| Age | Commit message (Collapse) | Author | Files | Lines |
|---|
|
reproducible→unreproducible status change
|
|
|
|
|
|
* new column in the database, schedule table
* turn $SAVE_ARTIFACTS a true boolean, 0→no artifacts, 1→yes artifacts
* $NOTIFY in the build script is text containg a keyword:
+ if empty no notification will be sent
+ if "debbindiff" a "$DBDVERSION had trouble with these" will be sent
+ if "failure" a "no artifacts could have been saved"-like will be sent
+ any other value trigs a normal message
* adapt the maintenance script and the manual scheduler to cope with the new
column
|
|
jenkins log
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
another running build wherever possible, using a lockfile named after the package name/suite
|
|
|
|
|
|
file permissions
|
|
|
|
what to put into it
|
|
|
|
reproducible but 1) there is .buildinfo 2) there is .debbindiff.html
|
|
goes wrong
|
|
|
|
very start: we use it from the start, cant' be defined in a function
|
|
and rewrite it all
|
|
the scope
|
|
to avoid pollute the global env
|
|
call to the main script, out of build_rebuild(). way clearer now
|
|
packages
|
|
to avoid pollute the global env
|
|
+ better message generation
|
|
update_db_and_html()
|
|
|
|
|
|
|
|
evenience
|
|
kgb-client call
|
|
the case
|
|
output also to the stdout, aka jenkins log
|
|
|
|
either having .debbindiff.html or not having .buildinfo
|
|
strip out handle_ftbr(). in all these strips I removed a loads of checks and something broke, let's fix
|
|
strip out handle_reproducible()
|
|
|
|
if/elif/elif/fi
|
|
strip out dbd_timeout()
|
|
strip out init_debbindiff() where we check for locks
|
|
|
|
build_rebuild() and the rest to handle_ftbfs()
also split more lines, and use more local variables, instead of pollute the global environment
|