Rebasing guile-daemon branch onto master

[Sandeep Subramanian]

Hi all,

I was trying to rebase the guile-daemon branch onto the master branch
and I have some trouble doing it.

The interactive git rebase showed me that there are 932 commits to be
picked but most of them had no relation to the daemon code (Most commits
were package addition/update commits.) I also tried `git cherry master
guile-daemon`
and that too showed 932 "+ commits" and 0 "- commits".

The graph looks something like:

(78a5205) (2472f7a)
(master-HEAD)
*----------------*--------------------*-----------------------------*
\ (c4395e7) \
\ \
(e338c9b)
\
*-----------------------*-* (6dee54f - guile-daemon HEAD)
\ (36cc971)
/
*-----------------------*--------------------------------*
(654c8a7) (5e27bfc) (f5dfbaa)


A lot of commits on the guile daemon section of the branch are also found
between 78a5205 and 2472f7a. For example both c4395e7 and 5e27bfc
are the exact same patches and have the same patch-id. And almost all the
932 commits have a duplicate patch.

I think this is the reason why both rebase and cherry are showing a lot of
redundant commits.

To verify, I created an orphaned branch "orphan" from the 78a5205 commit and
and did `git cherry-pick 2472f7a..guile-daemon`. Then a `git cherry master
orphan`
revealed 30 "+ commits" and 902 "- commits". `git diff guile-daemon orphan`
was empty
and the 30 "+ commits" corresponded to the 30 commits by Caleb Ristvedt.

Shall I now rebase by picking only these 30 commits?

I don't know if what I am doing is right. I could use some guidance and
criticism now.

--
Sandeep (uniq10)

[Sandeep Subramanian]

Post by Sandeep Subramanian
I was trying to rebase the guile-daemon branch onto the master branch
and I have some trouble doing it.
The interactive git rebase showed me that there are 932 commits to be
picked but most of them had no relation to the daemon code (Most commits
were package addition/update commits.) I also tried `git cherry master
guile-daemon`
and that too showed 932 "+ commits" and 0 "- commits".
(78a5205) (2472f7a)
(master-HEAD)

*----------------*--------------------*-----------------------------*

Post by Sandeep Subramanian
\ (c4395e7) \
\ \
(e338c9b)
\
*-----------------------*-* (6dee54f - guile-daemon HEAD)
\ (36cc971)
/

*-----------------------*--------------------------------*

Post by Sandeep Subramanian
(654c8a7) (5e27bfc)

(f5dfbaa)

Post by Sandeep Subramanian
A lot of commits on the guile daemon section of the branch are also found
between 78a5205 and 2472f7a. For example both c4395e7 and 5e27bfc
are the exact same patches and have the same patch-id. And almost all the
932 commits have a duplicate patch.
I think this is the reason why both rebase and cherry are showing a lot of
redundant commits.
To verify, I created an orphaned branch "orphan" from the 78a5205 commit and
and did `git cherry-pick 2472f7a..guile-daemon`. Then a `git cherry master
orphan`
revealed 30 "+ commits" and 902 "- commits". `git diff guile-daemon orphan`
was empty
and the 30 "+ commits" corresponded to the 30 commits by Caleb Ristvedt.
Shall I now rebase by picking only these 30 commits?
I don't know if what I am doing is right. I could use some guidance and
criticism now.

Fixed the graph formatting.

(78a5205) (2472f7a) (master-HEAD)
*---------*--------*--------------------*
\ (c4395e7) \
\ \ (e338c9b)
\ *--------------*-*(guile-daemon-HEAD)
\ (36cc971) /
\ /
*---------------*--------------*
(654c8a7) (5e27bfc) (f5dfbaa)

--
Sandeep (uniq10)

[Leo Famulari]

Post by Sandeep Subramanian
Shall I now rebase by picking only these 30 commits?
I don't know if what I am doing is right. I could use some guidance and
criticism now.

Yes, I think that picking only those 30 commits is the right thing to
do.

Please let us know how it goes :)

[Sandeep Subramanian]

I rebased the guile-daemon branch. I had to remove a bunch of
duplicate commits. The diff between the master merged to guile
daemon and the rebased guile daemon is empty and so I think
it went well.

You can find the result at: https://github.com/uniq10/guix

--
uniq10

Post by Leo Famulari

Post by Sandeep Subramanian
Shall I now rebase by picking only these 30 commits?
I don't know if what I am doing is right. I could use some guidance and
criticism now.

Yes, I think that picking only those 30 commits is the right thing to
do.
Please let us know how it goes :)

[Ludovic Courtès]

Hello,

Post by Sandeep Subramanian
I rebased the guile-daemon branch. I had to remove a bunch of
duplicate commits. The diff between the master merged to guile
daemon and the rebased guile daemon is empty and so I think
it went well.
You can find the result at: https://github.com/uniq10/guix

Awesome! Did you have troubles building it or running the test suite?
(See <https://www.gnu.org/software/guix/manual/html_node/Running-the-Test-Suite.html>.)

Thanks,
Ludo’.

[Sandeep Subramanian]

Hi,

Post by Ludovic Courtès
Awesome! Did you have troubles building it or running the test suite?
(See <

https://www.gnu.org/software/guix/manual/html_node/Running-the-Test-Suite.html

Post by Ludovic Courtès
.)

The normal `guix environment guix` gave me errors while building
the guile-daemon branch, but `guix environment guix --ad-hoc guile-sqlite3`
fixed those and gave a successful build.

A few tests are failing and I am looking into it.

--
uniq10

[Ludovic Courtès]

Heya,

A few tests related to call-with-container are failing. I have made only
little
progress on this end since my university finals are going on. I hope to fix
them over the weekend.

OK, let us know how it goes—the end of the community bonding period is
approaching.

Thanks,
Ludo’.