Add new streamFileWithStatus() methods to FileRepo and
MediaTransformOutput that can be used to get more detailed error
information on failure. The historic streamFile() methods become sinple
wrappers to the new methods. Thumb.php is changed to use the
streamFileWithStatus() methods so that failure reasons can be logged.
Change-Id: I3088cde2044a7ff00841e53ca252d0b222c8b518
The callback is named doCachedWork, not getCachedWork. The typo
made it so that if multiple people try to render a file at once,
MediaWiki would not check if the last person to go already did the
needed work.
This might not make as much difference as it would sound, as the
pool counter lock is for the entire file, not a specific size,
and I'm unsure if swift has an equivalent of "slave lag", where the
check to see if a cached version is available might happen before
the write is totally visible.
Also merge the fallback handler into the error handler, so we get
actually error messages reported.
Change-Id: I12d228961de39cffc3d492554d93e30bd741adc9
Also consistently use self:: instead of BagOStuff:: for constants
referenced within the BagOStuff class.
Change-Id: I20fde9fa5cddcc9e92fa6a02b05dc7effa846742
* This add noise to logs and graphs that can look like outages
due to a single crawler.
Bug: T110109
Change-Id: Ie2a1edd0330b0f858729fcc56c828c11d9c7476c
Without the Content-Length header, the response is sent with
Transfer-Encoding: Chunked, which is somehow mangled by
mod_fastcgi. Varnish then claims the response is malformed and declines
to process it, sending the client a 503 instead.
This is a followup of 0ee9e717f4
Bug: T84842
Change-Id: Ia610a43789e6ff14cfc0964f285bbec39c890152
Without the Content-Length header, the response is sent with Transfer-Encoding:
Chunked, which is somehow mangled by mod_fastcgi. Varnish then claims the
response is malformed and declines to process it, sending the client a 503
instead.
Change-Id: I977387bf3e039926df58e3e5a14d0d0bd55201c4
Functions must not contain multiple empty lines in a row
(Squiz.WhiteSpace.SuperfluousWhitespace sniff).
Bug: T102774
Change-Id: Ib706e4fc34b95c0d7c887e54af0ea61227767d8f
wfSuppressWarnings() and wfRestoreWarnings() were split out into a
separate library. All usages in core were replaced with the new
functions, and the wf* global functions are marked as deprecated.
Additionally, some uses of @ were replaced due to composer's autoloader
being loaded even earlier.
Ie1234f8c12693408de9b94bf6f84480a90bd4f8e adds the library to
mediawiki/vendor.
Bug: T100923
Change-Id: I5c35079a0a656180852be0ae6b1262d40f6534c4
Also:
* Update wfHttpError() to use uppercase DOCTYPE, to match other code
such as Html.php, wfThumbError(), HttpError.php, etc.
Change-Id: I4027e7fe1a138b03f78797b6d1bfe7bd1064d360
* Convert existing use of WebResponse::header() for HTTP status headers
to use this new statusHeader() method.
* Extend unit test forFauxResponse.
I'm not calling HttpStatus::header directly in code. We keep the abstraction
layer of WebResponse so that responses can continue to be mocked/fauxed without
affecting the outer HTTP response.
Change-Id: I8a536e16659fa88b54cffa1457efb889efa5fcd6
* Added doPreOutputCommit() and doPostOutputShutdown(),
which most entry points just using the later
* Also fixed problem where text profiling did not show up
* Avoid calling triggerJobs() in the file streaming
entry points
Bug: T100127
Bug: T100085
Change-Id: Ibc7e768fd483389a01847f08cdeba4058c853d3f
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
* Use ->parse() instead of ->text() for wikitext messages that were
being treated as HTML
* Explicitly specify ->parse() if no output format was set
* Document that wfThumbError() takes HTML
Bug: T76686
Change-Id: Id6e7548b2e081cfda7803772ed0395a15feb1f84
- use tab as indent instead of spaces
- Added space after closures "function"
- Added spaces around string_concat
- Added newline inside empty blocks
- Removed four spaces after comma
Change-Id: I4425b0c6a69b36f40acfea6511b8950cf09ce2b2
Follows-up 5ca94d2d26.
In thumb.php an undefined variable ($width) was used, triggering
fatal errors while processing requests that were passing thumbnail
width with px (pixel) suffix.
Change-Id: I2dc4f4d2979f8225d2a6211ec50cdc79438539a9
Otherwise if there is a broken file, and somebody fixes the file,
it could take an hour before thumbs are generated again. This
might confuse the user, and they may revert back to the broken
version thinking their fix didn't work.
Originally reported by Fae on wiki.
Change-Id: I024702391e5f7ccf10a406a29485601f64069405
Depends on Ic62984e0f4a761642b2bdd1bfa362301ed94c284
and If40e223c328ddd15310d13f2a49000bf6e08200e
Bug: 65691
Change-Id: Ic737673b5a31974d866b9bf5afa4b4036baa03bb
A large percentage (40-50%) of the 500s that are emitted in production
are for a single URL,
https://commons.wikimedia.org/w/thumb.php?f=Crystal_Clear_action_viewmag.png&width=21px
The reason this fails is because thumb.php expects width to be "21", not
"21px", and it currently tries to fetch (and generate) the "21pxpx"
thumb size, which is obviously an invalid size. (an invalid size
shouldn't result in a 5xx but rather to a 4xx. though; that's a separate
bug that needs to be fixed).
This URL is embedded by a gadget, Gadget-searchbox-js, that is copied in
a lot of our wikis, including a big one, frwiki. mwgrep reveals that
there are a bunch of other URLs in various Gadgets that have width
values with "px" in them, so this presumably worked at some point in the
past.
While we could in theory fix all those URLs in these dozens of gadgets
across wikis to not suffix width with "px", this sounds like a herculean
effort and we're probably better off adding this compatibility branch to
thumb.php that strips the "px" suffix, if existent.
Change-Id: I3a00c9634b1c6af49fb8503cc3ff4cafdaff6b43
* Generally, either the height or width is used for each description page
link when using $wgImageLimits. This depends on the $wgImageLimits values
and the file dimensions. Description page links should better pass the
wfThumbIsStandard() method now, and thus have weaker rate-limiting.
Change-Id: Id1c3b0fc57f8ed3c14929a1a1661842a069b805d
