merge with master

possibly working multithread contents writer that has udebs disabled

etch-m68k

disable uploads to m68k etch.

Signed-off-by: Joerg Jaspert <joerg@debian.org>

apt.conf

remove etch-m68k from apt.conf. No need to always regenerate its files.
There hasnt been an upload to it for months.
And starting today, etch-m68k is also untouchable. We should reject uploads for it

Signed-off-by: Joerg Jaspert <joerg@debian.org>

Merge commit 'mhy/master' into merge

Merge update10 changes

Signed-off-by: Mark Hymers <mhy@debian.org>

src_uploaders deletions now cascade from source

Signed-off-by: Mark Hymers <mhy@debian.org>

fix update script

Signed-off-by: Mark Hymers <mhy@debian.org>

add update to fix up src_uploaders

Signed-off-by: Mark Hymers <mhy@debian.org>

Merge commit 'mhyftpmaster/master'

todo

fixed an item.

Signed-off-by: Joerg Jaspert <joerg@debian.org>

process-accepted

fix the mess in stable-install.
Its still not optimal, but it should no longer need manual changes just to do
an oldstable point release.

Signed-off-by: Joerg Jaspert <joerg@debian.org>

Various

Fix a number of epydoc errors we had.
Make dakdb/ have a much nicer looking api doc :)

Signed-off-by: Joerg Jaspert <joerg@debian.org>

Merge branch 'master' of ssh://ftp-master.debian.org/srv/ftp.debian.org/git/dak

* 'master' of ssh://ftp-master.debian.org/srv/ftp.debian.org/git/dak:
  fix undefined cursor
  Fix the is_dm check to deal with the fact we have typing in our DB layer

TODO

add p-a oldstable mess

Signed-off-by: Joerg Jaspert <joerg@debian.org>

fix undefined cursor

Signed-off-by: Mark Hymers <mhy@debian.org>

Revert "add suite wrapper class"

This reverts commit a26b794c1190cacae572e4c416fd8b2450e4bd33.

Conflicts:

daklib/dbconn.py

Signed-off-by: Mark Hymers <mhy@debian.org>

Merge commit 'mhyftpmaster/merge'

Merge commit 'ftpmaster/master'

Fix the is_dm check to deal with the fact we have typing in our DB layer

This code is still fundamentally wrong (assuming unknown UIDs are DDs is
idiotic; of course we know that the keys have had to pass the sig check
against the keyring so it's not a security issue thankfully; it just
might give a one shot limited window for DMs to upload non-DM packages)

Signed-off-by: Mark Hymers <mhy@debian.org>

Merge branch 'master' into merge

Merge commit 'origin/master' into merge

Merge commit 'origin/master'

Merge commit 'djpig/djpig' into merge

* commit 'djpig/djpig':
  cruft_report: Hack: skip debian-installer on kfreebsd
  cruft_report: Fix syntax error

Signed-off-by: Joerg Jaspert <joerg@debian.org>

cruft_report: Hack: skip debian-installer on kfreebsd

Doesn't exist yet.

cruft_report: Fix syntax error

Merge branch 'master' of ssh://ftp-master.debian.org/srv/ftp.debian.org/git/dak

* 'master' of ssh://ftp-master.debian.org/srv/ftp.debian.org/git/dak:
  set STAMP variable during dinstall (mhy)

dinstall

also generate the arch-space/pkg-nums stats here

Signed-off-by: Joerg Jaspert <joerg@debian.org>

set STAMP variable during dinstall (mhy)

signed buildd incoming dir

moved all the code creating this into the common file, new function
cron.unchecked now just calls this function (and so sources the common file)
cron.dinstall also calls this function, immediately after process-accepted did run.
This should get the times where buildds get 403 on files down to at max. the runtime
of process-accepted.

Signed-off-by: Joerg Jaspert <joerg@debian.org>

dinstall

tell merkel that it should sync the dd accessible parts.
do that after the long run of apt-ftparchive clean, so we have a large
delay between the earlier "sync projectb" and this, so to not overload
merkel too much.

