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

[Michael Poole]

Adrian Irving-Beer writes:

> On Wed, Feb 09, 2005 at 09:19:12AM -0500, John S. Yates, Jr. wrote:
> 
> > I understand the model.  I understand the purity/simplicity/rigor
> > arguments. But I also recognize that in the real world things are
> > not nearly as clean as you all want to make it.
> 
> Actually, in the 'real world', things are even cleaner, since everyone
> knows you can't go back to fix your mistakes, and there's no mailing
> list to request that feature on. ;)
> 
> But of course you mean software management in the real world.  Do you
> have examples of when this proposed undo functionality is required?
> Examples where simply committing a reversal isn't sufficient?

One real-world example of needing that sort of operation was when bad
RAM caused a corrupt patch to be committed on a project I maintain.
(I think it was triggered by the pristine tree becoming corrupt; a
diff inside the patch was incorrect.)  Trying to apply the patch to a
correct tree made tla panic.  After some work, I did replace the patch
with a correct version.

The now-fixed bug of renaming (and changing?)  an old file at the same
time as creating a different file with the previous name could also
generate a corrupt patch.  Since I did not know as much about the
archive format when that bug bit me, I could not find a better way to
fix it than to create a new archive.

I hope creating a new version or branch will not always the approved
way to fix an invalid (invariant-violating) patch.

Michael Poole