Problem with cg-diff <file>

Problem with cg-diff <file>

[Marcel Holtmann]

Hi Petr,

your latest changes are breaking the cg-diff <file> functionality. Now
cg-diff creates a diff against all local not committed changes.

Regards

Marcel

Re: Problem with cg-diff <file>

[Petr Baudis]

Dear diary, on Mon, May 30, 2005 at 01:15:55AM CEST, I got a letter
where Marcel Holtmann <marcel@holtmann.org> told me that...
> Hi Petr,

Hi,

> your latest changes are breaking the cg-diff <file> functionality. Now
> cg-diff creates a diff against all local not committed changes.

oops, thanks for the report. Hopefully fixed and pushed out.

Eek, apparently not entirely fixed yet. git-diff-tree (in contrast to
git-diff-cache) won't take the pathspec as its trailing arguments,
causing cg-diff -r a:b still not to work 100% correct.

Junio, is there any specific reason for that, or is the end of
git-diff-tree argument list the right spot for the pathspec stuff?

Thanks,

-- 
				Petr "Pasky" Baudis
Stuff: http://pasky.or.cz/
C++: an octopus made by nailing extra legs onto a dog. -- Steve Taylor

Re: Problem with cg-diff <file>

[Junio C Hamano]

>>>>> "PB" == Petr Baudis <pasky@ucw.cz> writes:

PB> ... git-diff-tree (in contrast to
PB> git-diff-cache) won't take the pathspec as its trailing arguments,

PB> Junio, is there any specific reason for that, or is the end of
PB> git-diff-tree argument list the right spot for the pathspec stuff?

Baffled.  Are you at Linus tip?

Linus correctly decided that diff-tree does not have to call
diffcore_pathspec(), which may be what confused you to make the
comment "... in contrast to git-diff-cache) won't take ...".
But it does not call it only because it does not need to.  It
filters the filepairs itself on the input side using the
trailing arguments; since diffcore_pathspec filters as the first
one in the chain as the input filter, calling it from diff-tree
would not cull anything further.

Here is what I am getting from the Linus tip binary, between my
HEAD and Linus tip:

$ git-diff-tree -r linus HEAD   >.all    ;# everything
$ git-diff-tree -r linus HEAD t >.t-only ;# limiting to the test suite
$ wc -l .all .t-only                     ;# count results
  12 .all
   3 .t-only
  15 total
$ cat .t-only				 ;# show what's in "t" output
:100644 100644 a51985518251f6c3f677438c3cb51b9716c20296 5ac29d1f98438d3530bbc8b0bdaced58200aca37 M	t/t4005-diff-rename-2.sh
:100644 100644 518892b90c7cbb3fb193d6bfb622046aff0f4431 76ae7201f0d19b7933ca44958b7c468193ec9778 M	t/t4007-rename-3.sh
:000000 100755 0000000000000000000000000000000000000000 01d276692669f2241471b8ad611b17d51e2a98ab N	t/t4009-diff-rename-4.sh
$ head -n 1 .all                         ;# prove that "t" filtered. 
:100644 100644 f85a605f0a336f506cf5cf46476a43e4c56b3e66 1d92a01a02543e55d0feb3541ee594fbc638136c M	Documentation/diff-format.txt

Re: Problem with cg-diff <file>

[Petr Baudis]

Dear diary, on Mon, May 30, 2005 at 02:19:09AM CEST, I got a letter
where Junio C Hamano <junkio@cox.net> told me that...
> >>>>> "PB" == Petr Baudis <pasky@ucw.cz> writes:
> 
> PB> ... git-diff-tree (in contrast to
> PB> git-diff-cache) won't take the pathspec as its trailing arguments,
> 
> PB> Junio, is there any specific reason for that, or is the end of
> PB> git-diff-tree argument list the right spot for the pathspec stuff?
> 
> Baffled.  Are you at Linus tip?
> 
> Linus correctly decided that diff-tree does not have to call
> diffcore_pathspec(), which may be what confused you to make the
> comment "... in contrast to git-diff-cache) won't take ...".
> But it does not call it only because it does not need to.  It
> filters the filepairs itself on the input side using the
> trailing arguments; since diffcore_pathspec filters as the first
> one in the chain as the input filter, calling it from diff-tree
> would not cull anything further.

Ok, so this is what you get when you mix: sleepiness, performing only
mental experiments not verified in practice, and inattentive reading of
the code.

I'm sorry for bothering. Instruct yourself from my bad example, please.
:-)

-- 
				Petr "Pasky" Baudis
Stuff: http://pasky.or.cz/
C++: an octopus made by nailing extra legs onto a dog. -- Steve Taylor

Re: Problem with cg-diff <file>

[Junio C Hamano]

>>>>> "PB" == Petr Baudis <pasky@ucw.cz> writes:

PB> Ok, so this is what you get when you mix: sleepiness, performing only
PB> mental experiments not verified in practice, and inattentive reading of
PB> the code.

PB> I'm sorry for bothering. Instruct yourself from my bad example, please.
PB> :-)

If you forbid people to ask for help when the person who is
asked might feel the question groundless or based on "only
mental experiments not verified in practice and inattentive
reading of the code", the value to have a community diminishes.

We ask questions and ask for help because we know others know
more about things we do not know offhand, not necessarily
because we would not ever be able to figure them out ourselves.

If you know somebody else would know the answer immediately for
something that may take you a day or so to figure out, asking
for help is the right thing to do --- your time is better spent
on what you do best (e.g. improving Cogito).  I should not feel
bothered by your questions, and I am certainly not feeling
bothered at all (well, at least until seeing the last sentence,
and wondering what you really meant ;-)).

Always glad to be of help.

Re: Problem with cg-diff <file>

[Petr Baudis]

Dear diary, on Mon, May 30, 2005 at 09:54:13AM CEST, I got a letter
where Junio C Hamano <junkio@cox.net> told me that...
> >>>>> "PB" == Petr Baudis <pasky@ucw.cz> writes:
> 
> PB> Ok, so this is what you get when you mix: sleepiness, performing only
> PB> mental experiments not verified in practice, and inattentive reading of
> PB> the code.
> 
> PB> I'm sorry for bothering. Instruct yourself from my bad example, please.
> PB> :-)
> 
> If you forbid people to ask for help when the person who is
> asked might feel the question groundless or based on "only
> mental experiments not verified in practice and inattentive
> reading of the code", the value to have a community diminishes.
> 
> We ask questions and ask for help because we know others know
> more about things we do not know offhand, not necessarily
> because we would not ever be able to figure them out ourselves.
> 
> If you know somebody else would know the answer immediately for
> something that may take you a day or so to figure out, asking
> for help is the right thing to do --- your time is better spent
> on what you do best (e.g. improving Cogito).

Well, you definitively have a point. I just felt guilty that if I
would've spent ten seconds actually testing it, you wouldn't have to
spent ten minutes (or who knows how long) replying to my silly question.

> I should not feel bothered by your questions, and I am certainly not
> feeling bothered at all (well, at least until seeing the last
> sentence, and wondering what you really meant ;-)).

I meant you all on the mailing list, not you specifically. English is
sometimes annoying. ;-) (And I was just trying to be funny, in my usual
awful way.)

> Always glad to be of help.

Thanks. :-)

-- 
				Petr "Pasky" Baudis
Stuff: http://pasky.or.cz/
C++: an octopus made by nailing extra legs onto a dog. -- Steve Taylor