The extraneous whitespace in the return value from wfShellExec() causes
multiplying $size to trigger the newly introduced "A non well formed
numeric value encountered" warning in PHP 7.1+.
Work around that by using trim() to get rid of the whitespace.
Bug: T186299
Change-Id: I3d47ef6cc7fb99b4d4840dc847d150c3939ee535
The two lualib/ustring generation scripts run independently of MediaWiki, so
the new wfIsCLI() isn't usable there.
Bug: T184043
Change-Id: I217657d12e16a7b76dc814be5fed03540c461e7c
This should make sure the tests still work, even if the running
node is heavily overloaded.
Bug: T143389
Change-Id: Ic40c8d76c8799c2e9d11f53945276747c199fd02
Iaa880531 added extra frames in the call stack, so the frame being
tested by the "setfenv invalid level" and "getfenv invalid level" tests
was no longer invalid.
Bug: T175065
Change-Id: Id1028e7c8bbb92fb9d7d01ebeabd94e8ba284b1c
The core {{urlencode:}} parser function doesn't encode various
characters in WIKI mode that it does in other modes. mw.uri.encode
should match that.
Bug: T174239
Change-Id: I2be0811cf39c02c5c0ad3433e4b0ef9030350e24
Make the language cache size configurable, and increase the default from
20 to 30. It needs to be fairly small on default installations, but can
be essentially unlimited if $wgLocalisationCacheConf['manualRecache'] is
true.
Bug: T85461
Change-Id: Idb17691b30b0d2565a1624e5159df7d9b795764d
I252ec046 noticeably broke things by adding a dependency on the pcntl
functions, which tend not to be present under Apache.
It also subtly broke exit handling by using proc_close()'s return value,
which PHP mangles in such a way that we can't tell the difference
between an actual XCPU kill and exit( SIGXCPU ). This one wasn't noticed
because the pcntl functions interpret everything proc_close() is going
to return as a signal kill and we didn't test the 'exited' code path.
I'm not sure what was going on in I57cdf8aa since it provides no details
about what it was trying to fix, but that would have broken signal
handling in the other way: Ibf5f4656 worked because proc_open() on Linux
executes the command by passing it to /bin/sh -c, and that shell is
going to turn any signal that kills Lua (e.g. the SIGXCPU) into an exit
status of 128+signum.
To avoid proc_close()'s broken return value while also avoiding the
race, we can loop on proc_get_status() until $status['running'] is
false.
To have signals that kill Lua actually be interpreted as signals, we
have two options: add an "exec" in front of the command so proc_open()'s
/bin/sh -c is execed away, or detect shell-style signal reporting and
convert it. We may as well do both.
Bug: T128048
Change-Id: I8a62e1660fe1694e9ba5de77d01960c1ab4580aa
When investigating a test failure on T155600 I found the title
'scribuntotitletest:Module:TestFramework' to be rather obscure. Had to
browse the code to findout 'scribuntotitletest' is an interwiki prefix.
Update TitleLibraryTest php/lua and change the interwiki prefix to the
more meaningful 'interwikiprefix'.
Bug: T155600
Change-Id: Iec63734c82b0835a7f2444ea9af35617876831d5
https://www.mediawiki.org/wiki/Extension:Scribunto says that master
requires 1.25+, so let's remove checks for stuff that was added before
that.
* PPFrame::getTTL() was in 1.24.
* PPFrame::setTTL() was in 1.24.
* PPFrame::isVolatile() was in 1.24.
* Parser::fetchCurrentRevisionOfTitle() was in 1.24.
* ObjectCache::getLocalServerInstance() was added in 1.27, so restore the call to ObjectCache::newAccelerator() as BC.
This also removes BC with the php-luasandbox extension older than 1.6, which
was released before MediaWiki 1.22.
Bug: T148012
Change-Id: I36e37f3b65d0f167e1d28b00e0842d9721feee31
An empty pattern isn't "safe" since it could match in between the
bytes of a UTF-8 character.
Also, it turns out there's a bug in PHP <5.6.9 preg_replace() that we
need to work around too.
