mozilla-releng / scriptworker / 212

100%

master: 100%

push

travis-ci

0.8.0: gpg homedir creation (#16)

* update verify git signed commit logic

* add git repo config items

* config tweaks

* update_signed_repo

* add context.cot_config; start build_gpg_homedirs_from_repo

also create scriptworker.constants and change various things around
before pulling in jordan's download_files fixes in and rebasing.

* move stuff into cot_config

* add gpg_homedirs to cot_config

* flesh out schema; add missing docstrings

* finish build_gpg_homedirs_from_repo; rebuild_gpg async

* flake8 fixes

* create initial gpg homedirs function

* create_initial_gpg_homedirs console_script

* async sign_key.

sign_key calls gpg, which will lock the keyring, so parallelizing
sign_key in the same gpg homedir will not make sense.

however, a) we're going to be creating multiple gpg homedirs in
parallel, and while each sign_key must be run synchronously within
each gpg homedir, there's no reason why we can't run them in parallel
across gpg homedirs; and b) we're going to be creating these gpg
homedirs in the background while we're also signing firefox, so
releasing control to the event loop while we're waiting for sign_key to
finish will allow us to continue signing firefox without blocking on gpg
homedir creation.

* add worker loop to recreate gpg homedirs

* 100% config test coverage

* 100% worker coverage

* 88% gpg coverage

* fix travis?

* rename lockfile; remove _function docstrings; make
build_gpg_homedirs_from_repo more testable

* fix sign_key test failure; update requirements-*.txt with dephash

* test_build_gpg_homedirs_from_repo_lockfile

* bump sign_key timeout 30s->5min due to intermittent unit test failures

* 94% gpg.py coverage

* create_initial_gpg_homedirs tests

* fix: 96% gpg.py coverage

* mocker instead of kwargs

* 100% gpg coverage

* fix

* 0.8.0-alpha1

* update changelog

* update README

* update CHANGELOG to mention the new chain of tru... (continued)

199 of 199 new or added lines in 7 files covered. (100.0%)

1160 of 1160 relevant lines covered (100.0%)

1.0 hits per line