[Kst] Re: eparse.cpp, memory leak, and lex

[Barth Netterfield]

A minor update: running flex on escan.l does return lex.yy.c which is
compatible (essentially identical - except for improvements from being
generated by flex 3.5 rather than flex 3.3) with escan.cpp, except for the
win32 changes that Peter made.

It still has the memory leak.

On Sat, Jan 15, 2011 at 3:16 PM, Barth Netterfield <
netterfield at astro.utoronto.ca> wrote:

> Hi,
>
> There is a memory leak in escan.cpp, and I'm looking for some advice:
>
> -It looks like escan.cpp was at some time generated from escan.l
> -It looks like eparse.cpp was at some time generated from eparse.y
> -These cpp files were committed by fidler in commit 726751.
> -changes to the .l and .y files were committed after this by fidler
> (731128), but it is possible the .cpp and .h files were generated from
> these.  Or maybe not.  Most of the changes are much older than this.
> -I don't think that lex/yacc have been part of the build system for a very
> long time, if ever.
> -kuemmel made some changes to the cpp files recently to allow it to build
> under windows.
>
> We are now unable to re-generate the cpp files from the .l/.y files without
> hand editing the output.
>
> 1) Is this OK and so should we just continue to modify the cpp files for
> bug fixes as they come along?  If not... ?
>
> 2) eparse.cpp and escan.cpp are is listed in the cmake files as
> kst_dont_merge.  What does kst_dont_merge do/mean?
>
> --
> C. Barth Netterfield
> University of Toronto
> 416-845-0946
>
>
>


-- 
C. Barth Netterfield
University of Toronto
416-845-0946
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.kde.org/pipermail/kst/attachments/20110115/1bc8c7a9/attachment.htm