Like an earlier patch in the series, this uses `Split()` to return dependent
substrings into the blocklist string, rather than copying each individual
entry out for parsing. We also, in passing, eliminate a copy from the
converted UTF8 string into an entirely separate string.
Depends on D89169
Differential Revision: https://phabricator.services.mozilla.com/D89170
Instead of splitting the entire entry into an `nsTArray`, we can use
`nsTSubstring::Split` to provide dependent strings into the entry. We copy
the actual keys and values later, so this should be perfectly safe.
Depends on D89168
Differential Revision: https://phabricator.services.mozilla.com/D89169
There's a lot of string splitting here, which needlessly copies, and then
there's a lot of copies out of the split arrays. This patch at least
addresses the latter by using references where possible.
Differential Revision: https://phabricator.services.mozilla.com/D89168
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
Summary of major changes:
* Bookmarks, history, and tabs restriction chars now enter search mode. I added
a method to UrlbarProviderHeuristicFallback to return a result with a keyword
when one of these is used.
* This fixes other bugs like recognizing aliases that are entered at the
beginning of non-empty search strings, and not quasi-re-entering search mode
when search mode is already entered and you type another alias.
* The heuristic now determines whether we enter search mode, similar to how it
also determines whether we autofill. When the heuristic has a keyword but no
keyword offer, and the keyword is one of the recognized search mode keywords,
then we enter search mode, cancel the current query, and start a new query
with the remainder of the search string after the keyword.
* I slightly changed how we detect an alias, but only when update2 is
enabled. Now, an alias must be followed by a space; otherwise, the alias is
not recognized and instead just remains part of the seach string. Because if
we don't do that, then you end up in a strange situation after typing an alias
but before pressing space: The heuristic says "Search with <engine with the
alias>", but we haven't entered search mode yet because you haven't typed a
space yet. This is true for both @aliaes and non-@aliases.
* A consequence of the previous point is that we can still autofill @aliases
with a trailing space, which IMO is important. Then, once the user types any
char (space or not), we immediately enter search mode with the query being
whatever char they typed. This is less important after bug 1658605 landed, but
it's still good to have.
* Previously, `UrlbarView.onQueryResults` called UrlbarInput in order to
autofill after the first result is received. This is circuitous becaue the
input already has an `onFirstResult` method, which I now use to enter search
mode when appropriate. So I moved the autofill call from UrlbarView to
`UrlbarInput.onFirstResult`.
* As I mentioned, I improved some test framework and simplified some related
product (non-test) code. For example:
* I removed `UrlbarUtils.KEYWORD_OFFER.NONE` in favor of just leaving
`keywordOffer` as `undefined`.
* `tailOffsetIndex` can now be `undefined` if it's not relevant.
* I removed empty-string `icon` properties from payloads in favor of
`undefined`.
* In tests, I ignore `undefined` but present properties in payloads so they
don't count when comparing payloads with `deepEqual`.
* We weren't previously comparing `result.source` and `result.type` in
xpcshell tests, and that's important IMO, so I added checks for those and
updated tests.
* `isSearchHistory` is redundant, so I removed it. For form history, we
should be checking `result.source == HISTORY` and `result.type == SEARCH`.
* A bunch of tests needed to be updated for this new behavior.
Differential Revision: https://phabricator.services.mozilla.com/D87944
This stores slots/element ranges as object and start index rather than having
two separate representations (one used during marking, one when we yield to the
mutator during marking). It's possible that this could affect marking
performance but we'll catch that by checking the mark rate telemetry. This
should also reduce the size needed for the mark stack a little.
Differential Revision: https://phabricator.services.mozilla.com/D88723
* Build dictionary of paper sizes at init and whenever the printer changes
* Fetch async margin properties on the selected paper
* Find a paper match by size to selected the correct paper when opening a print dialog or switching to a different printer
* Add virtual properties for paperSize and currentPaper
* Refresh margin settings when the paperName changes as these may be computed
Differential Revision: https://phabricator.services.mozilla.com/D87943
It seems that the clang toolchain has trouble optimizing away the tag access in
a `mozilla::Variant`. Use of a `CompactPair` seems to be better when `E` is
small (all cases where `UnusedZero<E>` is true are small right now anyway).
While this doesn't change `sizeof(Result<V, E>)` in most cases, it makes
the generated code simpler/smaller: As of now, this reduce the overall VM size
of libxul.so by ca. 59K.
Without additional effort, this requires and leads to default-construction of
a `V` even in cases where it is never accessed, so this is restricted to
trivially default-constructible `V` for now.
This could be avoided by replacing `CompactPair<V, E>` by
`CompactPair<AlignedStorage2<V>, E>`.
Differential Revision: https://phabricator.services.mozilla.com/D88393
We want to try to align 'push-interval' tasks to the 'backstop'. This way
we have greater confidence in our backstop pushes, and it will allow us to
simplify a lot of our backstop logic.
Depends on D89055
Differential Revision: https://phabricator.services.mozilla.com/D89056
Each file represents a new taskgraph generation in the tests. But now that these exist, we
can add new assertions to the existing files without worry.
Depends on D89054
Differential Revision: https://phabricator.services.mozilla.com/D89055
Under this model, each test file should correspond to a static set of parameters. Each
set of parameters will only ever generate the taskgraph a single time (so adding a new
file will have a perf hit, but adding new test functions to an existing file will not).
Differential Revision: https://phabricator.services.mozilla.com/D89054
Android doesn't use sessionstore-windows-restored and currently doesn't use the search service (which fails for it at the moment).
Differential Revision: https://phabricator.services.mozilla.com/D89062
Scott
Mihai Alexandru Michis
Nathan Froyd
Micah Tigley
Andrei Oprea
Mirko Brodesser
Drew Willcoxon
Nan Jiang
Jeff Gilbert
Dan Mosedale
Sam Foster
Valentin Gosu
Razvan Maries
Jon Coppeard
Simon Giesecke
Emilio Cobos Álvarez
Jonathan Watt
Bob Owen
Henri Sivonen
Andrew Halberstadt
Agi Sferro
Mark Banner