The following errors are fixed:
* PHP warning and wrong return value with empty pattern and plain
* Incorrect offsets returned when init is larger than the string length
* Incorrect captured offsets returned when init is excessively negative
Bug: 47365
Change-Id: I9741418287dc727747326d6a19678370ce155a2b
We had intended to use the constant CONTENT_FORMAT_TEXT, but
accidentally used that as a string instead.
Change-Id: I93a2c02d48d3fd7b73530562165d817965e272d1
Two related issues:
* The package module was inheriting the loaders from the outer sandbox,
so loaded modules were being loaded into the outer sandbox's
environment.
* mw.loadData was using the outer sandbox's require(), so again loaded
modules were being loaded into the outer sandbox's environment.
Bug: 47300
Change-Id: I48d8dd4784c9a890e3abb6389f96f38e1420dbbb
The documentation, and the expectation of users, is that
lang:parseFormattedNumber() should actually return a number, not a
string.
Bug: 47268
Change-Id: Ieabddd0d9192f1fd8ef7e890d5d6268be9636f38
One of the design goals of Scribunto is that each #invoke should be
independent. Creative use of mw.log and mw.getLogBuffer can get around
this, passing information from one #invoke to the next.
This patch takes the simple solution of removing mw.getLogBuffer from
modules' environments. For good measure, it also removes
mw.clearLogBuffer and mw.executeModule.
Some minor cleanup of the console code is also included.
Change-Id: I30d73928bade4a6bdd7c00ffcd58a3858ff55698
https://meta.wikimedia.org/wiki/Lua_deployments/Localization_of_Module
Four languages (ka, ms, zh, zh-yue) are missing the translation for
Module talk, as there wasn't one on the Meta page.
Full list of newly added languages (including the ones with partial
translations): an, ar, ast, ba, bn, ca, crh, da, dv, es, et, eu, fa,
gl, id, it, ja, ka, ko, min, ms, nn, no, pl, ru, sk, sl, uk, zh,
zh-yue.
Change-Id: I504ce98a2a430aa98f116186051331fbbfcc57d5
This uses the timezone support added to core Language::sprintfDate in
Iea1f7842 to add support for various timezone-related formatting
characters, and to correct the output for 'c', 'r', and 'U' in local
mode.
This is related to bug 33454, which requests the same for
ParserFunctions' {{#timel:}}.
Bug: 33454
Change-Id: I1b92dc671051a6cc53e35ebd74d383448e16696c
On Windows for LuaStandalone, the lua executable's standard output is a
text-mode file handle, even if the pipe is opened from PHP with the
binary flag. Which means that when Lua returns a "\n", it gets silently
rewritten to "\r\n" and the unserialization fails.
So, change the protocol for Lua→PHP messages to encode \r and \n (and \
itself, as the escape character) to avoid this issue.
Bug: 46294
Change-Id: I73b5f44e8aa0334f5fd03013dc027d1a57318349
Users are reporting disappointment that using pairs on title objects
doesn't let them see the fields available. It's easy enough to add a
__pairs metamethod to allow pairs to work on title objects, so let's do
that.
The same can be done for mw.uri objects.
For mw.message objects, we can easily enough change the implementation
to be like mw.language objects, which doesn't have this problem.
For mw.site.stats, we may as well just remove the load-on-demand feature
since it will be demanded as soon as the environment is cloned for the
first module.
Change-Id: Ie5a3b5684b2bb6c090f9994baa03977687688929
The logic for recognizing absolute versus relative paths needs to take
into account the possibility of a Windows drive letter.
Bug: 46635
Change-Id: I3a43acac2f6e8b481807e1babe5a261b9eb1fe23
Use the SoftwareInfo hook to add the versions of LuaSandbox and Lua to
Special:Version.
Bug: 39655
Change-Id: I912197efee0211066677c4d46e638fb546a410c6
mw.title.new( pageid ) should not throw an error for an nonexisting
pageid, just return nil. Similarly, it should always return nil for 0,
rather than returning the last non-existent title created.
Change-Id: I3cdbb24fc785aef0f8e75fba1feccd26ac5b7370
This exists for some common text-processing functions that aren't
included in string (and therefore also aren't in mw.ustring), as well as
a logical place for the "unstrip" function requested in bug 45085.
Bug: 45085
Change-Id: I47356215fcc8ddeed5f901cd933a30021394bd78
On my test system, ScriuntoHooks::invokeHook takes about 4 seconds of
CPU time during the parse of a page containing 1000 #invokes of a module
containing a single do-nothing function. And about 2.3 of those seconds
appear to be spent in ScribuntoEngineBase::fetchModuleFromParser.
ScribuntoEngineBase::fetchModuleFromParser already assumes that if
Parser::fetchTemplateAndTitle returns the same $finalTitle then $text is
also the same and we can reuse the same module. But it seems just as
likely that the same input $title is always going to give the same
$finalTitle, so we may as well skip calling
Parser::fetchTemplateAndTitle entirely if we've seen the input $title
before.
This change takes the CPU time spent in ScriuntoHooks::invokeHook in the
test described above down to about 1.5 seconds. And since it's very
likely that $title and $finalTitle are the same (Modules don't support
redirects, so it's basically TemplateSandbox that would cause that), it
probably doesn't even increase the size of the module cache.
Change-Id: I87ad8b85d0f82791f49158d62effa6dc7c20f058