On 03/18/2013 07:34 AM, Roger wrote:
> No problem. I just wanted to let you know for your release candidate.
So, it seems like this was the problem...
All my .c files include things "sufficiently" - that is, they included
frame.h ahead of fooframe.h when that was needed.
But the process of creating the python bindings (ctypesgen) only looks
at the headers - so if fooframe.h should have included frame.h but
didn't, the C code in fooframe.c would compile, but the ctypesgen
process might or might not work - depending on what order it processed
the header files.
I believe this new script may process them in directory order - so it
might work for one person and not another.
I should create a "lint" process that processes them individually and
makes sure they are all going to compile individually - which is only
good practice anyway - and I should probably sort the header files so
the build process is more repeatable from person to person.
This is especially true since it turns out that ctypesgen doesn't
complain when you have duplicate conflicting definitions of the same
symbol. This can be bad news if you have a file named
SAVEoldcruftyframe.h which is an obsolete version of frame.h. Very
confusing (the voice of experience speaking here).
--
Alan Robertson <[email protected]> - @OSSAlanR
"Openness is the foundation and preservative of friendship... Let me claim
from you at all times your undisguised opinions." - William Wilberforce
_______________________________________________
Assimilation mailing list - Discovery-Driven Monitoring
[email protected]
http://lists.community.tummy.com/cgi-bin/mailman/listinfo/assimilation
http://assimmon.org/
