No description
Find a file
Sergio Durigan Junior 51966d39ba Extending regression report.
We now report FAIL -> PASS and PASS -> FAIL, as well as new PASS and
new FAIL.
2017-09-14 16:35:23 -04:00
lib Extending regression report. 2017-09-14 16:35:23 -04:00
scripts Revamping the try_count approach 2017-09-05 09:43:15 -04:00
.gitignore Modifying .gitignore to not ignore new files in lib/ 2014-12-13 03:41:48 -05:00
LICENSE Initial commit 2013-08-01 10:37:22 -07:00
master.cfg Including comment on try build email 2017-09-13 01:03:48 -04:00
README.md update README.md 2013-08-01 11:51:57 -06:00

gdb-buildbot

gdb buildbot configuration

This is a buildbot configuration for gdb.

Currently it is a bit out of date. For example, native-gdbserver.exp is now in-tree and doesn't need any special configuration.

This configuration makes it easy to deploy slaves that test gdb in different scenarios. For example there is a canned way to run the test suite using gdbserver. It is also easy to add new configurations. This is important because gdb can be used in many different ways.

It collects the .sum files and establishes a baseline. The baselines are then "ratcheted": new PASSes are automatically added to the baseline for a given slave configuration, and any regressions are flagged as build failures.

This also provides a way to retrieve the .sum files for any given revision and configuration. This can be handy for doing local comparisons.

To do items include:

  • Add more configurations. For example, dwz, split debuginfo, etc.

  • Store the .sum files in a nicer way. Right now they accumulate and are not de-duplicated. Just de-duplicating would help a bit, but perhaps a real database of some form would be nicer.

  • Provide a way to do nice web-based comparisons across different configurations.

  • Host it somewhere :-)