Hi Antonio, Thanks for the quick response. > Hi Stuart, > > Il giorno 21/ago/2011, alle ore 02.38, Stuart Mentzer ha scritto: > >> Hello, >> >> Some kinds of HDF5-DIAG errors don't generate any exceptions making it >> difficult to do proper error handling. For example: >> >> HDF5-DIAG: Error detected in HDF5 (1.8.5) thread 0: >> #000: ..\..\..\src\H5Dio.c line 174 in H5Dread(): can't read data >> major: Dataset >> minor: Read failed >> >> doesn't raise an HDF5ExtError or any other exception. Is this something that >> can be fixed in PyTables? > I just performed a quick check on all calls to H5Dread and it seems to me > that in all cases there is a check for failures with relative HDF5ExtError > exception raising. With your encouragement that the intent is to raise all errors as HDF5ExtError I revisited the code in question and found the well-hidden place where the developer had masked exceptions. After fixing that I do find that it is indeed raising HDF5ExtError. > > Can you please be a little more specific on how to reproduce the issue? > I encourage you to open an issue on https://github.com/PyTables/PyTables > >> Also, can the HDF5-DIAG messages be suppressed? I know that the HDF5 API has >> a dontPrint call but I don't see a PyTables interface to that. If we can't >> detect these exceptions I at least want to mask them from end users. >> >> Thanks! > > Yes "dontPrint" is part of the C++ API, the C one has "H5Eset_auto" to should > perform a similar task. > Well I suppose we could provide a function to suppress HDF5 messages but IMHO > it is more important to fix the issue regarding exception. With the proper error handling using HDF5ExtError I am able to stop the code quickly when the first problem arises but I still get about a page of HDF error messages. It would be great to be able to turn that off for a production release, so I would vote for adding a dontPrint call. > regards > > -- > Antonio Valentino Thanks again, Stuart
------------------------------------------------------------------------------ uberSVN's rich system and user administration capabilities and model configuration take the hassle out of deploying and managing Subversion and the tools developers use with it. Learn more about uberSVN and get a free download at: http://p.sf.net/sfu/wandisco-dev2dev _______________________________________________ Pytables-users mailing list Pytables-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/pytables-users
