MSP-11671
The run times for cucumber are so much faster now that `spec --tag
content` is actually the slowest job (on 1.9.3). The job is so slow
that the two cucumbers can be merged together and they will take the
same time, but we'll consume 2 fewer job slots (1 for 1.9.3 and 1 for
2.1), so the longer per job time is better as long as it does become the
longest job time.
MSP-11671
Development group contains only yard+redcarpet, which is used to
generated documentation, which isn't done on travis-ci, and pry, which
is used for the msfconsole command, which isn't tested on travis-ci.
MSP-11616
Travis runs all entries in the script section even if earlier entries
exit with 1 and only fails the build as a whole once all entries have
run by &&ing together the exit codes. Running `bundle exec rake
$RAKE_TASK` if the db/schema.rb is missing updates is unnecessary, so
use `&&` in the script section to stop `bundle exec rake $RAKE_TASK`
from running if `git diff --exit-code` exits with 1.
MSP-11616
With `git diff --exit-code` in `before_script` it caused the build to
show as errored in travis and didn't turn the merge button red, so try
moving to the `script` section to see if that will make the merge button
red.
This change updates msftidy to be run automatically for new modules
added since the last tag release because we can't rely on folks using
tools/dev/pre-commit-hook before submitting a PR. Now, when one attempts
to open a PR with a non-tidy'ed module, the build will fail out of the
gate.
Related to the 100s of msftidy errors extant today.
[SeeRM #8498]
commit c894e52de5705a1133191be5e9caf3ebdee33621
Author: Tod Beardsley <tod_beardsley@rapid7.com>
Date: Fri Jan 31 14:17:02 2014 -0600
Add a jacked up title to test travis. Revert this!
commit 2f00c190be71aeb456a7a546071286fd6d670bc1
Author: Tod Beardsley <tod_beardsley@rapid7.com>
Date: Fri Jan 31 11:39:42 2014 -0600
Allow for checking and spotchecking.
commit db11e8dfad5381030b08c431a183dbafe7a5f304
Author: Tod Beardsley <tod_beardsley@rapid7.com>
Date: Thu Jan 30 17:16:37 2014 -0600
Whoops, need to exit an Integer always.
commit 12d131d3157a78ff11e597476138323ed0a062fc
Author: Tod Beardsley <tod_beardsley@rapid7.com>
Date: Thu Jan 30 16:59:35 2014 -0600
Allow for exit statuses from msftidy.
commit 2c3b294ff17416f49935472caf2b6be3dbdd93a4
Author: Tod Beardsley <tod_beardsley@rapid7.com>
Date: Thu Jan 30 15:36:43 2014 -0600
Be more dynamic about tag checking years
commit d5d8a0b05ac17fb18666a9c252dbb6928d6b5e56
Author: Tod Beardsley <tod_beardsley@rapid7.com>
Date: Thu Jan 30 14:36:44 2014 -0600
Don't warn when there's really nothing
commit fb44a3142fb01eb2647c1c240bb1cc2e7bf59120
Author: Tod Beardsley <tod_beardsley@rapid7.com>
Date: Thu Jan 30 14:21:50 2014 -0600
Revert the intentional failure
This reverts commit 99a7630b0da301b27ac495cb027009a8cd9e2caf.
Fun fact: Reverting a commit does not automatically sign with my current
aliases, one must git revert then git c --amend.
commit 99a7630b0da301b27ac495cb027009a8cd9e2caf
Author: Tod Beardsley <tod_beardsley@rapid7.com>
Date: Thu Jan 30 14:08:05 2014 -0600
Cause an exit status in precommit check
Maybe travis will see these and fail the build.
Don't forget to revert this commit @todb-r7 !
commit 5a3b2fcd9598fae51a0dd2c7c87680c703a85448
Author: Tod Beardsley <tod_beardsley@rapid7.com>
Date: Thu Jan 30 13:11:04 2014 -0600
Update msftidy pre-commit-hook for spotchecking
commit 3f255e36dad9ed3081aaf359f845525d96872ef0
Author: Tod Beardsley <tod_beardsley@rapid7.com>
Date: Thu Jan 30 12:35:16 2014 -0600
Travis should run msftidy via precommit hook
commit 0959d9d2d281590a94c0ac960e43b74354e4e21b
Author: Tod Beardsley <tod_beardsley@rapid7.com>
Date: Thu Jan 30 12:25:53 2014 -0600
Add SPOTCHECK_RECENT to msftidy.rb
Instead of depth=1, try depth=5. This way, if something gets landed
before travis had a chance to build, it'll still get tested, at least
post-facto.
This is the root cause of the build failure at
https://travis-ci.org/rapid7/metasploit-framework/builds/11436293
That commit is not available because the HEAD pointer moved past it with
the next commit; IOW, we landed too quick for Travis to rspec each
change, which is generating failed build messages.
If there was really a failure in there, it'd be a pain now to determine
which commit actually broke the build.
Add the .travis.yml so that travis-ci.org knows which rubies (1.8.7 and
1.9.3 to run rake against. Add the build status button to the README.md
so that people can see if the specs are passing from the main
rapid7/metasploit-framework page on github just like other open source
projects.
Luke Imhoff
Tod Beardsley
James Lee
darkbushido