In some cases restBASE seems to provide errorous responses
despite a 200 response code.
Check if all required fields exists and treat the response
as failed otherwise.
Bug: T349347
Change-Id: I3506401f8dd2962daf7fd5d30019c507f9de5def
- munder: if only one child munder is not added as wrapping mml tag
- msup: renders an empty mi if there is no content for the superscript
This prevents some warnings in firefox browser dev-console.
Bug: T349825
Change-Id: Ia5efc1f33e3368e71527153d56160eee4ba3a46d
* From Feedback: The {smallmatrix} example has its fences wrapped in too many mrows I think, so it renders without stretching them in both Firefox and Chrome. Ideally you should be able to emit the <mo>(</mo> followed directly by an <mtable>, and trailed by <mo>)</mo>, and only wrap those 3 into a parent <mrow>.
* removes mrows and adds OPEN / CLOSE attributes
* tests MMLFullcoverage 194 to ~196 which cover the bigXYZ macros look
ok
Bug: T348971
Change-Id: Ie444e5bb3e3f87fac915d2692b69e976d721582e
* From Feedback: There is a curious use of <mstyle> around spacing elements, as in <mstyle scriptlevel="0"><mspace width="-0.167em"></mspace></mstyle>. I am not sure that wrapper element is needed? The mspace ought to be successful standalone.
* this is in the 'bmod' case
Bug: T349822
Change-Id: I9229ad73891d5000679761d3d4b2028d4a51ddf1
Allow in depth analysis of rendering results in
native mode with the same mechanism that is also
used for other rendering modes.
Change-Id: I4927f9bc683445f83ef7c29dc2a9298aa85ff4d0
For old mysql versions utf8 encode decode was required to avoid
that math characters get messed up. This problem does not occur
with MariaDB 10.2. As now data is stored and retrieved as binary
by default, the data is returned in the same way as it is written.
Bug: T324173
Change-Id: Id29c9a8614156ff8170bc3148f2498b32c70fd8a
* Currently the native mode does always set render to true
there is no reason to do, as this was initially introduced
for testing.
Change-Id: I632ca5ea726d055f377439e439a39baca3bed385
* MathRenderer passes purge state, the purge state has to be rechecked
since implicitly it is true in the MathRenderer
* for rechecking a new flag is introduced to isPurge function
Bug: T348976
Change-Id: Ifdcea4b2c17ac963a50ee48e2d6437c0feafc602
* rollback some previous changes for ce env in Parserhookhandler
* some minor fixes for feedback in previous commit
* set tex mode for mhchem preprocessor, since MediaWiki Hooks add
explicit ce tags: i.e. frormula coming to TexVC look like
"{\displaystyle \ce{ C6H5-CHO }}"
Bug: T348975
Change-Id: Id06852b030f9ea77f016289d4f9e2258d5a9f0d4
* before an explicit fallback was used to standard MathML mode
* now, in addition native MathML mode can also be used
* added interface code activating texvc for chemical formulas correctly
* ParserHooksHandler removes implicit ce environment for native chem
rendering
Bug: T348975
Change-Id: If2bbc7223d253d78e837eec2e007dd8171aa6349
BaseChecker::getError may be null, thus handle the
case when the error is null and output
unknown_error in that case.
Bug: T348973
Change-Id: I072ebf200139436b4f27a90dfe940cb310886b28
* currently, \begin{align} generates <mtable columnalign="right" but it should be <mtable columnalign="left"
Bug: T348791
Change-Id: Ic0df45aba4fdb6ffbdc790b5e4bc56e7daeca802
* updating the braces since mkern, mskip etc are Fun1 functions in TexVC
* updating TexVC support flag in test
* added macro longrightleftharpoons to rendered cases
* nearly all tests render correct in HTML now
Bug: T329620
Change-Id: I78da466c88d41f794e81f7222c9a67c3f4d482d2
* Remove \strokeint which was probably a typo and should be \strokedint
* Add \intBar command
Not done:
* Implement rendering of \intBar and \intbar, it would make sense to wait for
the mathoid reference rending here, which is just around the corner
Bug: T137787
Bug: T348373
Change-Id: I19d24fc23f59dda87c7c39a265f17efddb9fee8f
* defaults mathchoice to displaystyle
* adds several tests and introduces new Fun4 element for mathchoice in
grammar
Bug: T340024
Change-Id: I37c8fd35303e3b8a00faad18a3a131a716c7bb47
* adds the recently implemented cases for mhchem specific tex to check full macro coverage
* fix a minor issue with tripledash
Change-Id: Id3033fab7b3c3432bed6e11f7b50f7ff8e3906ea
* "\ln c, \lg d = \log e, \log_{10} f" : c d etc should have be more extra space
* "\Pr j, \hom l, \lVert z \rVert, \arg z ": same as above, specifically for \hom
* fixes expected result in LocalCheckerTest
Bug: T315978
Change-Id: I842b45b2d20dc267d828dca2040a3500ce06f45d
* All recognized macros are in category nullary macros, so this adds a
rule
* In test "\bigcap_{i=_1}^n E_i"
* In test "\bigcup_{i=_1}^n E_i"
Bug: T315978
Change-Id: I09287879c1448e404473891985457349a62554a7
* Test example in MathNativeTests: " "\not\operatorname{R}"
* fixed forwarding state in Fun1nb
Bug: T315978
Change-Id: Ia651b14ba16f39cc8cbc6b42e0598d959bf72989
* In test MathNativeTests: "\sum_{k=1}^N k^2"
* also verify that textstyle variant works correctly
Bug: T315978
Change-Id: I8f6cd5ef1e4539616439542184cb7d9dc5cf2363
* also fixes a warning in phan that there is a tainted parameter in args in BaseParsing
* this is to pass CI
Bug: T343636
Bug: T347320
Change-Id: I91328bf26e208bd4e0e939c4628def2f40e2f0b6
* In test MathNativeTests: "\underbrace{ a+b+\cdots+z }_{26}"
* Underbrace statement appear to right side
Bug: T315978
Change-Id: I3b1e1290ac07e7aa70ab5931d513984f011d3075
* Pilcrow was very big, see test on MatTestNative page:
* " \amalg \P \S \% \dagger \ddagger \ldots \cdots "
* and add specific testcase
Bug : T315978
Change-Id: I3074f3782ef821e7f9813b79a570bbd51ead28db
* contains a draft of integrating mhchemParser as a preprocessor for texvc
* this contains also a draft of recognizing some additional statements
coming from mhchemParser
Bug: T340023
Change-Id: Icdfd0bfbfc04b0f1fe89837c2fcd184a77733b0e
* mhchemParser v4.2.2 in Typescript was used as blueprint for this:
* https://github.com/mhchem/mhchemParser
* The mhchemParserPHP component tests can be run locally on a machine with PHP without MediaWiki-Environment with the MMLmhchemTestLocal.php
* For the tests it is probably only necessary to review the json files (Mhchemv4mml.json, Mhchemv4tex.json).
