Two fixes here:
updateAppearance really depends on the tabbrowser.xml updateVisibility(), so
make that dependency explicit and remove the event handling from tabbrowser.
After the previous patch, TabsInTitlebar.init() runs much earlier, which means
that the initialization check that bails out from update() is not taken. Turns
out that this is definitely not an edge case, and that tons of initialization
code all over gBrowser.onDOMContentLoaded call updateAppearance(true), causing
tons of extra invalidation and reflows that aren't really needed.
Ensure we only do the work once.
MozReview-Commit-ID: 2E9b12aBast
Not doing can it cause perf issues, but old Linux distros aren't really happy
about it, and they barf when the chromemargin is removed dynamically, so prevent
that happening by default.
Right now it happens on the resize we do right before loading the window, but
that will stop working after bug 1439875.
MozReview-Commit-ID: 1Rhaz07fHaY
This gives the chance to code that relies on setting the XUL window attributes
to run before we actually size the window.
This should prevent the resizing on OSX and fix some other untested stuff that
the first commit probably broke...
MozReview-Commit-ID: DhCWgmCppek
This gives the chance to code that relies on setting the XUL window attributes
to run before we actually size the window.
This should prevent the resizing on OSX and fix some other untested stuff that
the first commit probably broke...
MozReview-Commit-ID: DhCWgmCppek
This gives the chance to code that relies on setting the XUL window attributes
to run before we actually size the window.
This should prevent the resizing on OSX and fix some other untested stuff that
the first commit probably broke...
MozReview-Commit-ID: DhCWgmCppek
This gives the chance to code that relies on setting the XUL window attributes
to run before we actually size the window.
This should prevent the resizing on OSX and fix some other untested stuff that
the first commit probably broke...
MozReview-Commit-ID: DhCWgmCppek
This removes the sync reflow from almost all cases. The only case where we keep it is when a keypress
caught in content triggers a sync message to the parent process. We should clean this up in bug 1371523.
I've tried to fix the tests, but a lot of them seem to be disabled anyway...
MozReview-Commit-ID: 9k36p7q8MKy
textContent is used to set the indicator label text here instead of the value attribute because the text set using the value attribute does not wrap when it exceeds the width of the panel, which in turn pushes the menulist and half of the indicator text out of view.
MozReview-Commit-ID: 1VBaQlbZwzQ
This resolves some confusion around the required callback for
`openInExtenalEditor` by converting it to instead return a Promise. This also
simplifies the flow of its callers as well.
MozReview-Commit-ID: EYoucELJLbu
We keep the XBL binding around for <content>, <constructor>, and <destructor>. This can
eventually be migrated to a Custom Element once we have platform support, but in the meantime
this is a way to get the many thousands of LOC into a JS class.
MozReview-Commit-ID: 1dCQp527yF9
* Code in XMLHttpRequestMainThread is converted to set the username and password individually. This is because when the parameters are empty, it ended up calling SetUserPass(":") which always returns an error.
MozReview-Commit-ID: 3cK5HeyzjFE
