Chronological Thread 

< Chronological >      < Thread >

---

On 08/04/11 10:24, Lucian Smith wrote:
> OK, I put CellML things on hold a bit, but am now getting back to them. I
> still have this question about Cevas:
>
>> So, in the past, I would report cevas errors as erros in reading in the
>> model, and assume something drastically wrong had happened, and exit.
>> However, in the new code, cevas errors include what I would call
>> validation errors:
>>
>> CellML Variable Emax in component EAtrium has in interfaces, but is not
>> connected (directly or indirectly) to any variable with no in interfaces;
>>
>> Are these 'proper' Cevas errors, and if so, are there different cevas
>> errors that mean I would not actually be able to parse the model? And if
>> so, how do I distinguish between the two? My current code just checks to
>> see if 'cevas->modelError' is an empty string or not.

Hi Lucian,

Find a variable with no 'in' interfaces is an essential part of what
CeVAS does, since it treats that variable as the 'primary' variable,
with all other variables connected to it (directly or indirectly) as
secondary. You can't use CeVAS with a model that has variables with no
'in' interfaces.

>
> Also, how are the binaries coming along, and/or the build system for
> Windows? Are they working yet?

The Windows build system is still waiting on our IT department to
provision the server; a release candidate including binaries, with
Windows binaries built on a virtual machine on my dev machine as an
interim measure is planned as soon as we can get all the tests to pass.

You can see the build status at:
http://autotest.bioeng.auckland.ac.nz/cellml-build/waterfall
The latest Windows problems are because it had Python support turned on
- I'm planning to turn off Python support on Windows for this release
(since all other binaries are built with MSVC, and I haven't been able
to find suitable pre-built Python binaries), which might mean the build
succeeds when I trigger it manually.

Automatic binary packaging also still needs some work, but I think I am
close to being able to put out a release candidate.

Best wishes,
Andrew

---

• [cellml-dev] Cevas errors, Andrew Miller, 04/08/2011

---

Archive powered by MHonArc 2.6.18.