Re: git merge banch w/ different submodule revision

Hi,

On Fri, Apr 27, 2018 at 3:37 AM, Middelschulte, Leif
[off-list ref] wrote:

Am Donnerstag, den 26.04.2018, 17:19 -0700 schrieb Elijah Newren:

quoted

On Thu, Apr 26, 2018 at 3:49 AM, Middelschulte, Leif
[off-list ref] wrote:

<snip>

quoted

quoted

Problem case: Merge either branch into the other

Expected behavior: Merge conflict.

Actual behavior: Auto merge without conflicts.

Note 1: A merge conflict does occur, if the sourced revisions do *not* have a linear history

Let me just note that I don't actually use submodules myself, and
rarely run across them, so as far as users expect submodules should
behave I may have to defer to others.  But it was particularly this
sentence of yours that caught my attention and got me to respond.  I
may have misunderstood which repository had the non-linear history,
but...

<snip>

quoted

But, there is some further smarts in that if either A or B point at
commits that contain the other in their history and both contain the
commit that O points at, then you can just do a fast-forward update to
the newest.

This particular paragraph, is relevant to your example; more details below.

quoted

You didn't tell us how the merge-base (cd5e1a5 from the diagram you
gave) differed in your example here between the two repositories.  In
fact, the non-linear case could have several merge-bases, in which
case they all become potentially relevant (as does their merge-bases
since at that point you'll trigger the recursive portion of
merge-recursive).  Giving us that info might help us point out what
happened, though if either the fast-forward logic comes into play or
the recursive logic gets in the mix, then we may need you to provide a
testcase (or access to the repo in question) in order to explain it
and/or determine if you've found a bug.

I placed two reositories here: https://gitlab.com/foss-contributions/git-examples/network/develop
The access should be public w/o login.

If you prefer the examples to be placed somewhere else, let me know.

So the only thing I see here is a single repository, which contains a
submodule with linear history.  (unless I was grabbing it wrong; I
just tried `git clone --recurse-submodules
https://gitlab.com/foss-contributions/git-examples`)  Do you also have
an example with non-linear history demonstrating your claim that it
behaves differently, for comparison?


Anyway, in this case you had both branches updating the submodule to
something newer (to a fast-forward update of what it previously was),
but one side advanced it further than the other side did (in
particular, to what turned out to be a fast-forward update of what the
other branch used).  That means the whole fast-forwarding logic of
commit 68d03e4a6e44 ("Implement automatic fast-forward merge for
submodules", 2010-07-07)) came into play.

I would expect that a different example involving non-linear history
would behave the same, if both sides update the submodule in a fashion
that is just fast-forwarding and one commit contains the other in its
history.  I'm curious if you have a counter example.