bzr log http+urllib does not work, http+pycurl is too slow

[Alexander Belchenko]

Vincent Ladeuil пишет:
>>>>>> "bialix" == Alexander Belchenko <bialix at ukr.net> writes:
> 
>     bialix> I have Trac running as separate python application on separate
>     bialix> win32 machine.
>     bialix> I put pack shared repo in the Trac's htdocs directory, so my branches are
>     bialix> available via http protocol (I use custom port 8000).
>     bialix> Well, at least `bzr revno http://host:8000/chrome/site/branches/Logic` works
>     bialix> as expected. But not log command.
> 
>     bialix> On my client machine I use custom bzr.exe 1.0rc3 compiled without pycurl, so
>     bialix> I encounter this bug.
> 
> <snip/>
> 
>     bialix> http readv of 930af4714be1e33318bb630b8d626540.pack  offsets =>
> 
> Using -Dhttp will give us the true ranges requested.

It seems that server each time respond with full text rather than 
requested range. I'll send you http.log in another mail.
Probably http transport should buffer the full text in such case to 
avoid further request for another range?

That's actually explain why it so slow. I'll try to use bzr:// 
smart-server more and more, and will test it on win32 at the same time :-)

One thing I notice and I think it's not right:
bzr log http:// tries to open working tree. why for? it's absolutely 
useless, even though don't cost too much comparing with pulling 100MB 
data 11 times.