The next revision in the page history isn't necessarily the previous
revision (due to selective undeletions, history merges, etc). This
passes the next revision to HistoryRevisionTools so extensions can check
if needed. Also, it passes the user to this hook and DiffRevisionTools
to avoid use of wgUser or having to retrieve context.
Change-Id: Ibc68f19040eebe3614e07f753f26bbfd376ae28d
These callers don't need to do purges, but can still perfectly
take advantage of this instance over a plain BagOStuff. Namely:
* Replication and snapshot lag awareness
* Preemptive regeneration
* Easy process cache support
The idea is for there to only be one caching class/factory
to use, instead of having rules for picking which one to use.
Change-Id: I8e362df451c0c28731fc853c044c4c4b8e097f01
@codingStandardsIgnoreFile is for whole files, use
@codingStandardsIgnoreStart instead
Also correct a MediaWiki.WhiteSpace.SpaceBeforeSingleLineComment.
SingleSpaceBeforeSingleLineComment sniff in the same file
Follows I06cdab4616b5bff47c85152df28f18c861730a23
Change-Id: I75e6e936e2fb4453bd56848ff39cabd92ae171fc
Graphite expects name components to be dot-separated, so our habit of using
dashes doesn't really make sense. Change metric names to be more compatible
with Graphite, except the job queue's, since that will require a gdash
dashboard definition migration.
Change-Id: I77d0ff7606a8fc88434e4352d23415a9a8f4725a
GNU 'diff -u' prefixes each line with either a space, a '+', or a '-'.
UnifiedDiffFormatter does the same, but it also adds an additional column of
whitespace between the prefix and the line. GNU diff only does that in
non-unified mode.
Fix this by implementing lines() in UnifiedDiffFormatter, overriding the parent
class implementation.
Bug: T100069
Change-Id: I1bf1b8e6d1d5aceb2c3836548f492f7edebe5a12
* This is a more specific form of the logic removed in 3c2bc32ae1.
It does not suffer the problem of causing constant master DB
queries due to a bad template reference or such.
* It will use the master if writes from the current thread
are pending or were recently committed. This deals with the
common problem of code that needs to read things it just wrote,
such as diffs on rollback or edit hooks.
* This commit reverts 8624e261f by making the hack obsolete.
Bug: T93866
Bug: T94407
Change-Id: Ib9ecb75e1236e767bdc86d124d5e22a03ae0fb5f
In DifferenceEngine we need to show a diff immediately after
an edit happened if someone did a rollback. To make that work
again, try loading a revision from master if we have more than
one database servers and we failed loading the revision from
a slave.
Change-Id: I08a16faa9e637c614c6d2fc9a10536c7b41685a8
Follows: 3c2bc32ae1
Bug: T93866
* #top, #mw-content-text and others exist, but all point to the
beginning of the diff, not of the actual page content.
* Pick mw- prefix and technical but familiar related name, in
the hope it's not confusing nor too hard to remember.
* Lowercase letters, similar to #top, #footer, #file etc. were also
considered and discard, inspired to usual names:
https://www.wikidata.org/wiki/Q10966628
Bug: T4313
Change-Id: Ia90f2541327480535e9943ac30bfa19791a19620
Revision->getRawUser()
=> Revision->getUser( Revision::RAW )
Revision->getRawUserText()
=> Revision->getUserText( Revision::RAW )
Revision->getRawComment()
=> Revision->getComment( Revision::RAW )
The body of Revision->getRawUserText() has been moved
into Revision->getUserText().
Every usage has been replaced.
Change-Id: Ic6fbfbc0507dcf88072fcb2a2e2364ae1436dce7
Seems stupid omission. Title has one. Why do I need to think how
to determine how to users objects point to the same user. Allows
more expressive code.
Also fixes a bug in multiple places where users "0" and "00" were
considered equal.
Change-Id: I682392e564b332b77ab489f2ad394fa2d28098a5
Xhprof generates this data now. Custom profiling of various
sub-function units are kept.
Calls to profiler represented about 3% of page execution
time on Special:BlankPage (1.5% in/out); after this change
it's down to about 0.98% of page execution time.
Change-Id: Id9a1dc9d8f80bbd52e42226b724a1e1213d07af7
The new Extension:WikEdDiff is a custom inline difference engine.
There is currently no hook to integrate custom difference engines.
This patch adds a new hook called 'GetDifferenceEngine' in
/includes/content/ContentHandler.php in function
'createDifferenceEngine()'.
