On Mar 24, 2010, at 12:40 AM, Stefan Behnel wrote: > Dag Sverre Seljebotn, 24.03.2010 08:20: >> Stefan Behnel wrote: >>> I only noticed recently that the codecs module doesn't support >>> universal >>> newline parsing. This means that parsing source files with DOS >>> line endings >>> will currently lead to this error: >>> >>> C:\Python31\hello.pyx:1:23: Unrecognized character >>> >> >> Totally amazing that we haven't got complaints about this before from >> Windows users. > > Same from here.
Given we haven't (even now, once you've brought it up) it seems it isn't a big issue (though should still be fixed). >> Is this something that's been there all the time or the >> result of a recent change? > > Ever since I implemented source encoding support, and that was long > ago, > certainly before 0.12 came out. > > The funny thing is that I only became aware of this due to a test > failure > in Hudson on the msvc docstrings test, which had been added with DOS > line > endings. I was puzzled enough to see this test fail on a Linux > machine, > while I can't reproduce this on my (Linux-)laptop at all. So I can > only > guess that this depends more on the Python build or some other > environmental factor. I didn't ever see any failures on my box either. (FYI, OS X apps typically default to (or at least handle seamlessly) Unix line endings, at least the ones I ever use.) It > Anyway, the error message above was taken from a recent post on > c.l.py, > pointing at the end of a totally innocent looking source line (first > in the > file), so for now I assume that this really impacts at least some > users. Can anyone actually reproduce this on Windows. The example in question was provided by a Windows user who clearly had no issues with it. If you want to do an emergency release, feel free to do so, but I don't see any evidence that it's necessary. - Robert _______________________________________________ Cython-dev mailing list [email protected] http://codespeak.net/mailman/listinfo/cython-dev