Signed-off-by: Joerg Jaspert <joerg@debian.org>

dinstall

name the pre dumps differently.
delete predumps after 2 days.
also compress the dumpall files
hardlink same files.

Signed-off-by: Joerg Jaspert <joerg@debian.org>

expire_dumps

really tell the full rules

Signed-off-by: Joerg Jaspert <joerg@debian.org>

Merge branch 'master' into merge

dinstall

move pgdakdev way later where it doesnt matter that it takes long to run

Signed-off-by: Joerg Jaspert <joerg@debian.org>

dinstall

actually call the stage function...

Signed-off-by: Joerg Jaspert <joerg@debian.org>

kfreebsd

add kfreebsd-[i386|amd64]

Signed-off-by: Joerg Jaspert <joerg@debian.org>

examine_package

Signed-off-by: Joerg Jaspert <joerg@debian.org>

examine_package

docstring
"clearup" relations variable after output.

Signed-off-by: Joerg Jaspert <joerg@debian.org>

examine_package

docstring
"clearup" relations variable after output.

Signed-off-by: Joerg Jaspert <joerg@debian.org>

process-new

and also call output_package_relations from process-new

Signed-off-by: Joerg Jaspert <joerg@debian.org>

examine_package

output a list of package relations at the very end of our run.
should make it easier to compare them for packages that split into MANY

Signed-off-by: Joerg Jaspert <joerg@debian.org>

examine_package

one more parameter to output_deb_info, the packagename.

Signed-off-by: Joerg Jaspert <joerg@debian.org>

examine_package

a little docstring foo and "one import per line"

Signed-off-by: Joerg Jaspert <joerg@debian.org>

utils.create_hash()

if we cant open the file, don't try to parse it, but continue with the next

Signed-off-by: Joerg Jaspert <joerg@debian.org>

generate-releases

also generate release files for the debian-installer directories.

Signed-off-by: Joerg Jaspert <joerg@debian.org>

generate-releases

move the actual writing of the "sub"-release files into a function.

Signed-off-by: Joerg Jaspert <joerg@debian.org>

p-n

this should fix the case where files moved into [o-]p-u-new are
unreadable for everyone but the one running p-u-new. The code
simply didnt set the permissions right, defaulting to 0660 instead.

Signed-off-by: Joerg Jaspert <joerg@debian.org>

Merge branch 'master' into content_generation

ddtp_i18n

import new revision of the i18n script

Signed-off-by: Joerg Jaspert <joerg@debian.org>

todo foo

move a few entries over, delete some others.

Signed-off-by: Joerg Jaspert <joerg@debian.org>

criteria

fixed html errors. made dsa paragraph more weasel compatible :)

Signed-off-by: Joerg Jaspert <joerg@debian.org>

Merge commit 'ftpmaster/master'

remove base

Signed-off-by: Joerg Jaspert <joerg@debian.org>

archive criteria

merge the "old" criteria together with the "new" rules into one document.

Signed-off-by: Joerg Jaspert <joerg@debian.org>

Add a todo

Signed-off-by: Joerg Jaspert <joerg@debian.org>

Merge commit 'godog/master' into merge

* commit 'godog/master':
  add guid element (same as link) to make feed validator happy
  html entity escape for Maintainer/Description/Changes

Signed-off-by: Joerg Jaspert <joerg@debian.org>

add guid element (same as link) to make feed validator happy

html entity escape for Maintainer/Description/Changes

Merge commit 'stew/master' into merge

* commit 'stew/master':
  change reject email template to solicit a reply if new files are uploaded

Signed-off-by: Joerg Jaspert <joerg@debian.org>

change reject email template to solicit a reply if new files are uploaded

Signed-off-by: Mike O'Connor <stew@vireo.org>

change reject email template to solicit a reply if new files are uploaded

Signed-off-by: Mike O'Connor <stew@vireo.org>

we're using something which understands typing now

Signed-off-by: Mark Hymers <mhy@debian.org>

Remove use of an undeclared variable.