Passed variables:
$context: IContextSource context to be used for diff
$old: Revision ID to show and diff with
$new: Either a revision ID or one of the strings 'cur', 'prev' or 'next'
$refreshCache: If set, refreshes the diff cache
$unhide: If set, allow viewing deleted revs
&$differenceEngine: output parameter, difference engine object to be used
for diff
If the hook handler returns false, a valid difference engine object is
returned in the passed-by-reference variable $differenceEngine.
If the handler returns true, the default engine is used as fallback.
The specified diff engine class will typically be an extension of the
class DifferenceEngine (includes/diff/DifferenceEngine.php) with
modifications, e.g. of function generateTextDiffBody() and
__construct() (without deprecated parameter $rcid).
Also fixes a missing declaration in DifferenceEngine that is required for
extending this class.
Bug: 71916
Change-Id: I9da63c1ceb339bfeba7beddc712be51977b95f65
Wrap links in page revision diff header with classes like page revision history
entry links.
Bug: 44242
Change-Id: I0066827bd57c2d40b4b8bd5634cb46c5ce306f28
- Swap "$variable type" to "type $variable"
- Added missing types
- Fixed spacing inside docs
- Makes beginning of @param/@return/@var/@throws in capital
- Changed some types to match the more common spelling
Change-Id: I7b65fe04db431342cc58b469dc48f41a50c4e891
It previously added only the page HTML, which meant that ResourceLoader
modules associated with the page and other metadata was lost. This
affected the new way of doing things in SyntaxHighlight (Idf4ad439).
We still do not show categories, language links and other things
on the pages (although internally they behave just like in wikitext)
Also needs a fix in DifferenceEngine because our code sucks so much.
Bug: 64859
Change-Id: I5867985693b27ccc9195fddde8b922e67463e836
These files have all had treatment before, and these occurrences have either
been missed or have been introduced after.
Change-Id: I06cdab4616b5bff47c85152df28f18c861730a23
Swapped some "$var type" to "type $var" or added missing types
before the $var. Changed some other types to match the more common
spelling. Makes beginning of some text in captial.
Change-Id: I8c9f30128b46086064326708a4878228ba459447
Also swapped some "$var type" to "type $var" or added missing types
before the $var. Changed some other types to match the more common
spelling. Makes beginning of some text in captial.
Change-Id: Ic36c8c7820a6c2d603f1138130670c6bf6a1ca59
- Added spaces after if/foreach/catch
- Added new line before end of file
- Added or removed spaces before/after parenthesis, comma
- Added spaces around string concat
Change-Id: I0590070f1b3542108e242730e8d9a3ba9831e94f
Reasons for touching this are:
* "@param type $var" were mixed in a lot of places. Both works but the
MediaWiki coding conventions suggest that specific order.
* Things like String and Bool aren't objects and shouldn't be uppercase.
* Tried to fill missing types in "@param $var".
* Tried to fill missing descriptions in "@return type" when I could.
* Removed duplicate descriptions if a @see is sufficend.
* Removed useless descriptions ("isUsefull returns true if usefull").
* Removed useless @return void.
* Replaces mixed[] with array (does have the exact same meaning).
* Tried to find better replacements for "varargs", phpDocumentor
suggest $var,...
* Order should be @since, @param, @throws, @return, @see. This is the
order Doxygen renders this.
There is always more to do but I think this is already much better
than before. Please feel free to put more change sets on top of mine
or request more changes by adding comments.
Change-Id: I05262ce06caabf79f68772302524ac91bbead1c6
When one user has made all of the intermediate revisions in a diff,
make it clear whether or not this is the same user that made the
latest revision.
Bug: 59833
Change-Id: I7db9a02ae8bc7a8e092fcdc257b5e17595de02b4
There's no good reason for everything that wants to render a page to
have to test whether the page is a redirect and then call
Article::viewRedirect to get the fancy rendering instead of using the
ParserOutput. This logic can easily be moved into
WikitextContent::getParserOutput so callers can just use the
returned ParserOutput.
At the same time, we can handle "#REDIRECT [[File:Foo]]" and
"#REDIRECT [[Category:Foo]]" the way people expect, by recording the
link in pagelinks rather than imagelinks/categorylinks (although this
means fixing ImagePage's bug about assuming anything in imagelinks that
is a redirect is a redirect to the image too).
And we can finally fix bug 14323, too.
Bug: 14323
Bug: 17259
Bug: 27621
Bug: 42642
Bug: 50488
Change-Id: Id44d566a7ca35a1b9579d0c0e947877c980b0686
Followup: I1c7582d1bf7ec4184a45b00154e3dd5b39dd444b
Followup: I3653b608941813a73281f4f0545bea2487d43964
Part of program to remove underscores from class names. Checked core and
600+ extensions for occurrences. All uses are in core in core are updated
in this patch. No uses in extensions.
