Commit Graph

2178 Commits

Author SHA1 Message Date
469893cd4a * The "latest" channel shouldn't contain inactive jobs. 2009-04-25 11:41:46 +00:00
000fffeb11 2009-04-25 11:27:46 +00:00
3c47a11bd9 * Don't disable fsync in production. 2009-04-25 07:48:30 +00:00
109cc35edf * Randomly permute the order in which builds are added. This is
mainly to prevent all those Nixpkgs builds named "kde*" from
  building at the same time.  Since they all have the same slow
  dependencies (qt, kdelibs) this tends to block the buildfarm.
2009-04-23 15:40:36 +00:00
c48ec3d340 * Autoflush stdout. 2009-04-22 22:59:54 +00:00
97a6011628 * Hack around those SQLite timeouts: just retry the transaction. 2009-04-22 22:43:04 +00:00
80691a39f5 * Don't log redundant build steps in case of cached failures. 2009-04-22 14:41:12 +00:00
018585dba8 * In the job status page and the channels, pick the build with the
highest ID rather than the highest timestamp.  Otherwise, if a build
  from revision N finishes after a build from revision N + 1, then
  the build from revision N will end up in the channel.  Thus, the
  channel contents will be out of sync.

  This is still not quite correct: if a revision *reverts* to an older
  build, the channel will still end up out of sync, because Hydra
  won't schedule the build again (after all, it has already done it).
  A better fix would be to add a separate timestamp denoting when the
  build was last "current" (i.e. corresponding to the "head revision"
  of its job).
2009-04-22 13:55:20 +00:00
4d699c97fd * Don't do pretty printing for large logs, because the XSLT processing
is rather slow.
2009-04-16 15:22:14 +00:00
d6bd9bab5f * Job status / error pages: show only active jobs. 2009-04-15 14:50:15 +00:00
2a05522189 * Don't index any of the logs (mostly because xsltproc can be a huge
resource hog).
2009-04-14 12:40:00 +00:00
28fde34b9c 2009-04-09 15:09:00 +00:00
10ae24fd3c * Fixed a broken redirect when deleting a release set. 2009-04-09 11:43:49 +00:00
607d5641b6 * Added a page to quickly see all failed builds and failed evaluations
per project or jobset or job or globally.
2009-04-08 22:08:00 +00:00
bf2a5a2629 * Allow omitting the product number if there is only one product in a
build.  This allows URLs like

    http://server/job/patchelf/trunk/deb_debian40i386/latest/download

  or

    http://server/release/patchelf/unstable/latest/deb_debian40i386/download
2009-04-08 09:37:05 +00:00
11d8421def * Support redirecting to some job of a release. For instance, this
redirects to the "tarball" build of the latest release of
  patchelf/unstable:

    http://server/release/patchelf/unstable/latest/tarball

  In conjunction with the other redirects this allows linking to the
  actual download of the build:

    http://server/release/patchelf/unstable/latest/tarball/download-by-type/file/source-dist
2009-04-08 09:03:45 +00:00
26ff5f1ff0 * Provide some redirects to build products by type so that we can for
example link to the latest tarball through a URI like:

