I am using Window 7 on a 64bit machine, Python 2.7.3, gfortran4.7.2, 32bits (from
www.equation.com/servlet), and opentelemac extracted from wingfortran32_V6p2. I am running telemac2d and found several issues:
• Similarly to many other users, to be able to run, I have to have the full path of runcode
“Python c:\opentelemac\v6p2\pytel\runcode.py telemac2d cas” instead of
“Python runcode.py telemac2d cas”
• My first trial was with Malpasset, and it would give #INF in the downstream portion of the model. I solved the problem by initializing to 0 the depth H in routine CORSUI. This is surprising because telemac2d is always very good at initializing variables (in case a compiler does not do it by default, such as the one I am using), except in this case.
• If I write on file 29, FORMATTED RESULTS FILE, it does not copy it back from the temp directory to my work directory, although the file T2DRFO is properly created in the temp. The last message is:
moving: result.slf
My work is done
without any mention of T2DRFO
• In many cases during quick initial trial runs, we run telemac, and check the results in Blue Kenue. We do not necessarily change the result file name. If a result.slf file is loaded in Blue Kenue, and we have a new run with the same result file name, it does not move the new RESULT FILE from the temp directory to the work directory.. The xxx.old file is created but the new file is not, with the message:
"uncontroled error from python:: It could not copy the output files back from the temporary directory:
c:\Temp\malpasset6.2\cas_gb_2.txt_2013-03-24-16h40min35s:
WindowsError(32, 'The process cannot access the file because it is being used by another process')"
This is something which did not happen with version 6.0 where the new run was completed and the result file created properly.
• The temp directory name, now, shows the time when it was created. This is much better than the process number. Very good idea. But the local time would be more useful than GMT time, for users outside the GMT time zone.
Thierry