Change-Id: I282e5909e4663461d60c5386693db182de2fd44c
Implement Content::prepareSave() to ensure that any content
directly passed to WikiPage::doEditContent() that doesn't run edit
filters will still be validated. We have to use prepareSave() instead of
Content::isValid() because validation depends upon the current Title.
Create a ScribuntoContent::validate() convenience function to hold the
logic for that and add a todo to use it in the EditFilterMerged hook.
Also, remove a parser test that depended upon being able to save invalid
modules directly, as what it is testing is no longer possible (unless it
pre-dates making valid syntax a requirement).
Bug: T145548
Change-Id: Ie57eff36100963f02899d669df7375577f7375e1
Provides a simple wrapper for PHP's hash() and
hash_algos() functions.
I will add docs to the Lua reference manual once
this is merged.
Bug: T142585
Change-Id: I6697463974a175e99f9b77428a1085247165ebc9
For the PHP implementation, return the codepoints as a table instead of
multiple return values that get table-ified in Lua, to avoid hitting
too-many-values stack limits.
For the pure-Lua version, inline most of ustring.codepoint instead of
calling it to avoid what's effectively "{ unpack( stuff ) }".
Bug: T118687
Change-Id: I105f388cc23ab55d4124739700ef89d5354b7dbc
Core strip markers were changed in T110143 to include characters that
are normally encoded in attributes, however we want to pass them through
here so they can be unstripped correctly in the output wikitext.
This fix makes "Strip markers in CSS" parser test pass again.
Bug: T110143
Bug: T135961
Change-Id: I1353931a53c668d8a453dfa2300a99f59fdb01c5
The following continue to be ignored:
* Generic.Arrays.DisallowLongArraySyntax.Found, because I'm not sure
Scribunto is ready to abandon old version support in master.
* MediaWiki.ControlStructures.AssignmentInControlStructures.AssignmentInControlStructures,
because it's overly strict for its purpose.
Squiz.Classes.ValidClassName.NotCamelCaps isn't ignored globally, we
just ignore it explicitly every place it's needed.
Change-Id: I307668da6ef7b3e23da19b1fd1e08914239b99b3
This also makes some updates to make-normalization-table.php to handle
the move of UtfNormal to a separate library.
Bug: T126427
Change-Id: Id4985c3ca441cf92f08ba1f1af85c762ba43d7d2
When caching results from frame:preprocess and frame:expandTemplate,
restrict the scope of the cache to the frame object that was used. This
allows the integrity of the empty-frame expansion cache to be maintained
while also allowing parent frame access. This change is the equivalent of
I621e9075 in core.
Change-Id: Iae4c00e7e19ba12cfdaac135be16c991d9d0cea1
The new Scribunto_LuaTitleLibrary::redirectTarget() method is
used by mw.title objects as read-only attribute 'redirectTarget'.
If the page does not exist or it is not a redirect, the value
of the attribute is `false`; otherwise, it is the target of the
redirect page, as mw.title object.
This is a proper alternative to parsing wikitext as it is done in:
https://en.wikipedia.org/wiki/Module:Redirect
Bug: T68974
Change-Id: Id4d9b0f8c1cd09ebc42c031d4d3fc0c33eea44aa
Apparently microtime is a bit unreliable for this test for some reason,
so let's use getrusage() to measure actual CPU time instead.
Bug: T125045
Change-Id: Ia1ab6b043b99abb4fc6c2989ad09a24d97dd09c3
HHVM has a bug in json decode https://github.com/facebook/hhvm/issues/5813 .
Skip the test until there is a fix upstream.
Bug: T103346
Change-Id: I7e44d98c29ba9b9f5443665fc046382f696193c9
Lua actually treats a close-bracket at the start of a bracketed
character class as a literal, rather than using it to close the
character class. Probably unintended behavior, but it happens.
Also, have the pure-lua version throw our more informative errors on
error even when falling back to string.find and the like, and fix some
other weird edge cases that came up in testing.
Bug: T95958
Bug: T115686
Change-Id: Iab783d4a3e58b1514cc09729d4a71c2cb1242ee8
A string with a dot pattern is only "simple" if
followed by +, - or *. The end of string condition was not checked
properly.
Change-Id: Ia10b9164caeabe464c76441cc82eef37a7013048
And now Ie9e334e broke the tests again. Why can't anyone be bothered to
fix things they should *know* by now are going to need fixing?
