[PATCH v5] Allow commit date before Unix epoch, clean fix for (issue2513)

Florent Gallaire fgallaire at gmail.com
Sun Apr 3 21:13:21 EDT 2016


2016-04-04 1:45 GMT+02:00 Pierre-Yves David <pierre-yves.david at ens-lyon.org>:
> Can we have a small inline comment that explain where this 0x7fffffff value
> comes from?

This is already done in the parsedate() function in the same util.py file.

> As I said earlier, we want stronger testing of this (or pointer that this is
> already tested properly testing in another patch),
>
> You are adding (okay, re-introducing) a space of possible date, we should
> add automated testing for values in this space including possible limits.

Add all tests you want.

> If we don't do this, I'll have a low confidence that this patch is not
> breaking windows again. And equally important, someone may as well silently
> break that feature in the future, reintroducing regression.

You can be confident, this patch doesn't break Windows (more than it
already is of course).

Cheers,

Florent

-- 
FLOSS Engineer & Lawyer


More information about the Mercurial-devel mailing list