Problem with ASCIIMathMLPlugin and 2.3.0 Import Function

8 views
Skip to first unread message

jpa

unread,
Dec 27, 2007, 8:16:13 PM12/27/07
to TiddlyWiki
Hi, I'm a recent convert to TiddlyWikis and as such, I'm loving
them. But in the last couple of days, I've been baffled by a strange
behavior from the import function which I was able to trace the
presence of an active ASCIIMathMLPlugin in my TW. The problem can be
reproduced like this:

1. Take a new, empty, 2.3.0 TW and save it in the computer. Launch on
it a new tiddler with the following content:

''Dissertação possível 1''
Acaso os presentes sabem o que significa esta expressão?
{{center{
`Cos²(x)(1+Tan²(x))`}}}

The problem is not exclusive of this content, but I found it to be
sensitive to small changes of it: If I take or change portions of the
text above, it may disappear or remain, but it is still not very clear
to me, what are the exact conditions to it to exist. For example, if I
cut the string "presentes sabem o que significa", it will disappear,
but cutting "presentes sabem o" or "que significa", it may not.

2. Take a second, empty, 2.3.0 TW and save it in the computer. Then
import to it the pluggin "ASCIIMathMLPlugin", save the TW and reload
it.

ASCIIMathMLPlugin is part of a system to able someone to write math
text easily in TW's. The procedure I described is not enough to do
that but it is sufficient to denounce the problem (so, it must be
related with this tiddler specifically).

3. Try to do an import of the first TW to the second. After click on
the open button, the TW will stop after showing the message "Getting
the list of available tiddlers" and will never show that list. Instead
the 2 statements

''Dissertação possível 1''
Acaso os presentes sabem o que significa esta expressão?

...will appear rendered in the main page.

If we disable the plugin, the problem disappear and the list will be
shown. This happens at least on a XP with a FireFox 2.x .

I guess this might be more of a problem for the ASCIIMathMLPlugin
developers, but my point here is other: If the presence of certain
pluggins are enough to disrupt the import process, pluggins which
otherwise would work very well, shouldn't there be a message present
on the first import's step, stating exactly that, that it is advisable
to disable them before any import operation? Ideally, the import
process should disable all the plugins before the operation but
considering that to disable one manually, we must save and reload the
TW manually, I suppose this might not be possible (or else the same
technique would have been applied here too). I'm not seeing this
problem as an ASCIIMathMLPlugin problem alone because, this and other
plugglins might work perfectly otherwise, never to show problems, and
then fail on the import process. And if that happens on a TW full of
Plugings it might be difficult to discover the guilty one, even
because it will hard be to suspect them. And then, why they would
produce problems when in importation, but not in other common uses?

Cheers,
JPA

PS.: Sorry if this problem was already described.I didn't found it.

Eric Shulman

unread,
Dec 27, 2007, 9:12:20 PM12/27/07
to TiddlyWiki
> 2. Take a second, empty, 2.3.0 TW and save it in the computer. Then
> import to it the pluggin "ASCIIMathMLPlugin", save the TW and reload
> it.

Where are you importing ASCIIMathMLPlugin from?

-e
Eric Shulman
TiddlyTools / ELS Design Studios

jpa

unread,
Dec 28, 2007, 8:18:49 AM12/28/07
to TiddlyWiki
On 28 Dez, 02:12, Eric Shulman <elsdes...@gmail.com> wrote:
> > 2. Take a second, empty, 2.3.0 TW and save it in the computer. Then
> > import to it the pluggin "ASCIIMathMLPlugin", save the TW and reload
> > it.
>
> Where are you importing ASCIIMathMLPlugin from?

Oops, I didn't noted it was not yet published. I got this plugin
from Paulo Soares. It is something with which he is working, but not
yet published for all I know. I'm not at ease to present it here in
this context (but luckily, he will see this post and will say
something), but it is basically a bunch of "config.formatters.push"
callings, a fraction of them already presented in his (published)
plugin "LaTeXMathMLPlugin", so it is very likely that this one will
suffer from the same problem too, as also any plugin running along the
same lines. The challenge then, is to understand how the import action
may interact with similar plugins to behave as I described (In my
experiments, I became convicted the problem resided in some loose
sequence of letters... a complex regular expression appears to me to
be the obvious suspect).

Sorry for not being of better assistance...

Cheers,
JPA

Paulo Soares

unread,
Jan 3, 2008, 9:55:38 AM1/3/08
to TiddlyWiki
I believe it has nothing to do with this particular plugin but,
instead, is some misbehaviour of the importTiddlers code.

The problem occurs when one tries to import from a TW file that
includes some tiddler that invokes a macro that throws an error in the
importing TW file. You can check this by accessing
http://www.math.ist.utl.pt/~psoares/aml.html and importing from
http://www.math.ist.utl.pt/~psoares/aml_1.html.

Cheers,
Paulo Soares

Eric Shulman

unread,
Jan 3, 2008, 10:02:28 AM1/3/08
to TiddlyWiki
> I believe it has nothing to do with this particular plugin but,
> instead, is some misbehaviour of the importTiddlers code.
> The problem occurs when one tries to import from a TW file that
> includes some tiddler that invokes a macro that throws an error in the
> importing TW file. You can check this by

Does this problem also occur with

http://www.TiddlyTools.com/#ImportTiddlersPlugin

or does it only ooccur with the TW core standard import mechanism?

-e

Paulo Soares

unread,
Jan 3, 2008, 10:36:24 AM1/3/08
to TiddlyWiki
Your plugin works fine. It's just the TW core version that has some
problem.
--
Paulo Soares

Martin Budden

unread,
Jan 8, 2008, 1:23:57 PM1/8/08
to TiddlyWiki
I've reported it as ticket446 and submitted a patch.

Note that there is an underlying error in ASCIIMathMLPlugin at:

http://www.math.ist.utl.pt/~psoares/aml.html

there is an undefined function parseMath() that caused the core bug to
manifest itself.

Martin
Reply all
Reply to author
Forward
0 new messages