git stash: add --index-only, or --keep-index should not stash index?

DISCUSSION Piotr Krukowiecki 3 participants 7 messages Jun 15, 2016

Share & tools

Piotr Krukowiecki

author

2016-06-15 22:50 UTC #0 lore raw

Hi,

I wanted to do something like "Testing partial commits" described in git-stash documentation (see end of mail for reference). I think this is a common scenario: you start working on some feature, then discover a bug, start fixing it, but realize it needs more work. So now you have two features that needs more work (both are not ready for committing).

The documentation says to use --keep-index in this case.

The problem is that --keep-index leaves changes in index intact, but at the same time it saves them in stash. So if I edit those changes I'm likely to get conflicts when applying the stash.

For example:

$ git init && echo a > a && git add . && git commit -m a
$ echo x > a && git add a && git stash save --keep-index
$ echo y > a && git add a && git commit -m y
$ git stash pop
Auto-merging a
CONFLICT (content): Merge conflict in a

Maybe --keep-index should not stash staged changes? This would fix this problem. And I can't think of a situation when would want to stash changes and at the same time keep them.

If --keep-index works correctly maybe a new option, for example --index-only (or --cached-only?) could be introduced?

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

Testing partial commits::

You can use git stash save --keep-index when you want to make two or more commits out of the changes in the work tree, and you want to test each change before committing:
+
----------------------------------------------------------------

# ... hack hack hack ...
$ git add --patch foo            # add just first part to the index
$ git stash save --keep-index    # save all other changes to the stash
$ edit/build/test first part
$ git commit -m 'First part'     # commit fully tested change
$ git stash pop                  # prepare to work on all other changes
# ... repeat above five steps until one commit remains ...
$ edit/build/test remaining parts
$ git commit foo -m 'Remaining parts'
----------------------------------------------------------------

Signature · 1 line

Piotr Krukowiecki

Neal Kreitzinger

2016-06-15 22:50 UTC #1 lore raw

Quoted 40 lines · Piotr Krukowiecki

Hi,
I wanted to do something like "Testing partial commits" described in git-stash documentation (see end of mail for reference). I think this is a common scenario: you start working on some feature, then discover a bug, start fixing it, but realize it needs more work. So now you have two features that needs more work (both are not ready for committing).
The documentation says to use --keep-index in this case.
The problem is that --keep-index leaves changes in index intact, but at the same time it saves them in stash. So if I edit those changes I'm likely to get conflicts when applying the stash.
For example:
$ git init&&  echo a>  a&&  git add .&&  git commit -m a
$ echo x>  a&&  git add a&&  git stash save --keep-index
$ echo y>  a&&  git add a&&  git commit -m y
$ git stash pop
Auto-merging a
CONFLICT (content): Merge conflict in a
Maybe --keep-index should not stash staged changes? This would fix this problem. And I can't think of a situation when would want to stash changes and at the same time keep them.
If --keep-index works correctly maybe a new option, for example --index-only (or --cached-only?) could be introduced?
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Testing partial commits::
You can use git stash save --keep-index when you want to make two or more commits out of the changes in the work tree, and you want to test each change before committing:
+
----------------------------------------------------------------

Quoted 12 lines

# ... hack hack hack ...
$ git add --patch foo            # add just first part to the index
$ git stash save --keep-index    # save all other changes to the stash
$ edit/build/test first part
$ git commit -m 'First part'     # commit fully tested change
$ git stash pop                  # prepare to work on all other changes
# ... repeat above five steps until one commit remains ...
$ edit/build/test remaining parts
$ git commit foo -m 'Remaining parts'
----------------------------------------------------------------

behind-the-scenes, git stash saves your working tree as a commit and your index as another commit. "git-stash apply" and "git-stash pop" only apply your stashed-index if you do "git-stash-apply --index". The default is to only apply your stashed-work-tree. You can create new branches from your stashes with "git-stash branch". You may find that much better to deal with for managing your work. Stashes aren't really intended to be the primary way to manage your work, but instead are a supplement. Branches are a better main tool for managing work. You can create a branch from your stash and when the branch is ready you can merge it into your other branch.

