bitbucket: 1.3 minutes to ask password, and 8 minutes to find out it is wrong ??

rupert THURNER rupert.thurner at
Sun Jun 19 10:41:02 CDT 2011

On Sun, Jun 19, 2011 at 16:41, Mads Kiilerich <mads at> wrote:

> Andreas Tscharner wrote, On 06/19/2011 04:08 PM:
>  On 19.06.2011 15:00, Mads Kiilerich wrote:
>>> rupert.thurner wrote, On 06/19/2011 01:19 PM:
>>>> when doing an initial push to an empty bitbucket or google mercurial
>>>> repository it takes nearly 10 minutes to get an "authorization
>>>> failed". only until mercurial asks the password one has to wait for
>>>> more than a minute.
>>> Obviously it works fine for everybody else. There must be something
>>> different between your setup and other setups. Please work
>>> systematically try to figure out what that could be - perhaps by trying
>>> to reproduce on another machine on another network connection - and
>>> perhaps another OS.
>> I did that just today and I can confirm that behavior.
> Ok, as Augie hinted I wasn't completely right. Pushing over http(s) works
> fine for others, with the caveat that the posted data has to be uploaded
> twice - as described on**bts/issue1876<>. That is no problem with small pushes or high bandwidth, but the opposite
> combination is worse.
>  I moved a small repository  (~150 commits, thg transferred about 12 MiB
>> data) to BitBucket. When I pushed, the transfer started and it took a while
>> (I didn't use a stop watch, but I'd say it was a minute) before thg asked
>> for username and password. During this time the transfer was paused. After I
>> provided the information, the transfer resumed.
> How could you see that the transfer paused / resumed? Didn't you see full
> network utilization all the time?
> my impression was that "search for changes" takes especially long when the
bitbucket repository is still empty.

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <>

More information about the Mercurial mailing list