This was causing us to reject packages which already had an orig
in the pool even if matched.

Signed-off-by: Mark Hymers <mhy@debian.org>

Merge branch 'master' into content_generation

convert binaries.type to a enum

Signed-off-by: Mike O'Connor <stew@vireo.org>

some tweaking to the contents generation queries

Signed-off-by: Mike O'Connor <stew@vireo.org>

todo

start a new todo list. we need to check the old and copy everything over
which is still valid, i just dont want to bury new items in a large number
of old and *maybe* invalid items.

Signed-off-by: Joerg Jaspert <joerg@debian.org>

strip ./ from front of paths during contents.bootstrap

Signed-off-by: Mike O'Connor <stew@vireo.org>

More quotes

Signed-off-by: Joerg Jaspert <joerg@debian.org>

tidy up output

Signed-off-by: Mark Hymers <mhy@debian.org>

fix clean_suites maximum option

Signed-off-by: Mark Hymers <mhy@debian.org>

Merge branch 'master' into merge

Signed-off-by: Mark Hymers <mhy@debian.org>

Merge commit 'origin/master'

only delete 10000 files per clean-suites

Signed-off-by: Mark Hymers <mhy@debian.org>

cron.monthly

sync with raff and enable on ries

Signed-off-by: Joerg Jaspert <joerg@debian.org>

quote

Signed-off-by: Joerg Jaspert <joerg@debian.org>

somehow missed two sections for security

Signed-off-by: Mark Hymers <mhy@debian.org>

localizations -> localization

Signed-off-by: Mark Hymers <mhy@debian.org>

add note about new sections

Signed-off-by: Mark Hymers <mhy@debian.org>

add new sections

Signed-off-by: Mark Hymers <mhy@debian.org>

and as if by magic, we'll get something in p-u's Sources file

Signed-off-by: Mark Hymers <mhy@debian.org>

err, wtf?

shutup this idiotic thing and make p-u work again. this does *not* seem
right, but it helps. :()

Signed-off-by: Joerg Jaspert <joerg@debian.org>

update web page

Signed-off-by: Mark Hymers <mhy@debian.org>

Merge commit 'ftpmaster/master'

needs to be a tuple

Signed-off-by: Mark Hymers <mhy@debian.org>

rename variable

Signed-off-by: Mark Hymers <mhy@debian.org>

remove unneeded database. reference

Signed-off-by: Mark Hymers <mhy@debian.org>

Merge commit 'ftpmaster/master'

Conflicts:

dak/init_db.py

Signed-off-by: Mark Hymers <mhy@debian.org>

fix import path

Signed-off-by: Mark Hymers <mhy@debian.org>

fix import path

Signed-off-by: Mark Hymers <mhy@debian.org>

Merge commit 'ftpmaster/master'

remove extra quote

Signed-off-by: Mark Hymers <mhy@debian.org>

fix up oldstable suite check

Signed-off-by: Mark Hymers <mhy@debian.org>

fix up SQL call (hackish)

Signed-off-by: Mark Hymers <mhy@debian.org>

fix up SQL call (hackish)

Signed-off-by: Mark Hymers <mhy@debian.org>

fix len() calls

Signed-off-by: Mark Hymers <mhy@debian.org>

Merge commit 'stew/content_generation' into merge

Signed-off-by: Mark Hymers <mhy@debian.org>

fix the stupid commit that preceeded

Signed-off-by: Mike O'Connor <stew@vireo.org>

fix things I broke yesterday with the contents merge

* alter the pending_content_associations table to include arch
* only allow debs with more than 3 ar members during bootstrap

Signed-off-by: Mike O'Connor <stew@vireo.org>

Merge branch 'master' into content_generation