supported yet, and incremental GC is still always
disabled regardless of the preference setting.
This also adds back in the options that were accidentally deleted when
\#10342 was rebased, which moved things from script_thread.rs to
script_runtime.rs.
Add some machinery to allow us to disable CSS properties/values per-product
Here's an initial stab at solving the issues @SimonSapin brought up in #10408.
<!-- Reviewable:start -->
---
This change is [<img src="https://reviewable.io/review_button.svg" height="35" align="absmiddle" alt="Reviewable"/>](https://reviewable.io/reviews/servo/servo/10461)
<!-- Reviewable:end -->
Fixes for insertion point caret rendering
This contains several fixes for the code to position and render the insertion point. The main effect is that the insertion point is now rendered correctly when in an empty input field. See the individual commit messages for more detais. r? @pcwalton
<!-- Reviewable:start -->
---
This change is [<img src="https://reviewable.io/review_button.svg" height="35" align="absmiddle" alt="Reviewable"/>](https://reviewable.io/reviews/servo/servo/10361)
<!-- Reviewable:end -->
Before this fix, if a TextRun containing an insertion point was split into
multiple ScannedText fragments, text layout would draw an insertion point
inside of each of the fragments. This patch records the insertion point
position at most once per TextRun, and copies it only into the appropriate
ScannedText fragment.
This PR mistakenly removed JIT enabling/disabling by preference:
https://github.com/servo/servo/pull/10342/files
Look for `get_pref`.
I'm putting the missing piece of code into the appropriate place in
script_runtime.rs
layout: Allow non-absolutely-positioned elements with `overflow: scroll` set to be scrolled.
This makes them establish stacking contexts, which is a CSS 2.1 spec
violation. However, we were already violating the spec here for
absolutely-positioned elements with `overflow: scroll`. It will probably
be easier to fix this spec violation once we either switch entirely to
WebRender or we have multiple layers per stacking context.
r? @mbrubeck
<!-- Reviewable:start -->
---
This change is [<img src="https://reviewable.io/review_button.svg" height="35" align="absmiddle" alt="Reviewable"/>](https://reviewable.io/reviews/servo/servo/10450)
<!-- Reviewable:end -->
Added --soft-fail option.
At the moment, wptrunner always runs the WPT tests with `servo --hard-fail`. To test hardening, we need to switch off `--hard-fail`. This PR introduces a `--soft-fail` option to do that.
<!-- Reviewable:start -->
---
This change is [<img src="https://reviewable.io/review_button.svg" height="35" align="absmiddle" alt="Reviewable"/>](https://reviewable.io/reviews/servo/servo/10420)
<!-- Reviewable:end -->
set to be scrolled.
This makes them establish stacking contexts, which is a CSS 2.1 spec
violation. However, we were already violating the spec here for
absolutely-positioned elements with `overflow: scroll`. It will probably
be easier to fix this spec violation once we either switch entirely to
WebRender or we have multiple layers per stacking context.
Closes#2742.
This fixes https://github.com/servo/servo/issues/10260 by pulling in
https://github.com/servo/ipc-channel/pull/61 (fix receive for messages
close to packet size) and https://github.com/servo/ipc-channel/pull/62
(properly handle ENOBUFS); where the latter is not critical per se, as
there was a workaround already -- but that workaround aggrevated the
first bug, resulting in the urgent issue...
This bump requires a tidy override for `uuid`: `ipc-channel` was updated
to `uuid 0.2` in https://github.com/servo/ipc-channel/pull/63 (don't
know why...), while other crates are still with `0.1`. That was blocking
this urgent bug fix; and according to a discussion with @mbrubeck on
IRC, the override should be OK in this case.
