Before this change, Servo used one code path that computed the position
of flows with `position: static` or `position: relative` and another
separate code path that computed the position of flows with `position:
absolute` or `position: fixed`. The latter code attempted to duplicate
the former code to determine the static position of hypothetical boxes,
but this was both fragile and incorrect in the case of hypothetical
boxes nested inside floats. In fact, it's impossible to determine the
static position of an absolute flow relative to its containing block at
inline-size assignment time, because that static position could depend
on a float that cannot be placed until block-size assignment!
This patch changes block layout to use the same code path for static
positioning of regular flows and static positioning of absolute flows
where applicable. This both simplifies the code and improves its
efficiency, since it allows the `hypothetical_position` field and
`static_block_offsets` data structure to be removed. Moreover, it
improves correctness in the above case (which the new reftest checks).
This allows the sidebar in Facebook Timeline to be positioned properly.
r? @glennw
<!-- Reviewable:start -->
[<img src="https://reviewable.io/review_button.png" height=40 alt="Review on Reviewable"/>](https://reviewable.io/reviews/servo/servo/5691)
<!-- Reviewable:end -->
Extracted out of #5649
* add more hyperlinks to associated specification for structs/methods
* follow redirects and update links
* replace broken links
* removal of WHATWG multipage page name since the page name is not
guaranteed to be stable
<!-- Reviewable:start -->
[<img src="https://reviewable.io/review_button.png" height=40 alt="Review on Reviewable"/>](https://reviewable.io/reviews/servo/servo/5693)
<!-- Reviewable:end -->
Extracted out of #5649
* add more hyperlinks to associated specification for structs/methods
* follow redirects and update links
* replace broken links
* removal of WHATWG multipage page name since the page name is not
guaranteed to be stable
flows and static positions of hypothetical boxes.
Before this change, Servo used one code path that computed the position
of flows with `position: static` or `position: relative` and another
separate code path that computed the position of flows with `position:
absolute` or `position: fixed`. The latter code attempted to duplicate
the former code to determine the static position of hypothetical boxes,
but this was both fragile and incorrect in the case of hypothetical
boxes nested inside floats. In fact, it's impossible to determine the
static position of an absolute flow relative to its containing block at
inline-size assignment time, because that static position could depend
on a float that cannot be placed until block-size assignment!
This patch changes block layout to use the same code path for static
positioning of regular flows and static positioning of absolute flows
where applicable. This both simplifies the code and improves its
efficiency, since it allows the `hypothetical_position` field and
`static_block_offsets` data structure to be removed. Moreover, it
improves correctness in the above case (which the new reftest checks).
This allows the sidebar in Facebook Timeline to be positioned properly.
This patch adds support for setting the line cap and join. However, it seems there's a problem on the azure-side, as the line cap setting doesn't work. Changing either the default values or using the new function has no effect. Line join works fine though.
<!-- Reviewable:start -->
[<img src="https://reviewable.io/review_button.png" height=40 alt="Review on Reviewable"/>](https://reviewable.io/reviews/servo/servo/5635)
<!-- Reviewable:end -->
All of the commits by brson have been reviewed. Just the android fixups and Rustup one (which also cleans up some of the duplication due to a complete rewrite of code that had been moved in the original PR).
<!-- Reviewable:start -->
[<img src="https://reviewable.io/review_button.png" height=40 alt="Review on Reviewable"/>](https://reviewable.io/reviews/servo/servo/5489)
<!-- Reviewable:end -->
The urllib version used a `FancyURLOpener` which use urllib. But urllib does not not handle proxies with SSL well.
For example, when adding the proxies to urllib.FancyURLOpener, I got this SSL error:
IOError: [Errno socket error] [Errno 1] _ssl.c:510: error:140770FC:SSL routines:SSL23_GET_SERVER_HELLO:unknown protocol
So I switched the function to urllib2.
A better solution would be to use `requests` but I prefer to stay with the "no-dependency" approach.
For example, this was my first solution with `requests`:
```python
def download(desc, src, dst):
print("Downloading(R) %s..." % desc)
with open(dst, 'wb') as handle:
response = requests.get(src, stream=True)
if not response.ok:
print("something went wrong downloading %s, :(" % desc)
sys.exit(1)
size = int(response.headers['content-length'])
block_size = 1024
recved = 0
for block in response.iter_content(1024):
recved += len(block)
pct = recved * 100.0 / size
print("\rDownloading %s: %5.1f%%" % (desc, pct), end="")
if not block:
break
handle.write(block)
print()
```
<!-- Reviewable:start -->
[<img src="https://reviewable.io/review_button.png" height=40 alt="Review on Reviewable"/>](https://reviewable.io/reviews/servo/servo/5678)
<!-- Reviewable:end -->
Fixes#5620.
Fix the TODOs and FIXMEs to comply with the spec.
Add test case for passing invalid invalid labels.
Update test metadata; three test cases have been resolved upstream and
will be fixed whenever the rust-encoding dependency is sufficiently upgraded.
The urllib version used a `FancyURLOpener` which use urllib. But urllib does not not handle proxies with SSL well.
A better solution would be to use `requests` but I prefer to stay with the "no-dependency" approach.
Extracted this out of #5649
This commit was created with the following commands:
```
find . -iname "*.webidl" -type f -print0 | xargs -0 sed -i '' 's/http:\(.*\)whatwg.org/https:\1whatwg.org/g'
```
```
find . -iname "*.rs" -type f -print0 | xargs -0 sed -i '' 's/http:\(.*\)whatwg.org/https:\1whatwg.org/g'
```
Extracted this out of #5649
This commit was created with the following commands:
```
find . -iname "*.webidl" -type f -print0 | xargs -0 sed -i '' 's/http:\(.*\)whatwg.org/https:\1whatwg.org/g'
```
```
find . -iname "*.rs" -type f -print0 | xargs -0 sed -i '' 's/http:\(.*\)whatwg.org/https:\1whatwg.org/g'
```
This addresses https://github.com/servo/servo/issues/3149. The immediate purpose is to support the constants in NodeFilter. The changes mostly follow the current Gecko Codegen.py. The main gist is that the generation of certain code artifacts is now gated by hasInterfaceObject() or hasInterfacePrototypeObject(), rather than by isCallback().
Available markers only:
Reflow
DOMEvent
Also need to implement:
Style marker
Paint marker
Javascript marker
frames reply, depends on getting javascript stack
I decided to make pull request before implemented another markers for getting feedback.
mb it would be better to create separated tasks.
Notices:
Marker doesn't fill stack and stackEnd
MemoryActor sends fake data because there is no memory profiler per tab
FramerateActor sends empty Vec, need implement http://mxr.mozilla.org/mozilla-central/source/dom/base/nsGlobalWindow.cpp#5240
