[diffoscope] branch master updated (53a2143 -> 9415dd7)

Ximin Luo infinity0 at pwned.gg
Wed Jul 20 18:06:00 CEST 2016

Holger Levsen:
> Hi,
> (sorry for nitpicking but I think this is important.)
> On Tue, Jul 19, 2016 at 01:13:00PM +0000, Ximin Luo wrote:
>> It says in the email description of the force commit, 1 new 1 discard.
> well, yes, that tells me I have to diff these commits, which I haven't
> done. I prefer not to get into such situations by avoiding --force
> pushes (of shared branches).
>> I'm sure myself too, yes.
> great.
> (I hope you also agree that --force pushes shouldnt be done and this was
> a mistake… "I hope" because you haven't commented on that part…!)

I don't agree that --force pushes "shouldn't" be done in general, but I can agree to not do those in the future if this is the policy for diffoscope.

You could also go into diffoscope.git/config and set [receive] denyNonFastforwards = true.

There are a range of situations where force-push is "better" or "worse", and I think the one I did was towards the "better" side. In general, if your local history is H and I push H <- A then force-push H <- A', this doesn't affect you at all (despite the scary email about a forced push). I did the force-push within like 3 minutes of the original, so I guessed that nobody had fetched A yet.


GPG: ed25519/56034877E1F87C35
GPG: rsa4096/1318EFAC5FBBDBCE

More information about the diffoscope mailing list