[MERGE] readdir, take 2.

[Robert Collins]

On Tue, 2008-09-02 at 00:28 -0400, Martin Pool wrote:
> Martin Pool has voted resubmit.
> Status is now: Resubmit
> Comment:
> My comments about the docstring and copyright still apply, since they 
> crossed with this in the post.

The copyright? Oh, the fact is says Bazaar-NG? sure, I'll fix. 

> I really think you should define this to actually return the list in 
> order.  Then you can test that it actually does so, and the 
> sorting/stripping won't be pushed into every place that uses it.  Other 
> code that wants to avoid the overhead of sorting will need more pyrex, 
> but can also avoid the wrapping/unwrapping, which may(?) be comparably 
> expensive.

Well, that pushes the sort down into the pyrex; os.listdir() is also
unsorted on output. This is also a an internal function. It seems to be
tying things too closely together to do this, and neither John nor
Andrew commented on this [perhaps they missed it, or perhaps they feel
its ok].

> I'm not sure what the other pyrex tests do, but this definitely seems 
> like a case where we should be indicating that some tests were not run.

Why? if the extensions aren't built, test stipple, or a failure, won't
get them built. PQM should probably require that they built successfully
up front in the make file, not down at this level.

-Rob
-- 
GPG key available at: <http://www.robertcollins.net/keys.txt>.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: This is a digitally signed message part
Url : https://lists.ubuntu.com/archives/bazaar/attachments/20080902/52803a9d/attachment.pgp