[petsc-dev] SuperLU link error on BGP FEN

[Matthew Knepley]

On Thu, Dec 2, 2010 at 4:14 PM, Jed Brown <jed at 59a2.org> wrote:

> On Thu, Dec 2, 2010 at 23:06, Matthew Knepley <knepley at gmail.com> wrote:
>
>> The problem with this is that we automatically determine PETSC_ARCH in the
>> course of configure, well
>> after we start logging. We do symlink it at the end.
>>
>
> Yes, but the file gets truncated by the configure process that fires up
> afterward.  You could
>

How exactly does that happen? I have just checked that I ran multiple
configures and did not clobber those logs.


> create it as a temporary file in $PETSC_DIR (mktemp)
> symlink configure.log to this temp file
> wait until you know PETSC_ARCH
> hardlink $PETSC_ARCH/conf/configure.log to the temp file
> redirect the symlink to $PETSC_ARCH/conf/configure.log
> remove the temp file
>

I do not think this works on NTFS.

   Matt


> This process should be atomic and always have a usable state even with
> concurrent configures and unpredictable failures (though I suppose the help
> message should ask the user to send the real file instead of the symlink
> which could point elsewhere).
>
> Jed
>



-- 
What most experimenters take for granted before they begin their experiments
is infinitely more interesting than any results to which their experiments
lead.
-- Norbert Wiener
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/petsc-dev/attachments/20101202/1df2d652/attachment.html>