Bug: T90018
Change-Id: I7f71dda4af38b23f89c64b6116c1145c44f5c523
The parser's output was changed in I562e437e to add clickable anchors
for section headers which made Scribunto parser tests fail.
Bug: T90018
Change-Id: Ia74877e0641f2763d42be4a8fad1aa8ecffe449e
Should be good to apply now after Wikibase has been adopted to
this.
This reverts commit 06f03fe09b.
Bug: T69483
Change-Id: I28073a891b8847fc6aa5ad026bf702e4015f7d4a
* For most tests, subclassing MediaWikiLangTestCase instead of
MediaWikiTestCase works.
* The Site library needs to not reuse its namespace cache if the
language changed.
* ScribuntoExceptions are still being localized for some reason; test
the message and args of the exception instead of using getMessage().
Bug: T69483
Change-Id: I1c0ca3b11b6228654b47ceb92c1b1be7bda6709a
* I0ede3169 caused Language::fetchLanguageNames() to use 'en' rather than
null when given an invalid language code. This didn't break Jenkins
since, without cldr, it always uses null anyway.
* I4bead5f5 caused Language::listToText() to start encoding
single-quotes in messages such as 'and'.
Bug: T85854
Change-Id: Ic2807be8dad3ef029652e4b0db05e4dd8a6631b9
This makes the interwiki map available to Lua modules. The code is
based on the API interwiki map code in core (the appendInterwikiMap
method of includes/api/ApiQuerySiteInfo.php.) Everything that the
API includes is added, apart from iw_api and iw_wikiid, which I
couldn't think of a use for from Lua modules.
Accessing the interwiki map would be useful for modules like
enwiki's Module:InterwikiTable,[1] as it would stop module writers
having to duplicate the data.
[1] https://en.wikipedia.org/wiki/Module:InterwikiTable
Change-Id: Ie8ad2582aaf5e422824f7da51714a347bb4041d1
When os.date("*t") or ("!*t") is called, instead of just setting the TTL
to 1 second, create a metatable that sets TTLs as the values are looked
at.
Change-Id: Id1e2df731f182f21cf19708738f9907fa927185c
Currently, mw.title.new always results in a database query, which holds up
the parse until it finishes. This changes it to not require a database
query if it's not actually necessary.
Bug: T68328
Change-Id: I62f347d4cd9176bd0440215dcbe804c1dc3d4c99
Add more information to error messages in mw.html. This includes the
error level, the function name, and the position of the argument in the
argument list. Where possible, use the functions in libraryUtil.lua to
do this.
Some functions in mw.html accept multiple types, so add a checkTypeMulti
function to libraryUtil.lua to make these kinds of functions easy to check.
And while we're at it, add test cases for libraryUtil.lua as well.
Change-Id: If9cf9a52bd4b1bb42cc7f9f1f1096828710cbc52
mw.text.unstrip is too broad, it's allowing for unstripping things that
cause problems when unstripped (e.g. bug 61268). Since the original
request was only for unstripping <nowiki>, let's add a function that
does only that.
We should also add an interface to StripState::killMarkers(), instead of
requiring everyone to roll their own work-alike.
Then, to fix the bug, we can make mw.text.unstrip be the combination of
the two. This is the most like the original behavior of mw.text.unstrip
(removes all strip markers, replacing them with text where applicable)
without causing issues.
Bug: 61268
Change-Id: I3a151fd678b365d629b71b4f1cb0d5d284b98555
Scribunto currently supports libraries with PHP callbacks that are
loaded on startup, and pure-Lua libraries that may be loaded from the
module with require().
This change allows for libraries with PHP callbacks to also be loaded
with require().
Change-Id: Ibdc1f4ef51b1c8644c3d4c98d57755b5c06447a5
Instead of just displaying "Script error" in wikitext, display the actual
error string. Currently, many users don't see meaningful error messages
generated by modules, and instead just assume that "Script error" means
the module itself has a bug.
Bug: 71038
Change-Id: Ie5358049975352e0fef92088a06b97ad94717000
It's not necessary, it makes the output bigger, and some pages have enough
elements with CSS that it does make an actual difference.
