I tried to benchmark this. It's not really a big bottleneck in this
codebase, but can still be seen. For example, the relative runtime
of the method CommentUtils::isCommentSeparator (the heaviest user of
this feature) goes down from 1.5% to 0.3%.
Depends-On: If9252a97562542e7a4bec29edcc6e8dee0fb8221
Change-Id: I6b94a2481b5c4f7983df78ee48d45c0d8a50e53b
urldecode() should be used for decoding URL query parameters,
rawurldecode() should be used for decoding URL paths.
Bug: T367977
Change-Id: I7a7b14da85fb89f612c701d2746803d830017842
The PHP code should never see a `<span class="mw-headline">` node
since MediaWiki change If04d72f427ec3c3730e757cbb3ade8840c09f7d3,
but we have to support it for our old integration tests (T363031).
Improve code comments to explain this and move the handling to
one place, so that it can be deleted more easily in the future.
Follow-up to 08f61b2609.
Change-Id: I5ab9d3373a6911c1456c30d844b66576b278a1b5
MediaWiki's PHPCS plugin requires documentation comments on all
methods, unless those methods are fully typed (all parameters and
return value).
It turns out that almost all of our methods are fully typed already.
Procedure:
1. Find: \*(\s*\*\s*(@param \??[\w\\]+(\|null)? &?\$\w+|@return \??[\w\\]+(\|null)?)\n)+\s*\*/
Replace with: */
This deletes type annotations, except those not representable
as PHP type hints such as union types `a|b` or typed arrays `a[]`,
or those with documentation beyond type hints, or those on
functions with any other annotations.
2. Find: /\*\*/\n\s*
Replace with nothing
This deletes the remaining comments on methods that had no prose
documentation.
3. Undo all changes that PHPCS complains about (those comments
were not redundant)
4. Review the diff carefully, these regexps are imprecise :)
Change-Id: Ic82e8b23f2996f44951208dbd9cfb4c8e0738dac
There's now a different rule for the same thing in
mediawiki/mediawiki-codesniffer v43.0.0. Also document the reason for
the override. Follow-up to 8b00546749.
Change-Id: I392ee10639ffda6de55b091555e8c3cadd2af485
We already supported plain headings without the 'mw-headline'
wrappers, but now we need to parse the 'id' from a different
attribute.
Needed-By: If04d72f427ec3c3730e757cbb3ade8840c09f7d3
Bug: T357723
Change-Id: If85f89c40834618f23dc0ace2e599efb3b6d5ed4
In certain cases the parser could go back rather than forward after
finding a signature, causing it to find the same signature forever
until it ran out of memory.
Test cases coming later in a separate patch.
Bug: T356884
Change-Id: I8ac72b05e5e4ed45e6119c012a69708c9d8eda29
This will be present in parser cache output and can
sometimes be mistaken for user page links.
Bug: T356142
Change-Id: I800b23d8466f72affcadfa336aab07abf7f8d79e
Reasons:
* Various other methods dealing with ranges already live there
* It would be neat if ContentThreadItem was just a value class
without a lot of logic, similar to DatabaseThreadItem,
particularly for writing unit tests
* The methods access global state through Title, which can't
be fixed while they're in ContentThreadItem (see I9dfccc83)
The computation is now always done, instead of only when needed,
but that's a small drawback, since it's fast (fast enough that
I don't see the difference in the time taken when running tests),
and we were already computing it for all comments in many places.
Change-Id: Ic718a964e309ae3a8e15e299081f46d4db860731
This might be a matter of personal preference. Not sure if it's
worth it. Both is well readable. On the other hand, the method
exists. Why not use it?
Change-Id: Id66fc6c888db6ae1cf28e60a51f90d9ae2cdb6ee
It does the same as before.
I think performance is not a concern here, and wasn't my motivation
either. But I hope this makes the code easier to read and to reason
with.
I added a pure unit test case (without involving an actual Language
object) to cover the previously uncovered digits feature.
Change-Id: I6a0fc86035817eabb42b55e58183ae094c052aa6
I was curious why running the CommentParserTest takes so long. I
found this is one of the bottlenecks because it's called so often,
but many link titles that are parsed as user names turn out to be
something else. This little hack speeds up the test by 15% and has
probably a similar impact in production scenarios.
Change-Id: I5a0b3a49ba5793c8a345baaa7118fed500c082b6
MediaWiki's PHPCS plugin requires documentation comments on all
properties, unless those properties are typed.
This has potential to introduce bugs – in particular, because typed
properties without a default value will throw an exception if their
value is accessed before it's defined, while previously they defaulted
to null. I fixed this when I found it (making them nullable and null
by default), but I may have missed some cases.
