mercurial@4165: 28 new changesets

[Christian Ebert]

* Thomas Arendsen Hein on Saturday, March 10, 2007 at 10:46:39 +0100:
> * Christian Ebert <blacktrash at gmx.net> [20070309 13:52]:
>> * Mercurial Commits on Friday, March 09, 2007 at 00:15:11 +0100:
>>> http://www.selenic.com/hg/rev/5c1e18bb804c
>>> changeset:   4164:5c1e18bb804c
>>> parent:      4162:b2d9e553cdc8
>>> user:        Michael Gebetsroither <michael.geb at gmx.at>
>>> date:        Thu Mar 08 22:36:12 2007 +0100
>>> summary:     hgweb: use the given revision in the name of the archive
>> 
>> I don't understand this one.
>> 
>> I got reponame-shorthash.tar.bz2 (where shorthash is hash of the
>> corresponding revision) already without the patch.
> 
> The patch should allow requesting reponame-1.2.3.tar.bz2 instead of
> reponame-shorthash.tar.bz2

But this doesn't work for me via a browser, I still get the hash.

What works is eg:

$ wget http://localhost/cgi-bin/hgwebdir.cgi/reponame/archive/1.2.3.tar.bz2

But then the file I requeste is, well, 1.2.3.tar.bz2 and not
reponame-1.2.3.tar.bz2

Whereas if I do:

wget http://localhost/cgi-bin/hgwebdir.cgi/reponame-1.2.3.tar.bz2

I retrieve the right filename, but it contains the web page.

>> Now, for tip I get reponame-tip.tar.bz2, which imho is not very
>> informative.
> 
> Yes, here the hash is really the only useful thing if the user
> didn't specify a tag which was manually created.

One could do something like (hgweb_mod.py lines 613,614):

    if cnode == id or id == 'tip':
        arch_version = short(cnode)

But I still don't get it working for named tags.

Probably I misunderstand the whole purpose of it.

c
-- 
_B A U S T E L L E N_ lesen! --->> <http://www.blacktrash.org/baustellen.html>