    http://server/job/patchelf/trunk/tarball/latest/download-by-type/file/source-dist
2009-04-08 08:09:39 +00:00
f9572e4d1e * Provide some redirects to the latest successful build for a job (or
project, or jobset, or globally, but that's not all that useful).
  This is useful if you want to link to the latest build.
2009-04-07 15:49:02 +00:00
16a84f4bf5 * Big speed-up of the job status page and the channel generation (such
as the manifest).  The builds are now determined in one SQL query
  rather than a zillion ones.
2009-04-03 15:37:21 +00:00
ca11aba074 * Fix an infinite recursion. 2009-04-03 09:13:34 +00:00
0ae979f652 2009-04-03 09:12:36 +00:00
09944ed3a2 * Forgot to commit. 2009-04-03 08:38:13 +00:00
753e56b6eb * Improved the navigation bar: don't include all projects (since that
doesn't scale), and include links for jobset/job specific pages.
  The main page now lists the projects.
* Overview pages for jobsets and jobs.
* Links to the channels.
* Jobsets are now defined and edited in a separate action.
2009-04-02 16:15:57 +00:00
db4ce0df06 2009-03-31 16:09:04 +00:00
c970fc8335 * Urgh. 2009-03-31 15:59:47 +00:00
c39a693afd * Generate SVG instead of PNG, which is much faster and smaller. I
tried this before but the text didn't fit in the boxes in Firefox.
  The solution is to use Graphviz' svg:cairo backend instead of svg.
  svg:cairo doesn't depend on client-side fonts.
2009-03-31 15:59:31 +00:00
a6c32b33db 2009-03-31 14:55:47 +00:00
156f40130e * Generate robots.txt in a more Catalystic way. 2009-03-31 14:14:45 +00:00
d6e996d01c * Generate a robots.txt. 2009-03-31 13:48:03 +00:00
7b7c3f2265 * Perl sucks. "getBuildLog $drvPath" doesn't mean the same as
"getBuildLog($drvPath)" if you call it in a hash, and quietly screws
  up the rest of the hash.
2009-03-26 15:32:19 +00:00
f0f9f12e42 * Use Nix's negative caching. 2009-03-26 12:53:39 +00:00
8f6f6ee53e * Allow authorised users to create projects. 2009-03-24 14:22:59 +00:00
c75977ce5a * Record the input containing the Nix expression (release.nix) in the
build inputs.  Otherwise we can't (for example) reproduce a build
  later.
2009-03-23 21:42:59 +00:00
dab25b30f0 * Fix the queue page. 2009-03-23 15:19:34 +00:00
12c1d90135 * Add some DB indices to make the /releases page much faster.
* Reduce the number of DB queries for the /all and /jobstatus pages.
* Show the Hydra version number.
2009-03-23 13:52:24 +00:00
0da0384753 * Top-level: don't exit on errors. 2009-03-23 01:13:37 +00:00
d9424b6364 * Show the input bindings in evaluation error messages. 2009-03-20 17:06:50 +00:00
3d26bb7fd1 * Allow `-' in release set names. 2009-03-20 15:19:17 +00:00
2755c895ff * In job inputs of type "build", allow the project and jobset names of
the input build to be specified, as well as constraints on the
  inputs of the inputs build.  For instance, you can require that a
  build has input `system = "i686-linux"'.

  This is important when one binary build serves as an input to
  another binary build.  Obviously, we shouldn't pass a build on
  i686-linux as an input to another on i686-darwin.  Hence the
  necessity for constraint.

  The constraint are currently quite limited.  What you really want to
  say is that the "system" input of the other build has to match the
  "system" input of this build.  But those require a bit more work
  since they introduce dependencies between inputs.
2009-03-20 14:50:09 +00:00
d1affbfe1a * Oops. 2009-03-20 10:26:50 +00:00
e076c0cff4 * Catalyst::View::Download::Plain doesn't like empty files. 2009-03-20 10:26:26 +00:00
d13af4f392 * Doh. 2009-03-20 10:19:46 +00:00
163f3de2fa * Urgh. `-' had special significance in HTML form parameters... 2009-03-19 15:41:33 +00:00
af1e60a043 * Allow `-' in project/jobset names. 2009-03-19 15:12:50 +00:00
ba7b94a852 2009-03-18 18:50:42 +00:00
9e4b029285 * A quick hack to list the contents of various types of files (RPM,
Debs, tars, ...).
2009-03-18 17:40:12 +00:00
b39e2c5e32 * Doh. 2009-03-18 17:13:13 +00:00
702da969a6 * Ordering by timestamp isn't a good idea here since a newer revision
might finish building before an older revision, and therefore have a
  lower timestamp.
2009-03-18 16:46:01 +00:00
a1848b08f2 2009-03-16 17:51:42 +00:00
bf1480cc25 * Doh. 2009-03-16 17:46:46 +00:00