Change-Id: I80d471899c7e04a8a4876c205198a8c0d0b1f281
If a function name contains an equals sign, it doesn't increase the index,
so indexOffset=1 makes Lua see arguments as starting with 0. This change
fixes that, so that indexOffset is only set to 1 if the function name has
an index.
Change-Id: Ifa2dfad4d2f7228c3fd5721ccfc4c235576b663d
Add a way to fetch cascading protection information from Lua without
needing to call the CASCADINGSOURCES parser function.
Change-Id: I1b3ac18af11d3066f78d27b31da8d6709a6a2631
The pure-Lua ustring pattern matching functions short-circuit to the
much faster string library when the pattern would match the same against
the raw bytes.
A pattern like "[^a-z]" can match a partial UTF-8 character when applied
bytewise, and so must be detected as unsafe.
Let's also directly test the pure-Lua module, instead of me having to
comment out lines in Scribunto_LuaUstringLibrary::register() whenever I
want to test them.
Change-Id: I91ed3374aadfea379b9db2e13b4248ab20df509e
Clean up trailing whitespace from all of our code, and add comments
indicating that apparently unused variables are ScopedCallbacks.
Change-Id: I8e5997797cc7b1c64c5351ec112a18f30edc8fef
Two similar bugs are handled here:
* mw.getCurrentFrame() doesn't work when the module is loaded (only when
a function is called), which breaks os.date and os.time at module
scope since I59ad364d.
* mw.getCurrentFrame() gives access to frame args from inside
mw.loadData, which allows for data leakage between #invokes.
Bug: 67498
Bug: 65687
Change-Id: I82dde43e2601b59c03c6ed4b9365829c40a953a5
From the PHP subclass of Scribunto_LuaEngineTestBase, Lua tests run by
Scribunto_LuaEngineTestBase::testLua() may be skipped by adding an entry
for the test name to $this->skipTests.
From the Lua code run by Scribunto_LuaEngineTestBase::testLua(), tests
may be skipped by calling the "markTestSkipped" function exported by the
TestFramework module.
Also, use this new mechanism to skip certain mw.language tests if
Extension:CLDR or a similar extension is not available.
Bug: 67343
Change-Id: I4f0b15073a84bf2f9d8a5d905c3c960941dd2b71
Some functions in mw.html accept numbers as arguments, but later fail when
constructing the string. This disallows numbers in attribute names, since
they aren't valid anyway, and fixes the remainder of the cases to properly
build the string.
Bug: 67201
Change-Id: Ie7bcbb9d8df580dd8793681f78a8b0719d8a287a
Lua's string functions tend to auto-convert numbers to strings. We
should do the same in mw.ustring.
Bug: 67201
Change-Id: Icd3c5e93bac19dafd78d737ec9b315daba9f1729
Certain wikitext, such as that containing Cite.php <ref> or <references>
or the #tag versions of the same, should not be cached. This uses the
isVolatile method added to PPFrame in I95b3cf87 to avoid caching the
preprocessed output of such wikitext from frame:preprocess and similar
methods.
Bug: 46815
Change-Id: I1084f87fd863eb22f2f3f3d3ff308b24e20a08ef
When os.date, os.time, or mw.language:formatDate are called, set the
appropriate TTL on the output. This needs I412febf3 in core to function at
all, and I3f5a80aa in core to function with formatDate.
Change-Id: I59ad364d502fc247500d94c5606516ad9f98a24d
Rather than calling error() when nils get passed to mw.html methods,
either remove whatever it was that the nil would go to (if that makes
sense), or just do nothing. The seemingly inconsistent use of "not x" and
"x ~= nil" is to allow any falsey value where it wouldn't be ambiguous
(such as class names), but not where it could be (such as attribute values).
Bug: 62982
Change-Id: I76773abbb4394aa9bb8c8a08445e019cade3b2bf
When tables are passed from Lua to PHP, their metatables are lost. Because
of this, they need to be kept inside of Lua to allow the __index
metamethod to return a method to be called by #invoke.
Bug: 64141
Change-Id: I0840bc12b25dee72828ec97d2b205812e4929f2b
From wikitext, $parser->callParserFunction() will always get an array of
strings with at least an element [0]. Let's match this from Scribunto:
stringify numbers, and require that [0] (although in Lua it'll be [1]).