Change-Id: I86b8c6f8702e661554c7b794df09892db94a84d1
Part of program to remove underscores from class names. Checked core and
600+ extensions for occurrences. All uses are in core in core are updated
in this patch.
Change-Id: Ib3094249ab192db80f1639c5e543ece84c688b9a
This allows derivative DifferenceEngine classes to
generate cache keys in a different format, as appropriate.
(e.g. for Wikibase, allow diffs to be cached by language
and fully localized)
Bug: 55667
Change-Id: I22bf4e70f86da832a86baf6790ad4a403fce4bf1
* Move some classes to separate files to avoid having a monstrous file with a
dozen classes.
* Remove weird underscores from class names.
* Instead of prefixing members with underscores, specify visibility explicitly.
* Rename c_style -> camelCase per coding conventions.
* Throw exceptions instead of fataling out with trigger_error().
* Remove pointless constant used only once.
Change-Id: Icac23c0f6259c73f5fe07f201b83b9c332ba0469
== Prelude ==
wl_notificationtimestamp controls sending the user e-mail
notifications about changes to pages, as well as showing the "updated
since last visit" markers on history pages, recent changes and
watchlist.
== The bug ==
Previously, on every view of a page, the notification timestamp was
cleared, regardless of whether the user as actually viewing the latest
revision. When viewing a diff, however, the timestamp was cleared only
if one of the revisions being compared was the latest one of its page.
The same behavior applied to talk page message indicators (which are
actually stored sepately to cater to anonymous users).
This was inconsistent and surprising when one was attempting to, say,
go through the 50 new posts to a discussion page in a peacemeal
fashion.
== The fix ==
If the revision being viewed is the latest (or can't be determined),
the timestamp is cleared as previously, as this is necessary to
reenable e-mail notifications for given user and page.
If the revision isn't the latest, the timestamp is updated to
revision's timestamp plus one second. This uses up to two simple
(selectField) indexed queries per page view, only fired when we
do not already know we're looking at the latest version.
Talk page indicator is updated to point at the next revision after the
one being viewed, or cleared if viewing the latest revision. The
UserClearNewTalkNotification hook gained $oldid as the second argument
(a backwards-compatible change). In Skin, we no longer ignore the
indicator being present if we're viewing the talk page, as it might
still be valid.
== The bonus ==
Comments and formatting was updated in a few places, including
tables.sql and Wiki.php.
The following functions gained a second, optional $oldid parameter
(holy indirection, Batman!):
* WikiPage#doViewUpdates()
* User#clearNotification()
* WatchedItem#resetNotificationTimestamp()
DifferenceEngine gained a public method mapDiffPrevNext() used
to parse the ids from URL parameters like oldid=12345&diff=prev,
factored out of loadRevisionIds(). A bug where the NewDifferenceEngine
hook would not be called in some cases, dating back to its
introduction in r45518, was fixed in the process.
Bug: 41759
Change-Id: I4144ba1987b8d7a7e8b24f4f067eedac2ae44459
It's elitist mathematical jargon. In all cases dealt with here, it adds
no additional meaning compared to "if", beyond what was already obvious
from context. Thus, its only purpose is to smugly demonstrate that the
author attended their second-year mathematics classes, at the expense of
causing confusion for everyone who doesn't have such a background.
If you really think you need to convey extra information beyond what
"if" gives you, the English language contains plenty of devices for doing
so, without resorting to neologisms.
Change-Id: Iae21095d02ec2935c10e94f532235c2671c115b1
wfDl() is a wrapper around dl(), which is an evil function and
basically only works from the command line of Zend. Luckily
no extension has ever used this thing, so let's just remove it
outright.
For comparison, here's a list of places it does not work:
- hhvm
- php as apache module
- php compiled with zts support
- safe_mode
- Basically any shared host that cares about security
Most callers are using it to check for extension support and are
actually failing gracefully when wfDl() returns false. In these
places we're just going to use extension_loaded().
While we're at it, clean up some of the test skip logic in the
media tests so we can bail as early as possible if we know we
can't complete the test.
This also immediately removes $wgLoadFileinfoExtension. It's been
enabled by default since 5.3 and falls back gracefully when the
support isn't available.
Change-Id: Ieb430dfc74483731dde51d6e20fa700d641ba1f4
Currently if the revision exists but its content is not found, it will make
loadText() return false; which in turn will make a call to showMissingRevision()
from showDiff(). However since the revision exists; it will not match the
condition in showMissingRevision() and thus an error message will appear without
any revision ID.
