threads / discuss / 55363

git diff --text does not work during rebase for binary files

Subject: git diff --text does not work during rebase for binary files

## tl;dr

4 messages between Mar 21, 2021 and Mar 26, 2021.

replies: 3people: 3as markdown or json

Дилян Палаузов· Mar 21, 2021, 17:30 UTC · lore
Hello,
I create a small text file ending with \0 and add it to git:
In .gitattributes I specify:
*.txt diff merge=text text
I create several commits by changing the small text file.
The result is at https://github.com/dilyanpalauzov/git-diff-biinary .
I do now
  git rebase -i HEAD~2
and before the “second commit” type E to edit that commit.
I change the small file.
git add a.txt && git rebase --continue
git says:
[detached HEAD bc00e34] second commit
 Date: Sun Mar 21 19:16:15 2021 +0200
 1 file changed, 2 insertions(+), 1 deletion(-)
warning: Cannot merge binary files: a.txt (HEAD vs. 04c77de (third
commit))
Auto-merging a.txt
CONFLICT (content): Merge conflict in a.txt
error: could not apply 04c77de... third commit
Resolve all conflicts manually, mark them as resolved with
"git add/rm <conflicted_files>", then run "git rebase --continue".
You can instead skip this commit: run "git rebase --skip".
To abort and get back to the state before "git rebase", run "git rebase
--abort".
Could not apply 04c77de... third commit
'git diff' and 'git diff --text' show:

diff --cc a.txt index 7a61015,dc817ec..0000000

--- a/a.txt
+++ b/a.txt

The file does not contain <<HEAD===>>> markers, despite having
merge=text attribute.

HOW CAN I force git diff to show the differences and git merge to
include the <<<===>>> markers?

I have to say, that at different occassions git diff --text does work
ass expected (e.g. when I call git diff --cached --text).

I call now 
git rebase --abort
git log -p

git shows the differences in the small text file ending with \0.  So
apparently log reads the option, but diff does not.

Using git 2.30.2.

Greetings
  Дилян
Bagas Sanjaya· Mar 22, 2021, 05:53 UTC · re: Дилян Палаузов · lore

Re: git diff --text does not work during rebase for binary files

On 22/03/21 00.30, Дилян Палаузов wrote:
> Hello,
> 
> I create a small text file ending with \0 and add it to git:
Did you mean `echo -e "\0" >> a.txt`?
Show 63 quoted lines
> In .gitattributes I specify:
> 
> *.txt diff merge=text text
> 
> I create several commits by changing the small text file.
> 
> The result is at https://github.com/dilyanpalauzov/git-diff-biinary .
> 
> I do now
> 
>    git rebase -i HEAD~2
> and before the “second commit” type E to edit that commit.
> 
> I change the small file.
> 
> git add a.txt && git rebase --continue
> 
> git says:
> 
> [detached HEAD bc00e34] second commit
>   Date: Sun Mar 21 19:16:15 2021 +0200
>   1 file changed, 2 insertions(+), 1 deletion(-)
> warning: Cannot merge binary files: a.txt (HEAD vs. 04c77de (third
> commit))
> Auto-merging a.txt
> CONFLICT (content): Merge conflict in a.txt
> error: could not apply 04c77de... third commit
> Resolve all conflicts manually, mark them as resolved with
> "git add/rm <conflicted_files>", then run "git rebase --continue".
> You can instead skip this commit: run "git rebase --skip".
> To abort and get back to the state before "git rebase", run "git rebase
> --abort".
> Could not apply 04c77de... third commit
> 
> 
> 'git diff' and 'git diff --text' show:
> 
> diff --cc a.txt
> index 7a61015,dc817ec..0000000
> --- a/a.txt
> +++ b/a.txt
> 
> The file does not contain <<HEAD===>>> markers, despite having
> merge=text attribute.
> 
> HOW CAN I force git diff to show the differences and git merge to
> include the <<<===>>> markers?
> 
> I have to say, that at different occassions git diff --text does work
> ass expected (e.g. when I call git diff --cached --text).
> 
> I call now
> git rebase --abort
> git log -p
> 
> git shows the differences in the small text file ending with \0.  So
> apparently log reads the option, but diff does not.
> 
> Using git 2.30.2.
> 
> Greetings
>    Дилян
> 
Does `file a.txt` shows it as being binary file? I do the similar and it just says "data".
-- 
An old man doll... just what I always wanted! - Clara
Дилян Палаузов· Mar 22, 2021, 08:45 UTC · re: Bagas Sanjaya · lore

Re: git diff --text does not work during rebase for binary files

Hello,
I mean:

git init echo -e "AB RA\nCADABRA\0" > a.txt && git add a.txt && git commit - m"Initial" echo -e "AB2RA\nCADABRA\0" > a.txt && git add a.txt && git commit - m"Second" echo -e "AB3RA\nCADABRA\0" > a.txt && git add a.txt && git commit - m"Third"

git log -pa   show the diffs and handles the file as text.
git rebase -i HEAD~2

I EDIT the “Second” commit (the one before “Third”, not the second in the editor).

master|REBASE 1/2$ echo -e "AB4RA\nCADABRA\0" > a.txt && git add a.txt && git rebase --continue 

