Follows-up I4c7d826c7ec654b, I1287f3979aba1bf1.
We lose useful coverage and spend valuable time keeping these accurate
through refactors (or worse, forget to do so). The theoretically "bad"
accidental coverage is almost never actually bad.
Having said that, I'm not removing them wholesale (yet). I've audited
each of these specific files to confirm it is a general test of the
specified subject class, and also kept it limited to those specified
classes. That's imho more than 100% of the benefit for less than 1%
of the cost (more because `@covers` is more valuable than the fragile
and corrosive individual private method tracking in tests that
inevitably get out of date with no local incentive to keep them up to
date).
Cases like structure tests keep `@coversNothing` etc and we still don't
count coverage of other classes. There may be a handful of large
legacy classes where some methods are effectively class-like in
complexity and that's why it's good for PHPUnit to offer the precision
instrument but that doesn't meant we have to use that by-default for
everything.
I think best practice is to write good narrow unit tests, that reflect
how the code should be used in practice. Not to write bad tests and
hide part of its coverage within the same class or even namespace.
Fortunately, that's generally what we do already it's just that we
also kept these annotations still in many cases.
This wastes time to keep methods in sync, time to realize (and fix)
when other people inevitably didn't keep them in sync, time to find
uncovered code only to realize it is already covered, time for a less
experienced engineer to feel obligate to and do write a low quality
test to cover the "missing" branch in an unrealistic way, time wasted
in on-boarding by using such "bad" tests as example for how to use
the code and then having to unlearn it months/years later, loss of
telemetry in knowing what code actually isn't propertly tested due to
being masked by a bad test, and lost oppertunities to find actually
ununused/unreachable code and to think about how to instead structure
the code such that maybe that code can be removed.
------
Especially cases like LBFactoryTest.php were getting out of hand,
and in GlobalIdGeneratorTest.php we even resorted to reminding people
with inline comments to keep tags in sync.
Change-Id: I69b5385868cc6b451e5f2ebec9539694968bf58c
Just methods where adding "static" to the declaration was enough, I
didn't do anything with providers that used $this.
Initially by search and replace. There were many mistakes which I
found mostly by running the PHPStorm inspection which searches for
$this usage in a static method. Later I used the PHPStorm "make static"
action which avoids the more obvious mistakes.
Bug: T332865
Change-Id: I47ed6692945607dfa5c139d42edbd934fa4f3a36
It misrepresents the users contribution to show empty text for a
revision when in fact the revision contained some text which we later
lost.
Also, errors from SqlBlobStore::fetchBlobs() did not stop a cache entry
from being written, so a subsequent cache hit would show the bad
revision as empty.
So, in Storage:
* Add BadBlobException, which is thrown by the Storage layer to
indicate that a revision is marked as bad.
* Have SqlBlobStore::getBlobStore() return an error for bad blobs
instead of an empty string.
* Duplicate the check for flags=error into SqlBlobStore::expandBlob().
This avoids an unnecessary cache fetch, and avoids making
decompressData() throw on error, which would be a b/c break.
* In SqlBlobStore::getBlob(), suppress the cache when there was an
error.
In Revision:
* Add BadRevisionException, to wrap BadBlobException in the Revision
layer.
* Return null from RevisionRecord::getContent() on a broader set of
errors. Make it mostly non-throwing.
* Add RevisionRecord::getContentOrThrow() which returns a non-nullable
Content.
* Note that SlotRecord::getContent() returns a non-nullable Content so
now throws in more cases.
In the UI:
* In Article::view(), catch the exception and show an error message.
* In DifferenceEngine, catch the exception and make a suitable error
message available via getRevisionLoadErrors(). In the diff page, show
the error message in a box.
* In ApiComparePages and the legacy rvdiffto, show a warning.
* In RawAction, show a 404 by analogy with other error cases.
* In EditPage, there was already handling for $content=null with an
appropriate error message (missing-revision-content). But having
$this->textbox1 = null caused PHP 8.1 deprecation warnings, so I fixed
that.
* In EditPage undo, there was already handling for null content, but I
improved the error message: "does not exist or was deleted" seems more
appropriate than "conflicting intermediate edits".
Change-Id: Idd1278d6d756ef37d64addb7b5f3be30747ea603
* Add a safe unserialize() wrapper for HistoryBlob classes
* Add a safe unserialize() wrapper for plain array data as used for
compressed internal storage by ConcatenatedGzipHistoryBlob and
DiffHistoryBlob.
* Fix tests broken by this.
* Fix unnecessary call to uncompress(), __wakeup() does this already.
Was a phan error now that we have more information about the type of
$obj.
* Add tests for successful unserialize and wakeup of WMF production
data.
Change-Id: Ic995dda16d9c6045b33f2fdae7f6575ac8329976
PHPUnit 3.7.0 was the first version to support @requires extension. At
the time checkPHPExtension() was added in a7901801b4, MediaWiki
still supported PHPUnit 3.6.7.
MediaWiki now requires intl and xml, so I removed checks for those
extensions rather than converting them to annotations.
checkPHPExtension() is removed without deprecation; it does not appear
to have ever been used (and is not likely to be used) in MW extensions.
This is explicitly permitted under the stable interface policy. Even if
it were not, only tests are affected, and they are supposed to fail
anyway if hard deprecated code is used.
Change-Id: I45f9b4c0e120683103cead916f4d4ef58bd11530
createMock() does the same, but is much easier to read.
A small difference is that some of the replacements made in this
patch didn't use disableOriginalConstructor() before. In case this
was relevant we should see the respective test fail. If not we can
save some CPU cycles and skip these constructors.
