Reimplement signature check without pefile

[catlee]

No description provided.

[coveralls]

Coverage decreased (-4.3%) to 61.475% when pulling ba5400d on catlee:master into 34a81bb on mozilla-releng:master.

[catlee]

Will add some tests for the new function

[coveralls]

Coverage increased (+0.5%) to 66.331% when pulling e82cce6 on catlee:master into 34a81bb on mozilla-releng:master.

[coveralls]

Coverage increased (+0.5%) to 66.331% when pulling 37fb213 on catlee:master into 34a81bb on mozilla-releng:master.

[escapewindow]

Schweet. I can stamp this if you'd like.

• we just pushed out 3.1.2 for bug 1380536, so we need to bump the version higher
• once this lands, we can remove pefile from the signing scriptworker deps in puppet [!]
• how confident are you in this? we could land this sooner so we have more baking time, and possibly see signing issues in Nightly, or we could hold off til post-57 and land early in the cycle... but it's not tree specific, so it would affect betas/chemspills right away. I think I just convinced myself we should land this sooner.

[coveralls]

Coverage increased (+0.5%) to 66.331% when pulling bd4bd50 on catlee:master into 99c8f40 on mozilla-releng:master.

[catlee]

I'm pretty confident. I ran this over all the exe and dll files I could find, and it matches pefile's behavior. Can we roll it out to the dep workers first?

[catlee]

Want me to rebase / squash instead of merging it in?

[escapewindow]

I'm pretty confident. I ran this over all the exe and dll files I could find, and it matches pefile's behavior. Can we roll it out to the dep workers first?

Cool! Puppet has the signing scriptworkers modules all configured in the same shared place, so we can't roll out separately without either refactoring or pointing some workers at a puppet env. Do you want that?

Want me to rebase / squash instead of merging it in?
I don't have strong opinions about rebase/squash vs merge in this repo.