[detached HEAD 78408d2] Second
 Date: Mon Mar 22 10:35:57 2021 +0200
 1 file changed, 0 insertions(+), 0 deletions(-)
warning: Cannot merge binary files: a.txt (HEAD vs. 149b67e (Third))
Auto-merging a.txt
CONFLICT (content): Merge conflict in a.txt
error: could not apply 149b67e... Third
Resolve all conflicts manually, mark them as resolved with
"git add/rm <conflicted_files>", then run "git rebase --continue".
You can instead skip this commit: run "git rebase --skip".
To abort and get back to the state before "git rebase", run "git rebase
--abort".
Could not apply 149b67e... Third
master|REBASE 2/2$
And now `git diff` and `git diff --text` both print:

diff --cc a.txt index ee95164,d7db2f0..0000000 Binary files differ

But git diff --text should show how the files differ.
Greetings
  Дилян
On Mon, 2021-03-22 at 12:53 +0700, Bagas Sanjaya wrote:
Show 77 quoted lines
> 
> 
> On 22/03/21 00.30, Дилян Палаузов wrote:
> > Hello,
> > 
> > I create a small text file ending with \0 and add it to git:
> 
> Did you mean `echo -e "\0" >> a.txt`?
> 
> > In .gitattributes I specify:
> > 
> > *.txt diff merge=text text
> > 
> > I create several commits by changing the small text file.
> > 
> > The result is at https://github.com/dilyanpalauzov/git-diff-biinary .
> > 
> > I do now
> > 
> >    git rebase -i HEAD~2
> > and before the “second commit” type E to edit that commit.
> > 
> > I change the small file.
> > 
> > git add a.txt && git rebase --continue
> > 
> > git says:
> > 
> > [detached HEAD bc00e34] second commit
> >   Date: Sun Mar 21 19:16:15 2021 +0200
> >   1 file changed, 2 insertions(+), 1 deletion(-)
> > warning: Cannot merge binary files: a.txt (HEAD vs. 04c77de (third
> > commit))
> > Auto-merging a.txt
> > CONFLICT (content): Merge conflict in a.txt
> > error: could not apply 04c77de... third commit
> > Resolve all conflicts manually, mark them as resolved with
> > "git add/rm <conflicted_files>", then run "git rebase --continue".
> > You can instead skip this commit: run "git rebase --skip".
> > To abort and get back to the state before "git rebase", run "git
> > rebase
> > --abort".
> > Could not apply 04c77de... third commit
> > 
> > 
> > 'git diff' and 'git diff --text' show:
> > 
> > diff --cc a.txt
> > index 7a61015,dc817ec..0000000
> > --- a/a.txt
> > +++ b/a.txt
> > 
> > The file does not contain <<HEAD===>>> markers, despite having
> > merge=text attribute.
> > 
> > HOW CAN I force git diff to show the differences and git merge to
> > include the <<<===>>> markers?
> > 
> > I have to say, that at different occassions git diff --text does work
> > ass expected (e.g. when I call git diff --cached --text).
> > 
> > I call now
> > git rebase --abort
> > git log -p
> > 
> > git shows the differences in the small text file ending with \0.  So
> > apparently log reads the option, but diff does not.
> > 
> > Using git 2.30.2.
> > 
> > Greetings
> >    Дилян
> > 
> 
> Does `file a.txt` shows it as being binary file? I do the similar and
> it just says "data".
>
Jeff King· Mar 26, 2021, 11:31 UTC · re: Дилян Палаузов · lore

Re: git diff --text does not work during rebase for binary files

On Mon, Mar 22, 2021 at 10:45:39AM +0200, Дилян Палаузов wrote:
Show 5 quoted lines
> And now `git diff` and `git diff --text` both print:
> 
> diff --cc a.txt
> index ee95164,d7db2f0..0000000
> Binary files differ

I think the "combined diff" code isn't ready to handle the "--text" flag.

An easier reproduction than rebase is to just do a merge with a conflict:

  git init
  printf 'base\0' >file && git add file && git commit -m base
  printf 'one\0' >file && git commit -am one
  git checkout -b side HEAD^
  printf 'two\0' >file && git commit -am two
  git merge master

Now "git diff --text" will try to show a combined diff, but it doesn't work. Likewise if you complete the merge:

  printf 'new\0' >file && git commit -a --no-edit

then "git show --text" will not show a binary combined diff, even though "git show --text HEAD^" will happily show a normal two-endpoint diff.

Curiously old versions of Git would show broken binary combined diffs, until 4d5f347199 (combine-diff: handle binary files as binary, 2011-05-23). So the binary detection there looks mostly sane, but it's just not respecting the "--text" flag.

Something like this helps a bit:
diff --git a/combine-diff.c b/combine-diff.c
index 06635f91bc..940c5c43ff 100644
--- a/combine-diff.c
+++ b/combine-diff.c
@@ -1128,6 +1128,8 @@ static void show_patch_diff(struct combine_diff_path *elem, int num_parent,
 
 	if (textconv)
 		is_binary = 0;
+	else if (opt->flags.text)
+		is_binary = 0;
 	else if (userdiff->binary != -1)
 		is_binary = userdiff->binary;
 	else {

but I don't think that's the full story. It looks like the combined-diff
code is too eager to treat the content as NUL-terminated strings. A
regular diff shows the NULs:

  $ git show --format= --text HEAD^  | cat -A
  diff --git a/file b/file$
  index d249428..a9382c3 100644$
  --- a/file$
  +++ b/file$
  @@ -1 +1 @@$
  -base^@$
  \ No newline at end of file$
  +two^@$
  \ No newline at end of file$

but the combined diff doesn't:

  $ git show --format= --text HEAD | cat -A
  diff --cc file$
  index a9382c3,f33a432..c984a04$
  --- a/file$
  +++ b/file$
  @@@ -1,1 -1,1 +1,1 @@@$
  - two$
   -one$
  ++new$

So I suspect the code is simply not read to handle the NULs, and the
current behavior of ignoring --text is better than producing truncated
or bogus output until those deeper problems are fixed. We could issue a
warning that --text was ignored, I suppose, but that doesn't really
solve your problem. :)

-Peff

← back to recent threads