Le mercredi 31 mars 2010 15:02:25, David Baelde a écrit :
> Hi,
Hi,
> I checked the code, and I have to apologize: from the beginning I was
> talking with a wrong idea in mind. Currently, destroying a requests
> pops all indicators, which can trigger the deletion of several
> temporary files.
>
> What I had in mind is to always have at most one file (and several
> "purely virtual" indicators before). It might be more complex, but has
> two advantages: (1) This would avoid allocating several temporary
> files for a long time. (2) It's in that mindset that I was talking of
> telling the post-processor whether the current file is temporary or
> not, so it can possibly directly change an already temporary file.
I have been looking at the possible modifications of my patch.
Honnestly, I believe this starts to become more than simply adding post-
processors.
If you want to add the possibility to pass a temporary flag to the post-
processor so that he can not touch the file or modify it in place, then you
have a redundancy issue with the "temporary" operator of the lang_builtin
function.
Also, it would be cool to be able to return a single indicator and have
indicator option as the return of a postprocessor. The problem being that this
is not possible in the script language..
Eventually, the problem of having a single real indicator on top is orthogonal
to my patch since it also concerns the file resolvers..
I would advice to commit a patch about the post-processor and look at the
issue of the virtual/real indicator lately in order to seperate the different
issues (and also allow me to not maintain a patch that is getting bigger and
bigger..)
Romain
------------------------------------------------------------------------------
Download Intel® Parallel Studio Eval
Try the new software tools for yourself. Speed compiling, find bugs
proactively, and fine-tune applications for parallel performance.
See why Intel Parallel Studio got high marks during beta.
http://p.sf.net/sfu/intel-sw-dev
_______________________________________________
Savonet-devl mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/savonet-devl
