Instead of hashing each file separately, hash them in a batch.
Previous research on this area of code has identified the suppressing
and restoring of warnings to have a measurable cost, which is why
this was optimised in 3621ad0f82. However, we never really made
use if it (aisde from the 2:1 change in that commit itself), because
we always call it with a single item, turn it into an array, do the
hash and then merge it again.
Instead, we now let it handle a single module's set of files all
at once. Given that this no longer exposes an array of hashes,
also update the (private) signature of getFileHashes to reflect this.
This means all file modules will have their version bumped during the
next MediaWiki release. In general this happens for most releases
and weekly branches already (due to localisation update, general
maintenance, linting changes, and other internal changes). But,
noting here for future reference as this might not be obvious from
the diff.
Change-Id: I4e141a0f5a5c1c1972e2ba33d4b7be6e64ed6ab6
All the magic functions needs public visibility to be callable from php
internals like garbage collector
Change-Id: I1baf04bf8ff787da880d46e4a6daa77f5a6de73f
Repeating the variable name doesn't do anything. Documentation
generators don't need it. It's more stuff to read that doesn't add new
information. And it can become outdated.
Note there are two types of @var docs. When used inline (and not on a
class property) the variable name is needed.
Change-Id: If5a520405efacd8cefd90b878c999b842b91ac61
And also update approximated counts, which for the most part are lower
than reported (hooray!)
Bug: T231636
Depends-On: Ica50297ec7c71a81ba2204f9763499da925067bd
Change-Id: I78354bf5f0c831108c8f606e50c87cf6bc00d8bd
All of these suppression prevent the detection of many common mistakes,
and could easily prevent things like T231488. Especially if there are
few issues of a given type, it's way better to suppress them inline,
instead of disabling them for the whole core.
This patch only touches the one with a lower count (although those
counts may be out of date).
Bug: T231636
Change-Id: Ica50297ec7c71a81ba2204f9763499da925067bd
This makes AutoLoaderStructureTest in PHPUnit and the
generateLocalAutoload.php maintenance script much faster.
On my machine, it made it 35X faster (or time spent reduced by 97%).
Bug: T225730
Change-Id: Ife959bd17ce9c2ae952dfbd158ddb3d8475e8cb2
* Title: phan false positive
* McrUndoAction: fixed improper use of @param
* UploadSourceAdapter: fixed wrong type
* XmlTypeCheck: Use null so phan doesn't think we're trying to call the
function ''
* Database: phan false positive
* SpecialBlock: Use phan's advanced type documentation so phan knows
specifically what's being returned
* ChangesListSpecialPage: phan false positive
* BatchRowUpdate: Have default callback take a parameter so phan doesn't
think too many arguments are being passed
* MimeAnalyzer: left FIXME for relying on PHP 7.1 unpack() signature
* LanguageConverter: Specify types for $mTables since phan couldn't
determine it automatically
* preprocessorFuzzTest: Implement User::load() method signature
Change-Id: I08080ab636c5fe67ea6a4e14b2212d7523606e21
This code is functionally identical, but less error prone (not so easy
to forget or mix these numerical indexes).
This patch happens to touch the Parser, which might be a bit scary. We
can remove this file from this patch if you prefer.
Change-Id: I8cbe3a9a6725d1c42b86e67678c1af15fbc5961a
* Introduce MSCompoundFileReader, which reads the CFB directory and
detects the file type from well-known names in the root directory
* Do not detect a ZIP file if the EOCDR is not at the end. Other
containers, especially CFB files, may contain ZIP files embedded
within them in the last 64KB, but this is not a security concern
unless the EOCDR is exactly at the end of the file.
Bug: T40432
Change-Id: Id5b1a258ccf3c3c8951e32f6b7a5b1bafe941082
In a nutshell:
This commmit removes the use of drifting microtime()
in favour of non-drifting time(). Then, we increase the size of
the counter by x1000, and consider every 1000 increments as 1ms
for the purposes of UUID.
Why:
This means we eliminate the whole code that can produce a fatal
exception when the clock drifts by more than we can wait for,
which puts us in a logical conundrum we can't get out of, hence
it previously fatalled.
Not aborting random end-user requests and jobs is good.
This also means we avoid the vast majority of cases where
MediaWiki would busy-loop for upto 10ms in a likely-to-fail
attempt to sync the clock. This means the method runs faster,
which is a nice win, albeit not the main objective.
Bug: T94522
Change-Id: Ia8a847078ec76d633854db6823a20f0961c80f80
Various miscellaneous clean ups with no change to any of the
logical code. This to make the next commit have a cleaner diff
for easier review.
* Make internal millitime() non-static.
* Improve documentation and add missing @covers annotations.
* Correct getTimestampedID88() documentation to state that
only two values need to be passed, not three.
This is significant because the Flow extension is actually
using this method in a subclass and passes only two values.
Bug: T94522
Change-Id: I2a0c51bea58df4cc0c253c1c10de3ac383f04c8e
Per discussion in T166010, we're going to handle class aliases (e.g. for
BC) by including the class_alias() call in the same file as the target
class. When the target class is a PSR4-compliant class, we still need to
pick up that alias for inclusion in autoload.php.
Thus, instead of excluding whole directories, we need to process the
files and filter out only those found classes that are PSR4 compliant.
Bug: T204983
Change-Id: I1c516998df368531c90ea54acc5be8be96e1db6c
Replace it all with random_bytes(), leave
only MWCryptRand::generateHex() as a convenience helper.
Change-Id: Ic30376a90e66d8f00dab86e7e6466fb3a750b87d
AutoloaderTest covers the AutoLoader class, and AutoLoaderStructureTest
covers the structure part of the test.
Change-Id: Ic4e7bfd670e1c3413631bda31260cc1cc825b1a2
PHP 7 introduces anonymous classes with a syntax like
$instance = new class() extends Foo { ... };
ClassCollector is incorrectly detecting this as a class named
"() extends Foo". This patch fixes that by having it ignore "new class"
in much the same way it currently ignores "::class".
Change-Id: I4d1985a9c04be71f7bea6cb7b61dcea74f44a6e2
Because it doesn't work. `phpcs:ignoreFile` is intended as a performance
shortcut, and gives up on the file as soon as that token is found.
Instead, use `phpcs:disable` which does support selectively disabling
some sniffs. And since disabling is local to the file, there's no need
to re-enable it at the bottom of the file.
<https://github.com/squizlabs/PHP_CodeSniffer/issues/1903> has some
relevant discussion and clarification from upstream on this.
And make the files that are now being re-enabled pass PHPCS.
Change-Id: Ia4d8c45045f61cf9f24eb1d8631ff98d99c82d69
It was never super popular anyway, APC was always the best option.
The project has no plans to move to PHP7, so it's quickly reaching
its end of life. Oh, and Fedora dropped it from their repos 2 years
ago.
Change-Id: Ia3257e86a6323d8943f04a5c05c72c0bd4c4b0a9
phpdbg is a gdb-style debugger for PHP that is run from the command
line. However, it has a different PHP_SAPI value, so it was impossible
to run maintenance scripts with it (until now).
To avoid having to check both PHP_SAPI values in a bunch of places,
introduce wfIsCLI() to easily check whether running from the
command-line or not.
We're (CI team) interested in generating code coverage with phpdbg
instead of xdebug, hence this patch.
Bug: T184043
Change-Id: Id1f994ca146d7858cd8bb6ab6cdbb7718ff524fb
