continuum-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Wendy Smoak <>
Subject Re: Distributed Build Working Directories
Date Wed, 07 Oct 2009 20:30:18 GMT
On Wed, Oct 7, 2009 at 1:13 PM, George Snyder <> wrote:
> 1. With distributed builds, a project may be scheduled to build on various
> hosts (and architectures), each with its own working copy.  However, the
> "Working Copy" tab on every Build Result page in its history takes me to the
> working copy of the *latest* build.  Is that as intended?  I had expected
> the tab on each Build Result page would go to the working copy on the host
> where that build ran.

The tabs are at the project level -- info, build results, working
copy.  I don't think it was meant that there should be a working copy
per build result.  That is, when you click on the Working Copy tab, it
has nothing to do with what build result you were viewing.  I think
this is working as intended, it shows the latest in the scm repo, or
at least as of the last update/build.  (We should clarify this in the

> 2. For non-distributed builds, clicking a .tar.gz file on a Working Copy
> page downloads the file.  Clicking on a .zip file shows junk in the text
> panel, but I can then use the "Download as Text" link to download it.  For
> distributed builds, neither of these work.

See if there's an "Unable to download..." issue in JIRA already, this
sounds familiar.

> 3. Is there a way to get a standard Jetty directory listing of working
> directories, instead of the formatted Working Copy page?  I tried creating a
> symbolic link in /apps/continuum (working copies are on a different
> filesystem), but opening it responds "The URL requested cannot be found."  I
> can 'ls' and 'cd' through the link as expected.  Is there a setting
> somewhere that prohibits references outside the document tree?

I'm not sure whether that should work, but I believe there is a
feature request to allow Continuum to host project websites.  Part of
that work might be useful here (ability to define a directory that
gets exposed "as is".)  This would probably be a directory on the
master server though.

You can always do it outside of Continuum by exposing a directory
through whatever web server you're running.


View raw message