* Re: git merge ignoring whitespace
2009-02-05 23:24 git merge ignoring whitespace layer
@ 2009-02-05 23:44 ` Junio C Hamano
2009-02-05 23:48 ` layer
2009-02-13 1:38 ` Johannes Schindelin
1 sibling, 1 reply; 4+ messages in thread
From: Junio C Hamano @ 2009-02-05 23:44 UTC (permalink / raw)
To: layer; +Cc: git
layer <layer@known.net> writes:
> Junio, on 3/25/07 you said:
>
> I know Linus does not like applying patches with whitespace
> fuzziness, but I personally think a merge that ignores
> whitespace changes would be a useful thing to have, without
> having to call an external specialized merge backend. In other
> words, I do not think it is so special a thing to want to, but
> it is rather an unfortunate norm. I am not opposed to give
> git-merge-recursive a command line option to tell it to merge
> ignoring the whitespace-only changes, when it does the 3-way
> file-level merge internally.
>
> Whatever came of this?
I do not know. When I say "I am not opposed to", I mean just that.
I may consider it a bad mode of operation, and I may hesitate to encourage
people to use it as their default. But I do not feel strongly enough to
veto it---as long as it is a rope you need to explicitly ask for to hang
yourself with, I think it is Ok if we allowed such an option.
It is not my itch to scratch, and I do not keep track.
Did you submit a patch that implements the new feature cleanly?
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: git merge ignoring whitespace
2009-02-05 23:24 git merge ignoring whitespace layer
2009-02-05 23:44 ` Junio C Hamano
@ 2009-02-13 1:38 ` Johannes Schindelin
1 sibling, 0 replies; 4+ messages in thread
From: Johannes Schindelin @ 2009-02-13 1:38 UTC (permalink / raw)
To: layer; +Cc: Junio C Hamano, git
Hi,
On Thu, 5 Feb 2009, layer wrote:
> Junio, on 3/25/07 you said:
>
> I know Linus does not like applying patches with whitespace
> fuzziness, but I personally think a merge that ignores
> whitespace changes would be a useful thing to have, without
> having to call an external specialized merge backend. In other
> words, I do not think it is so special a thing to want to, but
> it is rather an unfortunate norm. I am not opposed to give
> git-merge-recursive a command line option to tell it to merge
> ignoring the whitespace-only changes, when it does the 3-way
> file-level merge internally.
>
> Whatever came of this? I tried giving git merge a --ignore-whitespace
> argument and the error message that followed show no equivalent
> argument.
I can think of two issues:
- there was this -X patch by Junio, but I think it is a bad interface to
pass merge backend options, as it is not closely coupled with the
backend. Just imagine this:
git merge -X ignore-whitespace -s resolve,recursive
Does the "-X ignore-whitespace" impact resolve, recursive, or both? And
we'd need some way to specify the option to the backend. I thought that
git merge -s resolve,recursive/ignore-whitespace
would be a sane syntax, but at least _I_ did not have a need to pass
backend options yet.
- more fundamental, how should the merge ignoring whitespace work? Which
version would it take if the lines agree _except_ for whitespace
changes? This needs to be well designed, as we have not only merges
(where we have "ours" and "theirs"), but also rebases (where we have
"theirs" and _then_ "ours"), and of course recursive merges when there
is more than a single merge base (and there, it gets really nasty, so
you better have the semantics of the whitespace-ignoring merge put down
quite clearly).
> It would sure come in handy right now. Is there some other way to do
> it that I didn't find searching the mailing list?
Of course, you could try to "normalize" the whitespace, and perform the
merge after that.
Ciao,
Dscho
^ permalink raw reply [flat|nested] 4+ messages in thread