Change-Id: If5b1f4d542ce3e1b69327ee4283f7c3e133a62a0
https://wikipesija.org is currently using ISO 8601 as the default date
format. The format is xnY-xnm-xnd"T"xnH:xni:xns and 'xn', 'm', and 's'
need support added.
Change-Id: I235098a578eb92ddd23ea47fa23d60df4b28f590
Also merge setMwGlobals() calls because they are really expensive.
Also utilize the more readable str_contains() and related.
Change-Id: Iebde6aa17c2e366f0c0a98fe13a454f6a06c299b
This can be done within sections using CSS:
* mw-notalk
Or at a page level using a magic word:
* __NOTALK__
"notalk" suppresses all comment detection, treating the content as
not containing any comments even if there are signatures present.
Bug: T295553
Bug: T249293
Change-Id: Ic1d7294bafcf7071e16838e70684ecadd7bc6fd3
Change code to match the documented consensus formed on T321683:
https://www.mediawiki.org/wiki/Manual:Coding_conventions/PHP#Exception_handling
* Do not directly throw Exception, Error or MWException
* Document checked exceptions with @throws
* Do not document unchecked exceptions
For this extension, I think it makes sense to consider DOMException an
unchecked exception too (in addition to the usual LogicException and
RuntimeException).
Depends-On: Id07e301c3f20afa135e5469ee234a27354485652
Depends-On: I869af06896b9757af18488b916211c5a41a8c563
Depends-On: I42d9b7465d1406a22ef1b3f6d8de426c60c90e2c
Change-Id: Ic9d9efd031a87fa5a93143f714f0adb20f0dd956
* Detect comment separators at the end of comments too
* Consider TemplateStyles associated with ignored templates
This unexpectedly improves a lot of cases other than T313097 too,
mostly where <br> or {{outdent}} was used within a paragraph:
splitting comments that were previously jumbled together, or restoring
content that was previously ignored for apps / notifications.
Bug: T313097
Change-Id: I9b2ef6b760f2ffd97141ad7000f70919aeab7803
The implementation in Parser doesn't descend into sub-thread.
Re-use the getThreadSummary method in ThreadItem and traverse
the thread properly.
Bug: T298617
Change-Id: I318d9012eb83f37ccbe463923524ef2e9f995ced
We can no longer change IDs so easily, because they're stored in the
permalink database, so remove this mechanism to make sure it's not
accidentally used in the future.
Change-Id: I392ee1f49c48fc2f23d05e9a37c643438b4f2b9a
Rename ThreadItem to ContentThreadItem, then create a new ThreadItem
interface containing only the methods that we'll be able to implement
using only the persistently stored data (no parsing), then create a
DatabaseThreadItem. Do the same for CommentItem and HeadingItem.
ThreadItemSet gets a similar treatment, but it's basically only for
Phan's type checking. (This is sad.)
Change-Id: I1633049befe8ec169753b82eb876459af1f63fe8
Change the HeadingItem constructor to take a 'null' headingLevel
and store this internally with the constant. Change the JSON
serializer to convert this back to null.
Change-Id: I27508eed75d94b99c5189548919309f8da7deb75
While in many cases the class will never be sub-classed, it's easier
just to always use static:: and not worry about predicting which
classes might have problems in the future.
Change-Id: I23072a1701b5acf62bb3379a877de97627d8fcf3
We previously ignored them before timezone indicator (e9c401e3aa),
but they can end up in other places too, e.g. after the time.
Now we ignore them after every token. This is way overkill, but it
shouldn't hurt.
Bug: T308448
Change-Id: I20f7aaa34dba23f2a2faf1be258c1aea32ab770f
Phan can analyze them now and reports some issues with types.
* Add some assertions on types where we're sure that we're using an
Element or non-null, but Phan can't prove it
* Fix incorrect type hints on getFullyCoveredSiblings() and
getCoveredSiblings(), luckily it was harmless
Change-Id: I8cc12450378efa7434c4d66882378b715edd4a70
We call CommentUtils::getHeadlineNodeAndOffset() before constructing
the HeadingItem in CommentParser, so the range's startContainer
is always the headline node.
Change-Id: I2afb6ba9100e785cd91f31d82f4cea59fa8b5443
`tagName` is only defined on Element, and it returns its tag name.
`nodeName` is defined on Node, and it returns the tag name for Elements,
and a string like '#text' or '#document-fragment' for other types.
We were using both, which made it harder to reason about what types
we're dealing with.
Change-Id: I8e621e5872bdf78c84ec553cfbfcdbf0192f0589
