Much like BindToTree.
This will be useful because I need to pass more information through
UnbindFromTree() to speed up dir=auto for bug 1874040.
Differential Revision: https://phabricator.services.mozilla.com/D202215
Includes only support for dynamically loading style sheets from `<link>`
elements and `Link` header fields.
The remaining support, including preloading, is completed in part 3.
Differential Revision: https://phabricator.services.mozilla.com/D191745
Right now, when binding to the tree we queue a runnable to update the
stylesheet, even though mEnableUpdates is false.
Even though the redundant update is nowadays always cached, it's just
wasted work, and the code is simpler without it.
This will be tested by bug 1771113, which is what made me look at this.
We need to tweak a bit the dispatch of applicable state change events
for DevTools, because for a case like:
div.attachShadow({ mode: "open" }).innerHTML = `<style>...</style>`;
Before we'd go through the stylesheet cache here due to the redundant
update:
https://searchfox.org/mozilla-central/rev/fb43eb3bdf5b51000bc7dfe3474cbe56ca2ab63c/layout/style/SharedStyleSheetCache.cpp#161-165
But now we won't, and the code in StyleSheet.cpp wasn't quite correct /
didn't dispatch the event.
Nobody listens to style-sheet-applicable-state-changed, so remove that
code while at it.
Differential Revision: https://phabricator.services.mozilla.com/D185559
Same rg + sed shenanigans as the first patch.
There were two that could fail, both due to OOM:
* HTMLInputElement::AfterSetAttr: If we fail (only in the type=range
case) we end up with an old value without it being clamped by
min/max/step.
* HTMLBodyElement::AfterSetAttr: If we fail we won't peek up the
DocShell's frame margins and styling could be incorrect.
That seems better than having to deal with broken states after we've
already set the attribute.
Depends on D176069
Differential Revision: https://phabricator.services.mozilla.com/D176070
I don't think there's ever a way these should differ.
Got some include hell from removing ReferrerInfo.h from Document.h but
hopefully should be straight-forward to review.
Depends on D173154
Differential Revision: https://phabricator.services.mozilla.com/D173155
Which is the spec term. nsIStyleSheetLinkingElement is even more
confusing since it may not be an element at all (see: processing
instructions).
Differential Revision: https://phabricator.services.mozilla.com/D76071
The cast in InitWithNode is wrong. AsElement() asserts instead of
checking the flag, so we always pass an element (and if we didn't we'd
have type confusion problems). I audited the callers and we're fine.
Anyhow, always require an element, and add two convenience constructors
for C++ code.
Differential Revision: https://phabricator.services.mozilla.com/D73636
The cast in InitWithNode is wrong. AsElement() asserts instead of
checking the flag, so we always pass an element (and if we didn't we'd
have type confusion problems). I audited the callers and we're fine.
Anyhow, always require an element, and add two convenience constructors
for C++ code.
Differential Revision: https://phabricator.services.mozilla.com/D73636
BindContext was going to have way more information at first, but then I realized
that most of the things I wanted to know were basically a flag away using the
parent node.
Still I think it's worth it, now experimenting with BindToTree will only mean
adding a field to a struct that's included from a couple cpp files, instead of a
massive pain.
I also think this is clearer, and doing this highlights quite a few
inconsistencies in our code which I've left untouched, but commented with
FIXMEs.
Steps are:
$ for file in $(rg 'nsresult BindToTree\(' | cut -d : -f 1 | sort | uniq); do sed -i 's#nsresult BindToTree(Document\* aDocument, nsIContent\* aParent,#nsresult BindToTree(BindContext\&, nsINode\& aParent)#g' $file; done
$ for file in $(rg 'nsresult BindToTree\(' | cut -d : -f 1 | sort | uniq); do sed -i 's# nsIContent\* aBindingParent) override#override#g' $file; done
$ for file in $(rg '::BindToTree\(' | cut -d : -f 1 | sort | uniq); do sed -i 's#::BindToTree(Document\* aDocument, nsIContent\* aParent,#::BindToTree(BindContext\& aContext, nsINode\& aParent)#g' $file; done
$ for file in $(rg '::BindToTree\(' | cut -d : -f 1 | sort | uniq); do sed -i 's#nsIContent\* aBindingParent)##g' $file; done
$ for file in $(rg '::BindToTree\(' | cut -d : -f 1 | sort | uniq); do sed -i 's#::BindToTree(aDocument, aParent, aBindingParent)#::BindToTree(aContext, aParent)#g' $file; done
$ ./mach clang-format
Then manual fixups.
