clone from URL

Paul Boddie paul.boddie at
Tue Jun 14 07:41:31 CDT 2011

[Re-ordered quotes]

On 14/06/11 14:30, J.S. van Bethlehem wrote:
> On 06/14/2011 02:12 PM, norricorp wrote:
>> Hi,
>> still very new to mercurial but less new than last week.
>> I have set up Apache on red hat to serve a mercurial repository. This
>> works.
>> I can use the browser on various PCs and http://jazz/hg/p1 displays my
>> project1 project. I require authentication and this works. I can also use
>> tortoiseplink to run ssh comands including hg commands from the PC to the
>> linux box.

SSH commands and HTTP are not the same thing, however. It's quite 
possible for SSH-based operations to work and for HTTP-based operations 
to fail.

>> If I use tortoiseHG or the command line on a PC to clone a repository
>> C:\mercurial>hg clone http://jazz/hg/p1 .\clone_p1
>> http authorization required
>> realm: Mercurial repositories
>> user: xxxxxxx
>> password:
>> abort: HTTP Error 404: Not Found
>> Any idea what I am doing wrong? The passwords for logging on to linux and
>> htpasswd are the same.

Again, Linux passwords and htpasswd credentials are different, unless 
you're using some Apache authentication module which integrates with the 
system's authentication mechanisms (which isn't really htpasswd). 
However, this isn't likely to be the problem here.

> Have you checked the settings in the [web]-section. I'm not sure what
> the default settings are, but maybe you need to explicitly set
> 'allowpull = *' or something along those lines?
> See:

Cloning should more or less work without any extra configuration. In 
this case, I'd be more suspicious about whether hgweb is really being 
involved correctly in the transaction.

Here's the checklist:

1. What does the log file say? Which resource is serving the repository? 
Which resource is accessed when trying to clone?
2. What does the repository view look like in a Web browser? Like hgweb 
or like an Apache-generated directory listing?


More information about the Mercurial mailing list