summaryrefslogtreecommitdiffstats
Commit message (Collapse)AuthorAgeFilesLines
* Don't send out reports with quoted-printable encoding.HEADmasterJeffrey C. Ollie2009-01-051-4/+4
|
* Improve MIME-correctness of rawhide reports.Jeffrey C. Ollie2008-12-302-1/+108
|
* Pass a previous run to mash, so we can copy the repodata.Bill Nottingham2008-12-161-1/+1
| | | | Requires mash-0.4.3 or later.
* Use a unique extention for the mock chroot for rawhide composes.Jesse Keating2008-12-021-8/+8
| | | | | | This will help me if I forget to exit out of the mock chroot if I have been debugging during the day. Each day gets a fresh extension and thus no locking issues.
* Add a timestamp to the start of the rawhide report.Jesse Keating2008-10-061-1/+3
|
* Back to 'development' as the configJesse Keating2008-09-171-1/+1
|
* Call it 'rawhide' in mash.Bill Nottingham2008-09-151-1/+1
| | | | Corresponding mash change building.
* Remove the need to become rawhide to send mail. This probably means we canJesse Keating2008-08-281-1/+1
| | | | kill the rawhide user all together.
* Use sudo to call rsync and mail.Jesse Keating2008-08-281-3/+3
|
* No need to remove the rpmdb, we clean the chroot out anyway.Jesse Keating2008-08-271-1/+0
|
* Change mode to be executable.Jesse Keating2008-08-271-0/+0
|
* Do depchecks as unpriv too, since we output to nfs again.Jesse Keating2008-08-271-2/+2
|
* Don't worry about OLD, just compare tree to last pushed rawhide. Also do it ↵Jesse Keating2008-08-271-8/+2
| | | | | | unpriv so that it can write the output to the nfs store.
* Define that we want gpg sigsJesse Keating2008-08-271-1/+1
|
* Add new keys.Jesse Keating2008-08-271-5/+9
|
* Reference the newlocation for signing certsJesse Keating2008-08-271-3/+3
|
* Remove the failsafes in the rawhide script.Jesse Keating2008-08-261-6/+3
|
* Make repodiff compare against the last published rawhide. Less chances of ↵Jesse Keating2008-08-261-1/+1
| | | | | | missing changes that the public would see.
* Don't call mock@ in pungify arguments, pungify will do that for us.Jesse Keating2008-08-251-1/+1
|
* NFS server changed hostnamesJesse Keating2008-08-251-1/+1
|
* Don't bother updating after creating the chroot, since we create it fromJesse Keating2008-08-251-18/+0
| | | | the static-repos. This means we'll be as up to date as necessary.
* ssh out as the masher user, instead of 'mock'. This is so that we can writeJesse Keating2008-08-251-4/+4
| | | | to the /mnt/koji share during the rsync.
* Call pungify from the cwd, since we'll be in a fresh git checkoutJesse Keating2008-08-251-1/+1
|
* More fixes from testing rawhide production.Jesse Keating2008-08-201-1/+4
| | | | | | | Make sure we run mash as the user calling mock, which in our case is masher, which has write access to /mnt/koji Copy in the host hosts file so that we correctly resolve koji.fedoraproject.org
* We want a temp directory, not a temp file in buildrawhide.Jesse Keating2008-08-201-1/+1
|
* Use a "secure" tempdir for doing the cvs checkout in.Jesse Keating2008-08-201-2/+2
|
* Disable some things for testing actual cron usage.Jesse Keating2008-08-191-3/+6
|
* We're doing bashisms, make it a bash scriptJesse Keating2008-08-191-1/+1
|
* Clean out rpmdb caches to avoid db mismatches in the chrootJesse Keating2008-08-191-0/+2
|
* Use the --chroot command (non-interactive) instead of --shellJesse Keating2008-08-191-3/+3
|
* Make the mock config used for rawhide production a variableJesse Keating2008-08-191-6/+7
|
* Fix a non-shell typo, thanks dgilmoreJesse Keating2008-08-191-1/+1
|
* Send mail to the right list.Jesse Keating2008-07-261-1/+1
|
* Actually fill in the owner alias. *sigh*Jesse Keating2008-07-261-1/+1
|
* Update the docsJesse Keating2008-07-261-3/+2
|
* Call function to send direct mail. Send mail ourselves to list rather than asJesse Keating2008-07-261-15/+27
| | | | piped output to /bin/mail
* Add a function to send mail directly to those interested in a packageJesse Keating2008-07-261-0/+34
| | | | as well as the person who did the build that broke the upgrade path.
* Don't do inheritance, only consider the packages specifically tagged.Jesse Keating2008-07-261-1/+1
| | | | | | As long as we start with some -final tag, we will catch every build along the way. There is no need to compare a build to itself inherited in a later tag, and inheritance can make the updates tags look...funny.
* Add a builder sorted list of bad upgrade paths.Jesse Keating2008-07-241-5/+27
| | | | | This is getting to be pretty damn ugly code. I think it needs a better storage mechanism than dicts of dicts of lists of dicts of etc....
* Add some documentation about what we do and don't do.Jesse Keating2008-07-211-0/+8
|
* Add more about where this came from, and a footer leading readers to where ↵Jesse Keating2008-07-211-0/+5
| | | | the code is
* Add a blurb at the top of the report regarding what tags were checkedJesse Keating2008-07-211-0/+2
|
* Add a new script that checks e:n-v-r paths based on koji informationJesse Keating2008-07-211-0/+92
|
* Add more verbosityJesse Keating2008-06-061-2/+2
|
* Handle doing a sign and write when something is signed but not written out.Jesse Keating2008-06-051-12/+17
|
* Create a new function to be able to write sigs via multicallJesse Keating2008-06-051-4/+19
|
* Create new functions to be able to use multiCall when verifying and ↵Jesse Keating2008-06-051-3/+25
| | | | importing sigs
* Don't try to import the base package, those aren't signed. Waste of time.Jesse Keating2008-06-051-2/+2
|
* Use multicall when getting the rpms from build n-v-rsJesse Keating2008-06-051-5/+11
|
* Use multicall when getting the uncached signaturesJesse Keating2008-06-051-8/+7
|
generated by cgit (git 2.34.1) at 2024-04-26 01:22:09 +0000