Also fix an old broken unit test.
Bug: 63597
Change-Id: Ie7ac34ae4bce70cec455d90c3f02a658644f6866
When displaying a nosuchfunction or nosuchmodule error, include the name
of the nonexistent function or module.
Change-Id: I17fc2c68dc8267302a82eee3cb2c5df9b5a3c46c
It's possible to pass information between multiple #invokes on a page by
having the first call math.randomseed with one of a set of known seeds
and then having the second examine the output from math.random to
determine which of those known seeds was used.
Prevent that by calling math.randomseed( 1 ) when invoking (see the bug
for details on why that seed). But avoid doing so if e.g. a
frame:expandTemplate() call results in a recursive invoke.
Bug: 62291
Change-Id: Id01cb63eca52ced29bf4efebc38beb9f159b7b0e
Various methods are throwing exceptions when passed invalid language
codes. Those need to be caught.
And we should really add unit tests for the mw.language library, too.
Doing so exposed another bug (in lang:gender), which is also fixed here.
Bug: 62242
Change-Id: Ib7d257cbb1ce179c510273526910d6ac5f3cac5d
The LuaStandalone interpreter needs to keep a mapping from integers
returned to PHP to the corresponding function. But if it never releases
these functions when PHP no longer has any reference to them, it can
result in Lua running out of memory if a module with a large number of
functions is invoked many times in one page.
The fix here is to track which function ids are referenced from PHP, and
periodically send the list to Lua so it can remove any that are no
longer used from its cache.
This also takes care of another issue where having multiple interpreter
instances and passing function objects from one into another could call
the wrong function in Lua.
Bug: 51886
Change-Id: I4f15841051f7748d1d6df24080949e5cbd88f217
Message formats other than plain should have never been exposed in this
way, as they allow link tables, etc. to be bypassed and serve no useful
purpose. This removes them, and also removes title, as it serves no
purpose without them.
Bug: 60758
Change-Id: I96284ffbe986a9cd92d2bde1ffdb746029bad989
If we don't do this, then the section edit links point to the wrong page
if we expand a template that contains section headings.
Bug: 55525
Change-Id: I00bda935be3e8b9c0f86fd0f131814207fbb34a7
Include a protectionLevels variable in the output of the mw.title.*
functions, containing the contents of the title's mRestrictions
array (i.e., its protection levels)
Change-Id: I79c9fed64bacfc90aee1d411a3e1b47e44c99755
PHP can't handle having arrays/objects or functions as keys in its
arrays, so make sure we don't try to pass them from Lua. Booleans aren't
really well-handled either, so let's disallow them too.
Also, add tests for proper stringification of floats and infinities when
those are used as keys.
Note this behavior change is needed to match the change in LuaSandbox
for fixing bug 54527, but isn't itself a security issue.
Change-Id: I1e2951bbe8cb78358650ad377bf7119fcac4485d
A module for building complex HTML from Lua using a
fluent interface. The module is originally from enwiki,
but the authors allowed us to reuse it under GPLv2+
(as stated in the file).
The module will be loaded per default and comes with
unit tests.
As discussed on wikitech-l:
http://lists.wikimedia.org/pipermail/wikitech-l/2013-December/073320.html
Change-Id: I7c8d4378091c13d5ace0dd1fcbb4e27163e8c896
This field already exists in PHP with exactly the content requested in
bug 47089, so we may as well expose it on the frame object.
Bug: 47089
Change-Id: I672820589f6ebc7c4daad29b5eb156733a5bc5cc
It's already possible to detect whether the current template is being
substituted via ParserFunctions (see [[en:Template:Ifsubst]]), and a
similar trick works with frame:preprocess. So we may as well provide the
flag directly.
Bug: 47828
Change-Id: Id06d27c6283ee589a8830b78c04e56978e0ac6da
Specifically:
* String conversion in non-URL contexts (e.g. .prefixedText) uses spaces
instead of underscores.
* Setting .fragment now applies the same transformations that are done
(in PHP) by mw.title.new.
Bug: 56217
Change-Id: I12e354636bcde3327864088175fb4de61aecc81a