Change-Id: Ib98fb06e0fe753b7a53cb087a47e1159515a8ad5
Follows-up I361fde0de7f4406bce6ed075ed397effa5be3359.
Per T253461, not mass-changing source code, but the use of the native
error silencing operator (@) is especially useful in tests because:
1. It requires any/all statements to be explicitly marked. The
suppressWarnings/restoreWarnings sections encourage developers to
be "lazy" and thus encapsulate more than needed if there are multiple
ones near each other, which would ignore potentially important
warnings in a test case, which is generally exactly the time when
it is really useful to get warnings etc.
2. It avoids leaking state, for example in LBFactoryTest the
assertFalse call would throw a PHPUnit assertion error (not meant
to be caught by the local catch), and thus won't reach
AtEase::restoreWarnings. This then causes later code to end up
in a mismatching state and creates a confusing error_reporting
state.
See .phpcs.xml, where the at operator is allowed for all test code.
Change-Id: I68d1725d685e0a7586468bc9de6dc29ceea31b8a
It seems some tests are fixed in the meantime without removing the
skipped part for this.
Bug: T195807
Follow-Up: I58261dd70eea3581803987a4a7739c7d55558f42
Change-Id: Ie883b7056cdd0cb6bec81c2f7e4ed54177b4ecb2
The removal of the Revision class led to this test failing, since
tt:1000 started to have content - not sure why but I assume something
in the Revision tests was clearing the store - the text
was "UTContent" which is used in many tests via helper
method MediaWikiIntegrationTestCase::getExistingTestPage()
which can create pages with that text.
Change-Id: I04c669ad8c8a7166138eabc973140205b60536a5
My personal best practice is to not document @params when there
is a @dataProvider. I mean, these test…() functions are not
meant to be called from anywhere. They do not really need
documentation. @param tags don't do much but duplicate what the
@dataProvider does. This is error-prone, as demonstrated by the
examples in this patch.
This patch also removes @throws tags from tests. A test…() can
never throw an exception. Otherwise the test would fail.
Most of these are found by the not yet released I10559d8.
Change-Id: I3782bca43f875687cd2be972144a7ab6b298454e
Advertize the fact that production errors resulting from corruption in
the content storage layer can be remedied using findBadBlobs.php.
Change-Id: I4eab773d0d2da57a545f211c2fef36146331c570
The name change happened some time ago, and I think its
about time to start using the name name!
(Done with a find and replace)
My personal motivation for doing this is that I have started
trying out vscode as an IDE for mediawiki development, and
right now it doesn't appear to handle php aliases very well
or at all.
Change-Id: I412235d91ae26e4c1c6a62e0dbb7e7cf3c5ed4a6
This should be the exact same. Its more a style change than anything.
So why do it then?
* I believe this is much less confusing than code mentioning a weird
"standard class". Barely anybody knows what this is, and what the
difference between "object" and "stdClass" is.
* The code is shorter.
* It's even faster. In my micro benchmark it's twice as fast.
Change-Id: I7ee0e8ae6d9264a89b6cd1dd861f0466ae620ccc
This adds support for addresses starting with the prefix "bad:",
which will always resolve to an empty string.
The syntax for addresses supported by SqlBlobStore is relaxed
to be closer to RFC 3986 and make it easier to include meaningful
information after the "bad:" prefix.
Bug: T240686
Change-Id: I175b1880ce94ff8a9079bc50440cde188e57d95d
This is a TEMPORARY HACK to fix a problem with cached values
getting mixed up when using getBlobBatch(). As of the creation of this
patch, the actual cause is unknown. Caching should be put back once the
cause of the problem has been found.
Note that getBlobBatch is only used in maintenance scripts.
Removal of caching in that method has no effect on web requests.
getBlobBatch() was introduced in I56306c50a661 and is used by
RevisionStore::getContentBlobsForBatch(), which in turn is needed
by the Translate extension to remove its dependency on the deprecated
rev_text_id field.
Bug: T235188
Change-Id: I94c6f9ba7b9caeebaa9b055916f15f7bbdcd8fb6
Something is going wrong with the way we use WANObjectCache in
getBlobBatch. This patch introduces a test case that should help narrow
down the cause.
Bug: T235188
Change-Id: I2d06c09609874440992e6c90d3ebfa15400ac302
Constructing a Language object in order to initialize the
BlobStoreFactory service causes a circular dependency
(see T231866).
SqlBlobStore was using the Language object to all iconv.
But nothing language specific is done in Language::iconv,
so we can just inline the call.
Bug: T231866
Change-Id: I90c25decbcff10ea762a2c7474a12fd2041b3abc
* Inject settings and global instances as dependencies to the
ExternalStoreMedium instances. This includes the local wiki
domain, so that wfWikiId() calls are not scattered around.
* Create ExternalStoreAccess service for read/write logic.
* Deprecate the ExternalStore wrapper methods.
* Add some exception cases for bogus store URLs are used instead
of just giving PHP warnings and failing later.
* Make moveToExternal.php require the type/protocol to decide
which ExternalStoreMedium to use instead of assuming "DB".
* Convert logging calls to use LoggerInterface.
Change-Id: I40c3b5534fc8a31116c4c5eb64ee6e4903a6197a
So we can make the job voting, preventing other tests from regressing.
These tests can be re-enabled whenever they're made to pass.
Bug: T195807
Change-Id: I58261dd70eea3581803987a4a7739c7d55558f42
This introduces a convenience method for constructing a blob address
from a text ID. It's the inverse of SqlBlobStore::getTextIdFromAddress
Change-Id: I31b3ee5e40185c754fb2c119eb5edc50b903f5dc
