Last modified: 2014-02-16 20:35:00 UTC
MathJax renders more commands than texvc, for example \lhd. This will be a problem when users with "MathJax" option insert formulas which only produce an error message for users with "PNG". One possible solution for this problem is to pass all TeX code through texvc and reject it for all modes when an error occurs. When the TeX code for MathJax is output in a way which can't be achieved by wiki syntax (bug 35479, bug 35480) this will even avoid problems like https://github.com/mathjax/MathJax/issues/43. See http://leuksman.com/mw/index.php/User:Schnark/long for an example (WARNING: This will freeze your browser for a while, but you should be able to abort. Visit at your own risk.)
The solution proposed by Nageh is to simply ignore these for now: For instance: pagecolor: ['Macro','',1] // ignore \pagecolor{} if someone has or can make lists of such commands, that would be handy. Please add here.
Created attachment 10338 [details] List of commands From source code I extracted lists of both the commands allowed in texvc and MathJax, and diffed them. I hope that the attached list is complete. I didn't test it, so there may be false positives. Pay special attention to he command \href, which allows to circumvent the spam blacklist; there may be other things that shouldn't be available on public wikis. Additionally, these environments seem to be allowed only in MathJax: align* multline multline* split gather gather* alignat* alignedat aligned gathered subarray equation equation* eqnarray eqnarray*
*** Bug 44422 has been marked as a duplicate of this bug. ***
I think the additional commands are no longer allowed in MathJax. Since (serverside) MathJax is going to be the default rendering mode, we wanted to make sure that the list of commands is well defined. As a first step we used the list specified by texvc and extracted the whitelisting of those to a seperate program called texvccheck http://www.formulasearchengine.com/texvccheck
I think this bug can be closed.
(In reply to comment #5) > I think this bug can be closed. Feel free to do so?