v/r,
neal

Piotr Krukowiecki

author

2016-06-15 22:50 UTC #2 lore raw

Quoted 57 lines · Neal Kreitzinger

Quoted 44 lines · Piotr Krukowiecki
Hi,
I wanted to do something like "Testing partial commits" described in git-stash documentation (see end of mail for reference). I think this is a common scenario: you start working on some feature, then discover a bug, start fixing it, but realize it needs more work. So now you have two features that needs more work (both are not ready for committing).
The documentation says to use --keep-index in this case.
The problem is that --keep-index leaves changes in index intact, but at the same time it saves them in stash. So if I edit those changes I'm likely
to get conflicts when applying the stash.
For example:
$ git init&&  echo a>  a&&  git add .&&  git commit -m a
$ echo x>  a&&  git add a&&  git stash save --keep-index
$ echo y>  a&&  git add a&&  git commit -m y
$ git stash pop
Auto-merging a
CONFLICT (content): Merge conflict in a
Maybe --keep-index should not stash staged changes? This would fix this problem. And I can't think of a situation when would want to stash changes
and at the same time keep them.
If --keep-index works correctly maybe a new option, for example --index-only
(or --cached-only?) could be introduced?
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Testing partial commits::
You can use git stash save --keep-index when you want to make two or more commits out of the changes in the work tree, and you want to test each change before committing:
+
----------------------------------------------------------------
Quoted 12 lines
# ... hack hack hack ...
$ git add --patch foo            # add just first part to the index
$ git stash save --keep-index    # save all other changes to the stash
$ edit/build/test first part
$ git commit -m 'First part'     # commit fully tested change
$ git stash pop                  # prepare to work on all other changes
# ... repeat above five steps until one commit remains ...
$ edit/build/test remaining parts
$ git commit foo -m 'Remaining parts'
----------------------------------------------------------------

Quoted 9 lines

behind-the-scenes, git stash saves your working tree as a commit and your index as another commit. "git-stash apply" and "git-stash pop" only apply your stashed-index if you do "git-stash-apply --index". The default is to only apply your stashed-work-tree. You can create new branches from your stashes with "git-stash branch". You may find that much better to deal with for managing your work. Stashes aren't really intended to be the primary way to manage your work, but instead are a supplement. Branches are a better main tool for managing work. You can create a branch from your stash and when the branch is ready you can merge it into your other branch.

Thanks for explanation, I understand there is not much pressure on improving it.

But it still does not explain how leaving code in index (with --keep-index) while
still stashing it might be helpful?

I would understand the use of --index-only (I gave an example of use case), or even --workdir-only, but not --keep-index. If I'm missing something please correct me.

Signature · 1 line

Piotr Krukowiecki

Neal Kreitzinger

2016-06-15 22:50 UTC #3 lore raw

Quoted 71 lines · Piotr Krukowiecki

Quoted 57 lines · Neal Kreitzinger
Quoted 44 lines · Piotr Krukowiecki
Hi,
I wanted to do something like "Testing partial commits" described in git-stash documentation (see end of mail for reference). I think this is a common scenario: you start working on some feature, then discover a bug, start fixing it, but realize it needs more work. So now you have two features that needs more work (both are not ready for committing).
The documentation says to use --keep-index in this case.
The problem is that --keep-index leaves changes in index intact, but at the same time it saves them in stash. So if I edit those changes I'm likely
to get conflicts when applying the stash.
For example:
$ git init&&    echo a>    a&&    git add .&&    git commit -m a
$ echo x>    a&&    git add a&&    git stash save --keep-index
$ echo y>    a&&    git add a&&    git commit -m y
$ git stash pop
Auto-merging a
CONFLICT (content): Merge conflict in a
Maybe --keep-index should not stash staged changes? This would fix this problem. And I can't think of a situation when would want to stash changes
and at the same time keep them.
If --keep-index works correctly maybe a new option, for example --index-only
(or --cached-only?) could be introduced?
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Testing partial commits::
You can use git stash save --keep-index when you want to make two or more commits out of the changes in the work tree, and you want to test each change before committing:
+
----------------------------------------------------------------
Quoted 12 lines
# ... hack hack hack ...
$ git add --patch foo            # add just first part to the index
$ git stash save --keep-index    # save all other changes to the stash
$ edit/build/test first part
$ git commit -m 'First part'     # commit fully tested change
$ git stash pop                  # prepare to work on all other changes
# ... repeat above five steps until one commit remains ...
$ edit/build/test remaining parts
$ git commit foo -m 'Remaining parts'
----------------------------------------------------------------
Quoted 9 lines
behind-the-scenes, git stash saves your working tree as a commit and your index as another commit. "git-stash apply" and "git-stash pop" only apply your stashed-index if you do "git-stash-apply --index". The default is to only apply your stashed-work-tree. You can create new branches from your stashes with "git-stash branch". You may find that much better to deal with for managing your work. Stashes aren't really intended to be the primary way to manage your work, but instead are a supplement. Branches are a better main tool for managing work. You can create a branch from your stash and when the branch is ready you can merge it into your other branch.
Thanks for explanation, I understand there is not much pressure on improving it.

