Commit graph

45 commits

Author SHA1 Message Date
Eelco Dolstra ba0c531f63 * Remove an empty file. 2010-03-10 15:20:06 +00:00
Rob Vermaas 2c132a2748 * hydra: project members can do stuff that the owner can do 2010-03-10 10:02:25 +00:00
Eelco Dolstra 7daca03e78 * Store jobset evaluations in the database explicitly. This includes
recording the builds that are part of a jobset evaluation.  We need
  this to be able to answer queries such as "return the latest NixOS
  ISO for which the installation test succeeded".  This wasn't previously
  possible because the database didn't record which builds of (say)
  the `isoMinimal' job and the `tests.installer.simple' job came from
  the same evaluation of the nixos:trunk jobset.

  Keeping a record of evaluations is also useful for logging purposes.
2010-03-05 15:41:10 +00:00
Eelco Dolstra 79a1489c95 * Remove the triggers to simulate foreign key constraints on SQLite,
as SQLite now supports them natively (though we still need to turn
  them on in Hydra).
2010-03-05 10:33:36 +00:00
Rob Vermaas 7208f79828 inital version of links to diff in scm 2010-02-25 15:32:56 +00:00
Rob Vermaas 87ea0cb2b9 fix wrong comment 2010-02-25 09:22:00 +00:00
Eelco Dolstra d8cc0bbb5d * Make the "latest succeeded" query (used by the "latest" channel)
faster, from about 4.5s to 1.0s for the global "latest" channel.
  Note that the query is only fast if the "IndexBuildsOnJob" and
  "IndexBuildsOnJobAndIsCurrent" indices are dropped - if they exist,
  PostgreSQL will use those instead of the more efficient
  "IndexBuildsOnJobFinishedId" index.  Looks like a bug in the planner
  to me...
2010-02-12 14:49:32 +00:00
Rob Vermaas 68c60b4c66 * hydra: added index, actual build time (buildstep with same outpath, so without deps) of the build 2010-02-11 12:23:46 +00:00
Eelco Dolstra 14a1e6db61 * More indices. 2010-02-11 10:42:37 +00:00
Eelco Dolstra 59e4f65298 * Speed up the jobset index page. Especially the query to get the
inactive jobs was quite slow.
* "IndexBy" -> "IndexOn".
2010-02-09 13:47:20 +00:00
Eelco Dolstra 9f9e18e3a6 2010-01-26 20:27:33 +00:00
Eelco Dolstra 268364bfd8 2010-01-26 20:19:25 +00:00
Eelco Dolstra 0c5f8dfb47 * Another essential index. 2010-01-26 20:03:20 +00:00
Eelco Dolstra f468caf192 2010-01-19 17:59:13 +00:00
Rob Vermaas 8a01999220 hydra
* remove trailing spaces from email notification
 * option to disable email notification for jobset
2010-01-06 13:07:59 +00:00
Rob Vermaas 044edfb764 * email notification of evaluation errors to project owner (if desired) 2009-12-18 12:07:45 +00:00
Rob Vermaas 06dc6d8f86 * for git inputs, check latest revision of branch (defaults to master for now), if there is change, only use input if last checkout was > hour ago. 2009-11-19 08:15:49 +00:00
Rob Vermaas 40920935fc remove revision from cached cvs inputs, and added missing files 2009-11-17 15:24:51 +00:00
Rob Vermaas 2fb05b34bf add support for git as jobinput 2009-11-17 15:16:41 +00:00
Eelco Dolstra 7eda090e74 * Prevent repeated evaluation of a jobset with the same inputs. This
should make the Hydra scheduler a lot less CPU-intensive, since it
  won't run hydra_eval_jobs all the time.
2009-11-17 13:55:22 +00:00
Eelco Dolstra cb2493eca9 * Store the jobset's nixExprPath and nixExprInput fields in a build to
allow it to be cloned (re-executed with modified inputs) later and
  to provide some traceability.
2009-10-26 13:33:48 +00:00
Eelco Dolstra 929cbe7b7c * Adding persistant releases. A release is a named set of builds. 2009-10-21 15:44:17 +00:00
Eelco Dolstra 686b6271d2 * Cleaned up the foreign key constraints.
* Generate SQLite and PostgreSQL schemas from hydra.sql.
2009-10-21 12:25:43 +00:00
Eelco Dolstra cec3201720 * Renaming "release sets" to "views" (not finished yet). Having
releases as a dynamic view on the database was misguided, since
