News for hackers. Stuff that matters.
Re:Follow up to the investigation results by pyre on Monday April 28, 2014 @05:46PM (#7663326)
Depends on what kind of logs they have. Laying down a couple of assumptions:

1. They only have the git repo history to go on.

2. It's been long enough that errant objects have been garbage collected.

It's possible that the code she committed removed via a push that removed it from history entirely (e.g. `git push +old-master:master`). If this push was never reverted, then the commit object would become an orphan that was up for removal at the next garbage collection interval. After that, the incident would absolutely just be 'he said, she said,' and a lack of evidence would not be proof that she was lying.

Re:Follow up to the investigation results by kemayo on Monday April 28, 2014 @06:04PM (#7663427)
Unless the guy was hyper-vigilant about reverting her code (as in, within a few seconds of her pushing her own commit), wouldn't that be a ridiculously noticeable way to do it? Other people would have pulled, and would find themselves inconsistent with the origin repo when they next went to push...

Not saying it's impossible, but it seems like it'd have generated some documentation if it happened, probably in the form of people sending out "WTF? Who's pushing history-changes???" emails.

Edited to add: Also, the article says they reviewed git push logs, which I imagine would catch this.

Re:Follow up to the investigation results by pyre on Monday April 28, 2014 @07:20PM (#7663804)
> as in, within a few seconds of her pushing her own commit

Is it really only possible for a revert to happen 'within a few seconds' to not cause issues? That's a strong statement to make based on little information. It might be true, but the repository in question could also have little traffic, or this could have happened over a weekend.

> Also, the article says they reviewed git push logs, which I imagine would catch this.

I see. I guess overlooked that (probably just read it as 'commit logs').

In general, I don't understand why they didn't just ask her about the commits in question. If it did happen, she would likely be able to point them in the right direction and at least narrow their search (e.g. 'the revert happened in repository X around these dates'). If she couldn't, then they are no better off, but no worse off either.

Re:Follow up to the investigation results by kemayo on Monday April 28, 2014 @07:32PM (#7663858)
> That's a strong statement to make based on little information.

That's a fair point. It'd depend a great deal on how github has their repos set up, and I can't remember seeing anyone talk about that.

Re:Follow up to the investigation results by yawboakye on Monday April 28, 2014 @06:48PM (#7663643)
The investigation considered all possible commits surrounding the accusation of passive-aggressive code removal. One instance was found where the engineer updated and broke some CSS in an internal application, which was fixed in a later commit. The investigator determined this change did not appear malicious.

The nature of the commit/revert.

Re:Follow up to the investigation results by pyre on Monday April 28, 2014 @07:15PM (#7663773)
Your reply does not address what I wrote. In the scenario that I mentioned the commits would no longer exist, and therefore they could review 'all possible commits' without reviewing the commits in question.
Re:Follow up to the investigation results by judk on Monday April 28, 2014 @09:56PM (#7664456)
The company's only product is an enterprise-targeted version control system. If they lose track of commits so easily , that would be rather damning for their product.
Re:Follow up to the investigation results by pyre on Monday April 28, 2014 @11:33PM (#7664773)
I listed the assumptions that I made to demonstrate the possibility of a scenario. That said, you've grossly simplified the situation. What if it is an unreferenced commit from a year ago. Would you really find it damning that you can't go to Github and say:

  Hey guys,

  I had a branch `foo` in my repository `bar` a year or
  so ago. I deleted the branch (which had no references)
  and don't have any backups of it. Do you think that
  you guys could restore it for me?
Re:Follow up to the investigation results by encoderer on Monday April 28, 2014 @05:54PM (#7663369)
Yeah, if they have no backups.

Also, they said they looked at both the push log and the commit history.