I started reading François's patches in ticket #534. (I also read [1] and [2], but not yet [3] or [4].) I would love it if we could get those unicode fixes into v1.5, and François's approach of implementing only the easy part that we know that we'll want is a good approach, but so far I think it is insufficiently unit-tested -- the first patch [5] and the second [6] don't come with unit tests for the code changed or added in the patch.
Also it looks like, in patch [5], that it is designed for Tahoe to get filenames through the Python byte-oriented APIs (i.e. "os.listdir(bytestring)") and then pass them through fs_to_unicode(), which can switch as necessary by platform or encoding. I don't think this will work; if you get filenames using the byte-oriented APIs on Windows then you'll get something which (maybe!?) can't be properly decoded into unicode. Instead we need to wrap "os.listdir()" with something which switches on platform and uses the unicode-oriented os.listdir() on Windows and Mac and the byte-oriented os.listdir() on other. Regards, Zooko [1] http://allmydata.org/pipermail/tahoe-dev/2009-May/001716.html [2] http://allmydata.org/pipermail/tahoe-dev/2009-May/001718.html [3] http://allmydata.org/pipermail/tahoe-dev/2009-May/001721.html [4] http://allmydata.org/pipermail/tahoe-dev/2009-May/001728.html [5] http://allmydata.org/trac/tahoe/attachment/ticket/534/%20plumbing%20for%20unicode%20support.darcspatch [6] http://allmydata.org/trac/tahoe/attachment/ticket/534/%20tahoe%20manifest%20unicode%20support.darcspatch _______________________________________________ tahoe-dev mailing list [email protected] http://allmydata.org/cgi-bin/mailman/listinfo/tahoe-dev
