> >> due to a *major* bug in the tinyMCE integration, which anybody who has
> >> seriously used the SAge notebook has run into. Nobody has come up
> >> with a clean test case though. My worksheet is unfortunately
> >> completely scrambled, and I'll have to spend an hour sorting it
> >> through. Jason, since you wrote the tinymce integration into sage,
> >> any idea why it would randomly scramble things?
>
> > Yes, and it's always just before class is about to start. It's SO
> > annoying.
>
> Unfortunately, I just got done teaching my last class before the
> summer break, so now it'll be even harder to debug this!
>
>
>
> >> If *anybody* has any idea how to systematically replicate this "random
> >> scrambling" of worksheet contents, please post.
>
> > I can't quite do it, though I've tried many times. BUT I do know what
> > it does, I think. I believe that somehow TinyMCE is recording the
> > inputs and eventually gets too much input if you don't save the
> > worksheet. Then (I'm pretty sure, again can't replicate) everything
> > since your last save/the last time TinyMCE could handle it is appended
> > in REVERSE order at the bottom of the worksheet - check it out! I
> > hope this helps Jason or someone else replicate it.
>
> That makes some sense. Not having looked yet, I'm guessing that the
> TinyMCE code in Sage doesn't properly verify that code was actually
> inserted by the server into the server's copy of the worksheet. E.g.,
> when you insert a new cell in the notebook, you do *not* see that new
> cell inserted until the message goes to the server "insert this cell",
> then a message comes *back* saying "new cell inserted". Try the
> following:
>
> (1) start the notebook server
> (2) make a worksheet with some cells include some tinymce cells
> (3) kill the notebook server.
> (4) notice that you get a big error when you attempt to insert a cell
> (5) try to edit a tinymce cell. Notice that everything appears to
> work fine, but in fact it can't be since the server doesn't exist.
>
> The above "proves" that if the message "this text cell has changed"
> from the web browser to the server somehow gets dropped, then one will
> silently just have edits vanish. Maybe they get sent later, which is
> why they appear elsewhere in the document.
>
> In any case, irregardless of it fixing the reording bug or not, the
> tinymce integration code should be changed so that when you save your
> changes to a cell, it does a roundtrip to the server to make damn sure
> the changes really got sent.
Here are some possibly related issues which have come up on this list
and sage-support, which I record here for convenience.
1. Very often (but maddeningly not 100% reproducible, maybe 90% for
me) the second-to-last place one can Shift-Click will not create a
TinyMCE region, no matter how often one tries.
2. Open new worksheet, put text above and below the cell (don't bother
to put anything in the Sage cell). Now save/quit, and reopen; you will
have a new (2nd) Sage cell below your second text. This is
reproducible, though I'm not sure it's a bug.
3. If you click on a text area too many times in a row before it saves
(again, not consistently reproducible, but three often does it),
perhaps before it sends to the server? then you will get a weird-
looking text box which is NOT TinyMCE, but just a box with save and
cancel buttons that do nothing; also, all text is now just plain
text. At this point I usually have to just save the worksheet, reopen
it, and hope for the best.
- kcrisman
--~--~---------~--~----~------------~-------~--~----~
To post to this group, send email to sage-devel@googlegroups.com
To unsubscribe from this group, send email to
sage-devel-unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URLs: http://www.sagemath.org
-~----------~----~----~----~------~----~------~--~---
