- From: lpsmith at spod-central.org (Lucian Smith)
- Subject: [cellml-dev] servererror
- Date: Wed, 5 Jan 2011 19:07:09 +0000
Thanks--the change to the API plus the change to antimony did indeed seem
to fix the problem. I'll look into the memory leak as well.
-Lucian
* Andrew Miller <ak.miller at auckland.ac.nz> [2011-01-04 19:44] writes:
>
On 04/01/11 13:13, Lucian Smith wrote:
>
>So, on Unix at least, I now have Andrew's version of my code debugged and
>
>updated to the point where the translation machinery seems to be working
>
>as well as it did before, meaning I no longer should have to use
>
>xulrunner, which is great.
>
>
>
>However, something is still going wrong with the translator program: if I
>
>ask it to translate several models at once, it translates a random subset
>
>of them, telling me it cannot parse the remaining files due to
>
>'servererror'. If I send those files individually to the translator, it
>
>works fine.
>
>
>
>Presumably, something asynchronous is going on, but I haven't a clue where
>
>to start looking.
>
>
The CellML API provides an API for asynchronous loading, but the actual
>
loading in that case needs to be done with the program, so I don't think
>
that is involved here.
>
>
I suspect something is getting corrupted, or there is uninitialised
>
memory somewhere.
>
>
I found a bug in the reference counting for math results in the CellML
>
API, and a bug in the string processing in the Antimony translation -
>
with both fixes together I don't get any errors (although
>
antimony2cellml does seem to be leaking memory) - the solution is to
>
make sure that all calls which expect the caller to release memory
>
afterwards actually do so; unfortunately, a single missing release_ref()
>
can cause the entire model representation to leak.
>
>
Patch to get antimony2cellml working for me attached.
>
>
Best wishes,
>
Andrew
>
>
> Do I need to give the CellML parser time to do
>
>something after parsing a file? Do I query someting to find out if it's
>
>relaxed enough to perform another task, and if so, will querying it too
>
>often itself stress it too much?
>
>
>
>-Lucian
>
>_______________________________________________
>
>cellml-tools-developers mailing list
>
>cellml-tools-developers at cellml.org
>
>http://lists.cellml.org/mailman/listinfo/cellml-tools-developers
>
>
_______________________________________________
>
cellml-tools-developers mailing list
>
cellml-tools-developers at cellml.org
>
http://lists.cellml.org/mailman/listinfo/cellml-tools-developers
Archive powered by MHonArc 2.6.18.