Message41845
Logged In: YES
user_id=113328
I've now taken Neal's patch and integrated my changes. I'll
check it through again, and run a few more tests, and then
upload the result (I've had problems with cvs update, so I'm
downloading a new copy of CVS python first).
Most things look OK. There are a couple of issues with case-
insensitive filesystems.
The first is that the cache of os.listdir results is (case-
sensitively) keyed on directory name. I don't see this as an
issue as (a) I'd expect the case used to remain constant
(who's going to change a directory in sys.path at runtime to
the same name with a different case???) and (b) it's only an
efficiency hit, not an error, in any case.
The second issue is more problematic. Zip files on sys.path
are recognised (see add_zip_names()) by the extension ".zip"
(case sensitively). It's arguable that on case-insensitive
filesystems, this check should also be case insensitive. But
finding that a file is on a case-insensitive filesystem is, I
believe, not possible. And using a case-insensitive test all the
time isn't right either. I'd say that we should just *define* the
algorithm as only treating files ending in ".zip" (case
sensitive) as zip files. It doesn't seem to be a problem in
practice.
I'll try to find some way to fit that into the documentation...
Apart from this issue, I've seen no other problems with the
patch. |
|
Date |
User |
Action |
Args |
2007-08-23 15:18:40 | admin | link | issue645650 messages |
2007-08-23 15:18:40 | admin | create | |
|