Quoted 9 lines

But it still does not explain how leaving code in index (with --keep-index) while
still stashing it might be helpful?
I would understand the use of --index-only (I gave an example of use case), or even --workdir-only, but not --keep-index. If I'm missing something please correct me.

Keep in mind that a key to his workflow is "git-add --patch" method (which you did not include in your example). The workflow of the manpage assumes you will have to resolve conflicts in such a case. If you look at the "Pulling into a dirty tree" workflow of the same manpage its obvious that this workflow will almost always get conflicts, but they don't explicitly include conflict resolution in the example. You could resolve the conflicts for "Testing partial commits" in this way:

Neal's Version of Testing partial commits::
----------------------------------------------------------------

# ... hack hack hack ...
$ git add --patch foo            # add just first part to the index
$ git stash save --keep-index    # save all other changes to the stash
$ edit/build/test first part
$ git commit -m 'First part'     # commit fully tested change
$ git stash pop                  # prepare to work on all other changes
   (conflict resolution)
   $ edit conflicted file        # keep new version of the hunk already
                                 # committed in prior iteration
   $ git reset HEAD -- foo       # clear conflict from index
   $ git reset HEAD -- foo       # (again) reset index to match HEAD
# ... repeat above five (or eight) steps until one commit remains ...
$ edit/build/test remaining parts
$ git commit foo -m 'Remaining parts'
----------------------------------------------------------------

Maybe there's a better way to do this. Hope this helps.

v/r,
neal

Piotr Krukowiecki

author

2016-06-15 22:50 UTC #4 lore raw

Quoted 90 lines · Neal Kreitzinger

Quoted 82 lines · Piotr Krukowiecki
Quoted 58 lines · Neal Kreitzinger
Quoted 44 lines · Piotr Krukowiecki
Hi,
I wanted to do something like "Testing partial commits" described in git-stash documentation (see end of mail for reference). I think this is a common scenario: you start working on some feature, then discover a bug, start fixing it, but realize it needs more work. So now you have two features that needs more work (both are not ready for committing).
The documentation says to use --keep-index in this case.
The problem is that --keep-index leaves changes in index intact, but at the same time it saves them in stash. So if I edit those changes I'm likely
to get conflicts when applying the stash.
For example:
$ git init&&    echo a>    a&&    git add .&&    git commit -m a
$ echo x>    a&&    git add a&&    git stash save --keep-index
$ echo y>    a&&    git add a&&    git commit -m y
$ git stash pop
Auto-merging a
CONFLICT (content): Merge conflict in a
Maybe --keep-index should not stash staged changes? This would fix this problem. And I can't think of a situation when would want to stash changes
and at the same time keep them.
If --keep-index works correctly maybe a new option, for example --index-only
(or --cached-only?) could be introduced?
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Testing partial commits::
You can use git stash save --keep-index when you want to make two or more commits out of the changes in the work tree, and you want to test each change before committing:
+
----------------------------------------------------------------
Quoted 12 lines
# ... hack hack hack ...
$ git add --patch foo            # add just first part to the index
$ git stash save --keep-index    # save all other changes to the stash
$ edit/build/test first part
$ git commit -m 'First part'     # commit fully tested change
$ git stash pop                  # prepare to work on all other changes
# ... repeat above five steps until one commit remains ...
$ edit/build/test remaining parts
$ git commit foo -m 'Remaining parts'
----------------------------------------------------------------
Quoted 13 lines
behind-the-scenes, git stash saves your working tree as a commit and your index as another commit. "git-stash apply" and "git-stash pop" only apply
your stashed-index if you do "git-stash-apply --index". The default is to
only apply your stashed-work-tree. You can create new branches from your stashes with "git-stash branch". You may find that much better to deal with
for managing your work. Stashes aren't really intended to be the primary way to manage your work, but instead are a supplement. Branches are a better main tool for managing work. You can create a branch from your stash
and when the branch is ready you can merge it into your other branch.
Thanks for explanation, I understand there is not much pressure on improving it.
But it still does not explain how leaving code in index (with --keep-index) while
still stashing it might be helpful?
Quoted 7 lines
I would understand the use of --index-only (I gave an example of use case),
or even --workdir-only, but not --keep-index. If I'm missing something please
correct me.

