MathJax Test Results - 2013-05-02

13 views
Skip to first unread message

Frédéric WANG

unread,
May 2, 2013, 9:07:28 AM5/2/13
to mathj...@googlegroups.com
So I've looked more carefully about IE10's modes and there are now six
document modes:

1) Standards
2) Quirks
3) IE9
4) IE8
5) IE7
6) IE5 Quirks

It seems that modes 3-6 are the former document modes that were
available in IE9. When I switch to mode 1) or 2), I get the alert box
about MathJax not able to setup MathPlayer (even when I uninstall
MathPlayer!) and then the page renders normally. This unexpected alert
box prevents the testing framework to continue. That's why I couldn't
execute the Quirks mode in IE10, although I can probably execute 6) if I
modify the Python's scripts. IE10's Quirks mode seems a bit more
standard compliant, for example it passes
LaTeXToMathML/AMScd/amscd-1.html while IE7 and IE5 Quirks fails that
test at the moment.

I've executed the tests for IE9-SVG and IE8 again, as it looks like the
bad mode was selected yesterday (for example
LaTeXToMathML/AMScd/amscd-1.html was supposed to pass). As I said, the
document mode is done via keyboard simulation and is not always
reliable, but Selenium does not have any API to do that otherwise. This
keyboard simulation still seems to work in IE10, though. I don't know if
the IE7 mode was correctly selected, but I don't think there will be
much difference anyway.

http://mathjaxtest.s3.amazonaws.com/v2.1-candidate/2013-05-02/Windows_MSIE_IE8_TeX_HTML-CSS-1.html.gz
http://mathjaxtest.s3.amazonaws.com/v2.1-candidate/2013-05-02/Windows_MSIE_IE9_TeX_SVG.html.gz

--
Fr�d�ric Wang
maths-informatique-jeux.com/blog/frederic

Davide P. Cervone

unread,
May 3, 2013, 10:42:51 AM5/3/13
to mathj...@googlegroups.com
Thanks for the new tests. I've gone over them and found one important bug (which I fixed) and some zoom issues in IE (most of which I fixed). The rest of the issues are not critical (some are already fixed after you ran the tests, some are the standard pixel problems, and most of the rest are ones that run for me by hand -- may be an IE9 versus IE10 issue?)

In any case, I think we are ready for the beta release. I don't see anything else that needs to be fixed before that.

Davide
> Frédéric Wang
> maths-informatique-jeux.com/blog/frederic
>
> --
> You received this message because you are subscribed to the Google Groups "MathJax Development" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to mathjax-dev...@googlegroups.com.
> For more options, visit https://groups.google.com/groups/opt_out.
>
>

Frédéric WANG

unread,
May 3, 2013, 11:14:33 AM5/3/13
to mathj...@googlegroups.com
OK great. Was the "important bug" the problem with parameter
substitution in localization or is it another bug?

On 03/05/2013 16:42, Davide P. Cervone wrote:
> Thanks for the new tests. I've gone over them and found one important bug (which I fixed) and some zoom issues in IE (most of which I fixed). The rest of the issues are not critical (some are already fixed after you ran the tests, some are the standard pixel problems, and most of the rest are ones that run for me by hand -- may be an IE9 versus IE10 issue?)
>
> In any case, I think we are ready for the beta release. I don't see anything else that needs to be fixed before that.
>
> Davide
>

Davide P. Cervone

unread,
May 3, 2013, 11:28:46 AM5/3/13
to mathj...@googlegroups.com
No, I fixed that one Wednesday. This was due to the change in issue #373 that caused a regression in other tests. This was a change to the Get() function that looks up MathML attributes with inheritance (one of the core routines) in order to fix the \mathchoice problem, so I was a little worried that it might have unforeseen side-effects. It did, bit I have taken care of it. I removed the change that I was worried about, and it turns out not to have been necessary for issue 373 anyway, so all should be well.

Davide

Frédéric WANG

unread,
May 3, 2013, 11:31:46 AM5/3/13
to mathj...@googlegroups.com
On 03/05/2013 17:28, Davide P. Cervone wrote:
> No, I fixed that one Wednesday. This was due to the change in issue #373 that caused a regression in other tests. This was a change to the Get() function that looks up MathML attributes with inheritance (one of the core routines) in order to fix the \mathchoice problem, so I was a little worried that it might have unforeseen side-effects. It did, bit I have taken care of it. I removed the change that I was worried about, and it turns out not to have been necessary for issue 373 anyway, so all should be well.
>
> Davide
>
OK, so I think we have eliminated the major issues now. I still need to
more testing about the safe mode, localization and on other potential
fixes after the beta, but I guess we are ready to for the beta release.
It looks like issue 301 won't make it for the beta, though.

Davide P. Cervone

unread,
May 3, 2013, 11:47:09 AM5/3/13
to mathj...@googlegroups.com
I know; I'm sorry about that. I think it makes sense to include that with the other font-based changes like STIX-Web and the other ones, so we can do that fairly quickly after this release, I would think. I hope that is OK.

Davide

Frédéric WANG

unread,
May 3, 2013, 11:53:22 AM5/3/13
to mathj...@googlegroups.com
OK, no problem. My only concern was about MathJax and Wikipedia. If some
people start complaining about the fact that MathJax is slow on
Wikipedia, we could at least have suggested to switch to the native
MathML on Firefox. Without the change, we must additionally say that the
MathJax fonts should be installed, and that's a shame. However, there is
still the issue with horizontal braces anyway and it makes sense to put
everything in a font release.

On 03/05/2013 17:47, Davide P. Cervone wrote:
> I know; I'm sorry about that. I think it makes sense to include that with the other font-based changes like STIX-Web and the other ones, so we can do that fairly quickly after this release, I would think. I hope that is OK.
>
> Davide
>

Reply all
Reply to author
Forward
0 new messages