* src/TexVC/MHChem/MhchemParser.php contains the basic functionality, Patterns, StateMachines, Texify functions are located in dedicated classes
* added extracted mhchem commands which have intermediately have been introduced to the
texVC(PHP) grammar
Change-Id: I8cc3d04937b93339e352adc95c85a8a178b8825f
Bug: T329620
The VisualEditorFullRestbaseURL config variable is no longer used by
VisualEditor and is no longer set in Wikimedia production
configuration.
Bug: T344458
Change-Id: Ifd0b9f919f7a9bdceb8761125588fe72e8d1e50c
The use of "HookHandlers" attribute in extension.json makes it possible
to inject services into hook handler classes in a future patch.
Bug: T263364
Change-Id: Id8d9d0dcc6bb6ab60c5df3074803d132393415e9
Add a config option to enable passing special macros needed to
handle the output from the mhchem texify process.
Bug: T340023
Change-Id: I3553931c252184b54e7a13938e030825e5d45e59
texutil.json exists twice with identical contents in
test and main sections.
Delete the redundant copy in the test section and access
the main file in tests.
Change-Id: Ie1eba9ed599d6826f1efe2d37bf8370f8e9dd036
* Cache results for checked tex and MathML string in one cache.
* Remove access to parsetree
* Introduce run method to speed up service wiring
Note that the indirection table used in previous versions was
abandoned here. texvc does only little unification's of the
input string so that it is not expected that the overall savings
in space and compute time warrant the additional table.
Change-Id: Ib9ce3d2ab02bd9a2a0f9926db6b937435b7e5458
In preparation for caching we store the MathML fragment returned by
local checker. The rational is that the string serialization of the
parse tree is better chachable as the parsetree.
Change-Id: Ice2ef1f4f7b83ed187507d4d067f446603d0d6a5
Introduce new classes for checked exceptions, and use SPL exceptions
for the unchecked ones.
Bug: T328220
Change-Id: Ic44463e910911afe6395c38fd2f59b9bfa02a4f0
Remove additional /v1 when computing internalRestbaseURL from
virtualRestConfig as this will be added in getUrl in MathRestbaseInterface.
internalRestbaseURL was introduced in I2fbc441955a1fe417f264f31f3729ce0715f7f16.
Bug: T334842
Bug: T338381
Change-Id: If2719fb9106e1414ce8814a2fbafa5dcedb84e39
This new MathInternalRestbaseURL setting allows the internal URL of
Mathoid to be configured directly, instead of relying on
VirtualRestConfig.
This patch moves the fallback logic for the internal and external API
URL into an extension registration callback. This way, we can keep the
application logic simple, and we can issue config warnings that will
cause the updater to fail, forcing wiki owners to update their config.
Bug: T334842
Change-Id: I2fbc441955a1fe417f264f31f3729ce0715f7f16
This fixes the problems with If208753edfdb301 which was reverted
and adds a new test.
Bug: T335347
Co-Authored-By: <rrana-ctr@wikimedia.org>
Change-Id: I7ddb7b294cbccc4dab42e286fc270f438385bee5
This add new function named "getMultiHttpClient" to the
MathRestbaseInterface.
The function creates a MultiHttpClient object using the HTTP
request factory.
This function will be useful for making HTTP requests to Mathoid.
Add test case for getUrl() method.
Bug: T335347
Change-Id: If208753edfdb3017f722c97cea7653170d05f8a5
In native MathML we only display MathML, also get
rid of enclosing <div / <span container.
Bug: T182041
Change-Id: I26bc4ea88d360d53cc7a34a89bef78812b84b2ea
* checked this with: EnWikiFormulaeTest.php
* remove finding chem macro from BaseMethods::checkAndParse, because
this will be used somewhere else
* minor: remove mi from mathFont parsing function
Bug: T321599
Change-Id: I472e3ebdebfc91f97c9391268a384f788571612e