Depends on D32948
Differential Revision: https://phabricator.services.mozilla.com/D32949
$ for file in $(rg UnbindFromTree | cut -d : -f 1 | sort | uniq); do sed -i 's#UnbindFromTree(bool aDeep = true,#UnbindFromTree(#g' $file; done
$ for file in $(rg UnbindFromTree | cut -d : -f 1 | sort | uniq); do sed -i 's#UnbindFromTree(bool aDeep,#UnbindFromTree(#g' $file; done
$ for file in $(rg UnbindFromTree | cut -d : -f 1 | sort | uniq); do sed -i 's#UnbindFromTree(aDeep,#UnbindFromTree(#g' $file; done
$ ./mach clang-format
And fix the two callers and little use of the aDeep argument (see the "Manual
changes" patch attached to bug).
Differential Revision: https://phabricator.services.mozilla.com/D32898
...instead of forwarding to the sheet like HTMLStyleElement does.
I've proposed this behavior in:
https://github.com/whatwg/html/issues/3840#issuecomment-480894419
I think this is one of the sane behaviors we can have, what Blink / WebKit do
makes no sense to me.
Alternative potentially-sane behavior is making the initial value of the
stylesheet's disabled bit the same as the content attribute, and both reflect
and forward the attribute from the setter.
That means that setAttribute does something different than setting `disabled`,
which means that you can get into all sorts of funny states when reloading the
sheet... So I rather not do that.
Differential Revision: https://phabricator.services.mozilla.com/D26573
...instead of forwarding to the sheet like HTMLStyleElement does.
I've proposed this behavior in:
https://github.com/whatwg/html/issues/3840#issuecomment-480894419
I think this is one of the sane behaviors we can have, what Blink / WebKit do
makes no sense to me.
Alternative potentially-sane behavior is making the initial value of the
stylesheet's disabled bit the same as the content attribute, and both reflect
and forward the attribute from the setter.
That means that setAttribute does something different than setting `disabled`,
which means that you can get into all sorts of funny states when reloading the
sheet... So I rather not do that.
Differential Revision: https://phabricator.services.mozilla.com/D26573
Summary: Really sorry for the size of the patch. It's mostly automatic
s/nsIDocument/Document/ but I had to fix up in a bunch of places manually to
add the right namespacing and such.
Overall it's not a very interesting patch I think.
nsDocument.cpp turns into Document.cpp, nsIDocument.h into Document.h and
nsIDocumentInlines.h into DocumentInlines.h.
I also changed a bunch of nsCOMPtr usage to RefPtr, but not all of it.
While fixing up some of the bits I also removed some unneeded OwnerDoc() null
checks and such, but I didn't do anything riskier than that.
Various places in dom/ use the pattern:
already_AddRefed<NodeInfo> ni = ...;
which is supposed to be disallowed by our static analysis code, but
isn't, for whatever reason. To fix our static analysis code, we need to
eliminate instances of the above pattern.
Unfortunately, eliminating this pattern requires restructuring how Nodes
are created. Most Node subclasses take `already_AddRefed<NodeInfo>&` in
their constructors, and a few accept `already_AddRefed<NodeInfo>&&`. We
need to enforce the latter pattern consistently, which requires changing
dozens of source files.
Mostly automatic via sed. Only parts which I touched manually (apart from a
couple ones where I fixed indentation or which had mispelled arguments) are the
callers. I may have removed a couple redundant `virtual` keywords as well when
I started to do it manually, I can revert those if wanted.
Most of them are just removing the argument, but in Element.cpp I also added an
assertion for GetBindingParent when binding the ShadowRoot's kids (the binding
parent is set from the ShadowRoot constructor, and I don't think we bind a
shadow tree during unlink or what not which could cause a behavior difference).
Differential Revision: https://phabricator.services.mozilla.com/D2574
MozReview-Commit-ID: 2oIgatty2HU
I've kept the nsAutoStrings in the StyleSheetInfo class on the hopes that the
compiler does RVO, but if it doesn't I can remove I guess.
MozReview-Commit-ID: 2vN6BSEhYcw
Per spec, assinging an empty string to textContent or innerHTML should not
trigger any mutation observers, and therefore should not trigger a stylesheet
reparse, when the element is already empty.
Since we need to apply some special handling to innerHTML and textContent
assignments to <style> nodes, and therefore re-parse directly in the setter
rather than in response to mutation observers, we need to special-case this
scenario.
MozReview-Commit-ID: KdOYFs8ayT7
I moved the IgnoreErrors decl so it would come after the OOMReporter decl and I
could add the new conversion operator.
MozReview-Commit-ID: B1S6DXmZfvE
Summary: It uses two node bits that can be better suited for something else.
Reviewers: xidorn, smaug
Bug #: 1444905
Differential Revision: https://phabricator.services.mozilla.com/D709
MozReview-Commit-ID: HIPDtHm6xpM
This change captures the subject principal when a scripted caller sets the
textContent or innerHTML property of a <style> node, and uses it as the
triggering principal for the resulting stylesheet.
If the node contents are modified in any way other than through textContent or
innerHTML, the triggering principal is forgotten (which is an intentional
design feature).
MozReview-Commit-ID: GacZFIB5BzS
Cathie Chen