Change-Id: Idd8a5f20a3c082a3b02bd77557e0f4dc4c66e876
The "if" always exit, so there is no need to use else branches;
also added blank lines for better readability.
Change-Id: I7d8321652a90fbba99e53fa0c1fe018492883b8a
Add the moment the message is added within a class=diff-multi,
which is not true.
Bug: 53168
Follow-Up: I458fb688b0001fb674ece65b3fdabf56fc658a29
Change-Id: Ic3040ceca4ff1459181c84f041490e9e72b12802
This makes it easier to see the fact, because without a hint, the user
can mean, that there is something missing, but here the missing is okay
and for that, the message is helpful.
Change-Id: I458fb688b0001fb674ece65b3fdabf56fc658a29
This adds two very simple queries for every diff page view.
It could be made to only add one (loading tags for both revisions at
once), but it would be a little ugly.
It could even be made to add zero, but this would require either
rewriting and duplicating a lot of code here and constructing
Revisions by hand or making the Revision itself know about its tags.
Bug: 25824
Bug: 49602
Change-Id: Ic2ae58c703db7ceee5de4b320229d8c93810a73b
This prevents a gap from appearing, and the resulting code will be
slightly cleaner as well.
Shouldn't matter in core, but does matter for extensions displaying
diffs like AbuseFilter.
Change-Id: I9e3e74226a23c0a18db091bc1550b694d9d08118
In multi diffs we had a "wrong" value for rc_last_oldid
set (not the directly previous revision but the one
selected by the user). Due to that there weren't any
unpatrolled recentchanges rows found. As that is cached
in memcached these rows stay unpatrollable forever.
Bug: 49019
Change-Id: I5d6111032dba7d102e30d85718639697f44c036e
Affects whether these modules are loaded:
* mediawiki.searchSuggest ($wgEnableAPI only)
* mediawiki.page.watch.ajax ($wgEnableAPI, $wgEnableWriteAPI,
'writeapi' right)
* mediawiki.page.patrol.ajax (same as above)
Checking of $wgUseAjax has not been removed where it was
already present, in case some users have set the variable
to false to disable these specific features.
Bug: 30213
Change-Id: If2ec219cfbb94e7c9718c58b9b54a508d0e0c656
I've changed the logic in Article::showPatrolFooter to be able
to fetch the recent changes id and to only show the patrol link
in case the change hasn't yet been patrolled.
In case recentchanges patrolling is enabled this will try to
create a patrol link for the revision the user is currently
viewing. If only new page patrolling is enabled it tries to
create a patrol link for the first revision of the page.
Furthermore I've removed the passing around of &rcid parameters
within MediaWiki as those had several issues (some even security
related) and were only a workaround to protect the DB from some
queries, which is no longer needed.
This has already been partly implemented in a different manner in
r45778 but had to be reverted in r46542 due to performance issues.
This version shouldn't cause such issues as I'm only adding one or
two indexed database queries per page view.
I've written this new version of the patch with mostly
performance in mind and even tested the database queries it uses
against the replicated databases of enwiki on the toolserver. I'm
pretty sure this can't be implemented any faster without creating
a new index on the recentchanges table.
As I was on it I've implemented RecentChange::isInRCLifespan which
checks whether the given timestamp is new enough to may have a RC
row. That way we can avoid some DB queries for timestamps which are
older than the max RC age.
Fixes bugs:
(bug 15936) New page's patrol button should always be visible
(bug 35810) ! N pages non-patrol-able
(bug 36641) Patrol page link shows on non-existent revs
Change-Id: I1e24733cafbfdc51b7a5a9a1c1baf948e760fe1a
_split() copied two arrays N times, where N is number of lines in diff.
This was done by $a = array_merge($a, ...);
Instead of doing this, new words are appended to the end of array
using []= syntax.
Bug: 47989
Change-Id: I41338a2a82fbc20d7511f4c79581880febeeeea5
"Quickbar" was a feature of the Standard and CologneBlue skins that
allowed the sidebar to be displayed on left or right side of the page,
floated or fixed, and hidden on diff pages.
Standard was removed (Ia6d73c2d), and CologneBlue doesn't support this
anymore (bug 41246), so all things quickbar can now be safely removed.
* Removed user prefs option + interface
* Removed related messages
* Removed code for this in SkinLegacy
* Removed dead code in DifferenceEngine and Language
Change-Id: I5e6f7d48d6904a052a3a11547d3ebe6161463018
