Ignore encoding errors when opening scb files

[dfeeley]

Hi, thanks for creating this, its been very useful.

This pull request handles a case I encountered with windows-1252 encoded swimmer names in the .scb files causing the process to fail. Since we don't actually care about the swimmer names we can just ignore encoding errors when opening the scb file.

[JohnStrunk]

Thanks for the patch!
Could you send me an example scb that has the problem? I've been working on https://github.com/JohnStrunk/wahoo-results which has the same scb->dolphin functionality, and I'd like to have a test case that I can add to the parser.

[JohnStrunk]

@dfeeley What OS are you using?

After testing out that example file and digging around a bit, I think this is a platform-specific issue.
On Windows (or my machine, anyway), the default encoding is cp1252 based on the output of locale.getpreferredencoding(False). I've updated wahoo-results to always specify 1252 even though it's not going to run anywhere but Windows.

My thought is that the fix for this issue should be encoding="cp1252" instead of utf-8. If you get a chance, could you give it a try?

[dfeeley]

Hi John, for wahoo-results (where you care about names / not dropping characters), I agree cp1252 makes more sense than utf-8. I'm linux only myself, which is neither here nor there for this discussion ... I have the use of a swim club windows 10 computer for Dolphin / Meet Manager. On *that* computer I confirmed that encoding="cp1252" works fine. Interestingly, sys.getdefaultencoding() reports 'utf-8' for me on Windows 10, which I think tallies with something I read about Windows eventually moving away from the legacy encodings towards utf. For now looks like Meet Manager is cp1252 so thats the best encoding to use. Maybe overkill, but for future-proofing you could always try cp1252 first and then fall back to utf-8, e.g.: for enc in ('cp1252', 'utf-8'):

try: scb = open(path, "r", encoding=enc, errors='strict') lines = scb.readlines() break except UnicodeDecodeError: pass

thanks Dave

…

On Fri, 2 Jul 2021 at 15:19, John Strunk ***@***.***> wrote: @dfeeley <https://github.com/dfeeley> What OS are you using? After testing out that example file and digging around a bit, I think this is a platform-specific issue. On Windows (or my machine, anyway), the default encoding is cp1252 based on the output of locale.getpreferredencoding(False). I've updated wahoo-results to always specify 1252 even though it's not going to run anywhere but Windows. My thought is that the fix for this issue should be encoding="cp1252" instead of utf-8. If you get a chance, could you give it a try? — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#1 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAOPICC52NSGO4SMMDOZHVDTVY3NJANCNFSM47VAA76Q> .