Re: [Gnu-arch-users] Re: [GNU-arch-dev] Re: how to fix a bad log message?

[James Blackwell]

> Arch *could* certainly automate that kind of cheating and, in spite of 
> the rumours, I'm all for that.   That's not the same thing as being in
> favor of a poor design/implementation of this feature, though.
>

The way I'd do it then is this: 

for this--that--0--patch-56 to be disavowed: 

1. User gets patch-56
2. user fixe 56 however he wants
3. user does a "commit --replace patch-56"
4. arch then checks every single proceeding revision to ensure
   that future history (ohhh, I've wanted an excue to say that
   for years), is consistant.
5. If all is good, save patch-56-1 (or whatever). If not, refuse.
6. delete patch-56


For step 4, one can make later optimizations (for example, if only the
commit summary changed, one needn't replay). Though this won't ensure
safety for child branches, if the parent branch is able to cope with it,
then child branches may be able to cope with it too.

But I can't wait for the day that somebody starts shouting and screaming
the following: "I deleted this revision from my branch after a user
merged me. I merge him back, and the code is suddenly back".

-- 
 James Blackwell      |   Life is made of the stuff that hasn't killed
 Tell someone a joke! |   you yet.                       - yours truly
----------------------------------------------------------------------
GnuPG (ID 06357400) AAE4 8C76 58DA 5902 761D  247A 8A55 DA73 0635 7400