  doing thing like adding a new job to a release set will invalidate
  all old releases.  So we rename release sets to views, and we'll
  reintroduce releases as separate, static entities in the database.
2009-10-15 21:35:19 +00:00
Eelco Dolstra 6cedee5476 * Allow jobsets to be disabled. 2009-10-08 11:39:16 +00:00
Eelco Dolstra 158a746e94 * Don't bother with the Jobs.active column anymore. 2009-10-08 11:19:39 +00:00
Eelco Dolstra 16f2d003b2 * In the last succeeded / job status queries, use the Builds.isCurrent
column instead of Jobs.active.
2009-10-07 15:45:17 +00:00
Eelco Dolstra e9cf409d80 * Mark the "current" builds in a jobset, i.e. those corresponding to
the derivations that the jobset currently contains.  This is
  necessary to allow the "latest" channel to contain the correct
  builds when the sources of a jobset are reverted.
2009-10-02 16:06:28 +00:00
Rob Vermaas 87f3fbaab1 extra indexes 2009-09-30 12:17:55 +00:00
Eelco Dolstra 1aec78014d * In the job status and error pages, show when the status of a job
last changed.
2009-07-09 14:48:15 +00:00
Eelco Dolstra 5bdd5e7152 * Added a maintainers field to the Builds table.
* Regenerated the schema bindings with the latest DBIx::Class.
2009-07-07 13:59:59 +00:00
Eelco Dolstra 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
Eelco Dolstra 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
Eelco Dolstra 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
Eelco Dolstra d13af4f392 * Doh. 2009-03-20 10:19:46 +00:00
Eelco Dolstra 4404800ad8 * Release sets: need to include the jobset name to disambiguate
between jobs with the same name from different jobsets (e.g. trunk
  vs. branch).
2009-03-14 22:34:22 +00:00
Eelco Dolstra ae364b9e5f * Represent jobs explicitly in the DB. 2009-03-13 14:49:25 +00:00
Eelco Dolstra f2f586d842 * Disambiguate jobs by jobset name. I.e. jobs with the same name in
different jobsets are not considered the same job.
2009-03-12 23:46:17 +00:00
Eelco Dolstra a623ad157e * Add the name of the jobset to ReleaseSetJobs, otherwise we can't
distinguish between jobs with the same name in different jobsets
  (e.g. "trunk" vs "stdenv-branch" for Nixpkgs).
* Renamed the "attrName" field of Builds to "job".
* Renamed the "id" field of BuildSteps to "build".
2009-03-12 14:18:30 +00:00
Eelco Dolstra fec08462ee * In Sqlite "release" is now a keyword, so use "release_" instead.
Note: to upgrade old databases, do a dump with an old Sqlite first;
  dumping with a new Sqlite will silently discard (!) the contents of
  the ReleaseSetJobs table.
2009-03-12 13:42:44 +00:00
Eelco Dolstra 54a122b49c * DB integrity: when deleting a build, get rid of its build steps. 2009-03-10 08:35:39 +00:00
Eelco Dolstra fca7fb20c4 * Negative caching: don't perform a build if a dependency already
failed in a previous build.  This is essential for Nixpkgs: we don't
  want to keep doing the same failed dependency (say, Glibc) over and
  over again for a few hundred jobs.
2009-03-09 17:21:10 +00:00
Eelco Dolstra 36fdd7f37f * For products that are directories (like manuals), allow a default
suffix other than index.html to be declared.  E.g. if a build does

    echo "doc manual $out manual.html" >> $out/nix-support/hydra-build-products

  the default link for the product is

    http://localhost:3000/build/417/download/1/manual.html

  but other files are also accessible, e.g.
    
    http://localhost:3000/build/417/download/1/style.css
2009-03-06 13:34:53 +00:00
Eelco Dolstra dca6b943d0 * Allow scheduled builds to be cancelled. They're not removed from
the database, just marked as cancelled, because otherwise the
  scheduler would just add them again.
2009-03-06 12:49:01 +00:00
Eelco Dolstra 97ed2052ba * Move everything up one directory. 2009-03-05 13:41:57 +00:00