Please explain avoiding history simplifications when diffing merges

Please explain avoiding history simplifications when diffing merges

[Magnus Holmgren]

Hi (and thanks for a great free tool).

QGit was bitten by 
https://github.com/git/git/commit/0dec322d31db3920872f43bdd2a7ddd282a5be67

It looks like passing --simplify-merges to override the default solves the 
problem, but I still want to ask here because I'm not sure I fully understand 
the reasoning:

> the default history simplification would remove merge commits from 
> consideration if the file "path" matched the second parent.

I guess that means that even if there was a difference relative to the first 
parent, which you wanted to see, the merge commit would be omitted if there's 
no diff relative to the second parent? That's fine, and history simplification 
isn't trivial, but now every merge commit ever is included in the commit log 
of an individual file, even if that file was only added somewhere in the 
middle and never changed since. Is that really desirable? Shouldn't it be 
possible to include only merge commits that are not TREESAME to *all* parents?

Thanks,
-- 
Magnus Holmgren
./¯\_/¯\. Milient
(also holmgren@debian.org)

Re: Please explain avoiding history simplifications when diffing merges

[Magnus Holmgren]

Friday, 8 September 2023 11:09:20 CEST, I wrote
> QGit was bitten by
> https://github.com/git/git/commit/0dec322d31db3920872f43bdd2a7ddd282a5be67

Maybe I should link to the QGit issue:
https://github.com/tibirna/qgit/issues/129

> It looks like passing --simplify-merges to override the default solves the
> problem, but I still want to ask here because I'm not sure I fully
> understand
> the reasoning:
> > the default history simplification would remove merge commits from
> > consideration if the file "path" matched the second parent.

As I wrote at the above URL, I realized that the old git log output without --
simplify-merges and the output with --simplify-merges aren't quite the same. 
The old output indeed omits some interesting merge commits, which may explain 
why the change was made, but git log --simplify-merges does include them, so 
it seems a reasonable default to me.

However, QGit has a problem: git log --diff-merges=separate includes a 
separate diff for each parent, but only for each parent with differences 
compared to the merge commit, *and* there's no custom format placeholder for 
the current parent, only for the list of parents (%P/%p). How should one go 
about adding that? I figure the format_commit_context struct in pretty.c needs 
another field.

-- 
Magnus Holmgren
./¯\_/¯\. Milient

Re: Please explain avoiding history simplifications when diffing merges

[Bagas Sanjaya]

[-- Attachment #1: Type: text/plain, Size: 1616 bytes --]

On Fri, Sep 15, 2023 at 05:10:28PM +0200, Magnus Holmgren wrote:
> Friday, 8 September 2023 11:09:20 CEST, I wrote
> > QGit was bitten by
> > https://github.com/git/git/commit/0dec322d31db3920872f43bdd2a7ddd282a5be67
> 
> Maybe I should link to the QGit issue:
> https://github.com/tibirna/qgit/issues/129
> 
> > It looks like passing --simplify-merges to override the default solves the
> > problem, but I still want to ask here because I'm not sure I fully
> > understand
> > the reasoning:
> > > the default history simplification would remove merge commits from
> > > consideration if the file "path" matched the second parent.
> 
> As I wrote at the above URL, I realized that the old git log output without --
> simplify-merges and the output with --simplify-merges aren't quite the same. 
> The old output indeed omits some interesting merge commits, which may explain 
> why the change was made, but git log --simplify-merges does include them, so 
> it seems a reasonable default to me.

Can you provide examples?

> 
> However, QGit has a problem: git log --diff-merges=separate includes a 
> separate diff for each parent, but only for each parent with differences 
> compared to the merge commit, *and* there's no custom format placeholder for 
> the current parent, only for the list of parents (%P/%p). How should one go 
> about adding that? I figure the format_commit_context struct in pretty.c needs 
> another field.

What are you trying to accomplish with your proposed formatting verbs?

Confused...

-- 
An old man doll... just what I always wanted! - Clara

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 228 bytes --]

Re: Please explain avoiding history simplifications when diffing merges

[Magnus Holmgren]

söndag 17 september 2023 11:58:14 CEST skrev  Bagas Sanjaya:
> On Fri, Sep 15, 2023 at 05:10:28PM +0200, Magnus Holmgren wrote:
> > Friday, 8 September 2023 11:09:20 CEST, I wrote
> > 
> > > QGit was bitten by
> > > https://github.com/git/git/commit/0dec322d31db3920872f43bdd2a7ddd282a5be
> > > 67
> > 
> > Maybe I should link to the QGit issue:
> > https://github.com/tibirna/qgit/issues/129
> > 
> > > It looks like passing --simplify-merges to override the default solves
> > > the
> > > problem, but I still want to ask here because I'm not sure I fully
> > > understand
> > > 
> > > the reasoning:
> > > > the default history simplification would remove merge commits from
> > > > consideration if the file "path" matched the second parent.
> > 
> > As I wrote at the above URL, I realized that the old git log output
> > without -- simplify-merges and the output with --simplify-merges aren't
> > quite the same. The old output indeed omits some interesting merge
> > commits, which may explain why the change was made, but git log
> > --simplify-merges does include them, so it seems a reasonable default to
> > me.
> 
> Can you provide examples?

git init git-test; cd git-test
echo foo > test; git add test; git commit -m Initial
git branch testbranch
git branch irrelevantbranch
echo foobar > test; git commit -am Change
git switch testbranch
echo barfoo > test; git commit -am Diverge
git switch irrelevantbranch
echo whatever > otherfile; git add otherfile
git commit -m "Add other file"
git switch master
git merge -m "Merge branch 'testbranch'" -s ours testbranch
git merge -m "Merge branch 'irrelevantbranch' irrelevantbranch
git log test  # old git doesn't show either merge commit, new git shows both
git log --simplify-merges test  # both versions show the first merge commit
git log --diff-merges=separate -p --simplify-merges test # only in new git
# note that only one diff is generated because the other would be empty,
# although the man page says "Separate log entry and diff is generated for
# each parent."

> > However, QGit has a problem: git log --diff-merges=separate includes a
> > separate diff for each parent, but only for each parent with differences
> > compared to the merge commit, *and* there's no custom format placeholder
> > for the current parent, only for the list of parents (%P/%p). How should
> > one go about adding that? I figure the format_commit_context struct in
> > pretty.c needs another field.
> 
> What are you trying to accomplish with your proposed formatting verbs?

QGit builds annotated file histories by parsing git log output. To accomplish 
that, it needs to use a custom, more machine-readable format, and to handle 
merge commits correctly, it either needs one diff per parent in the order the 
parents are listed, or it needs to know which parent each diff is relative to.

Because old git log (without --simplify-merges) only included merge commits 
where the file in question is different to all parents, it didn't get confused 
by missing diffs - it was all or nothing - but some merge commits were missing 
in the file history view.

QGit should be rewritten to use libgit2, but regardless, I think all the 
information you can get with a standard format should be available to custom 
formats as well.

-- 
Magnus Holmgren
./¯\_/¯\. Milient