Quoted 2 lines

Keep in mind that a key to his workflow is "git-add --patch" method (which you did not include in your example).

I did - in my first mail.

Quoted 5 lines

The workflow of the manpage assumes
you will have to resolve conflicts in such a case. If you look at the "Pulling into a dirty tree" workflow of the same manpage its obvious that this workflow will almost always get conflicts, but they don't explicitly include conflict resolution in the example.

ok, for this workflow, when you stash away your changes to do a pull, you can't do anything to avoid conflicts.

I mean this case:

Quoted 21 lines

You could resolve the conflicts
for "Testing partial commits" in this way:

Neal's Version of Testing partial commits::
----------------------------------------------------------------

# ... hack hack hack ...
$ git add --patch foo            # add just first part to the index
$ git stash save --keep-index    # save all other changes to the stash
$ edit/build/test first part
$ git commit -m 'First part'     # commit fully tested change
$ git stash pop                  # prepare to work on all other changes
  (conflict resolution)
  $ edit conflicted file        # keep new version of the hunk already
                                # committed in prior iteration
  $ git reset HEAD -- foo       # clear conflict from index
  $ git reset HEAD -- foo       # (again) reset index to match HEAD
# ... repeat above five (or eight) steps until one commit remains ...
$ edit/build/test remaining parts
$ git commit foo -m 'Remaining parts'
----------------------------------------------------------------

Quoted 1 line

Maybe there's a better way to do this. Hope this helps.

How about following workflow?

   # ... hack hack hack ...
   $ git add --patch foo
   $ git stash save --keep-index
   $ build/test first part         # note I have removed the "edit" part

There are two possibilities now:

1. You're happy with the result - you commit your changes:

   $ git commit -m 'First part'
   $ git stash pop

Important: there will be no conflicts in this case, and committed changes won't be reverted/applied/conflicted, because you have not edited them!

I wasn't fully aware of this before (looks like I'm still new to git).

2. You're not happy with the result - maybe something does not build and you need to stage more changes, or maybe fix is not working yet.

You should first pop your stashed changes! This way you'll avoid conflicts, you'll be able to stage/modify other changes. So the next steps are like this:

   $ git stash pop
   $ edit/add -p
   $ git stash save --keep-index

and now you're back to the "build/test" part.

Quoted 5 lines

(conflict resolution)

  $ edit conflicted file        # keep new version of the hunk already
                                # committed in prior iteration
  $ git reset HEAD -- foo       # clear conflict from index
  $ git reset HEAD -- foo       # (again) reset index to match HEAD

I'm not really sure why/how this works :/

Signature · 1 line

Piotr Krukowiecki

Neal Kreitzinger

2016-06-15 22:50 UTC #5 lore raw

Quoted 168 lines · Piotr Krukowiecki

Quoted 90 lines · Neal Kreitzinger
Quoted 82 lines · Piotr Krukowiecki
Quoted 58 lines · Neal Kreitzinger
Quoted 44 lines · Piotr Krukowiecki
Hi,
I wanted to do something like "Testing partial commits" described in git-stash documentation (see end of mail for reference). I think this is a common scenario: you start working on some feature, then discover a bug, start fixing it, but realize it needs more work. So now you have two features that needs more work (both are not ready for committing).
The documentation says to use --keep-index in this case.
The problem is that --keep-index leaves changes in index intact, but at the same time it saves them in stash. So if I edit those changes I'm likely
to get conflicts when applying the stash.
For example:
$ git init&&      echo a>      a&&      git add .&&      git commit -m a
$ echo x>      a&&      git add a&&      git stash save --keep-index
$ echo y>      a&&      git add a&&      git commit -m y
$ git stash pop
Auto-merging a
CONFLICT (content): Merge conflict in a
Maybe --keep-index should not stash staged changes? This would fix this problem. And I can't think of a situation when would want to stash changes
and at the same time keep them.
If --keep-index works correctly maybe a new option, for example --index-only
(or --cached-only?) could be introduced?
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Testing partial commits::
You can use git stash save --keep-index when you want to make two or more commits out of the changes in the work tree, and you want to test each change before committing:
+
----------------------------------------------------------------
Quoted 12 lines
# ... hack hack hack ...
$ git add --patch foo            # add just first part to the index
$ git stash save --keep-index    # save all other changes to the stash
$ edit/build/test first part
$ git commit -m 'First part'     # commit fully tested change
$ git stash pop                  # prepare to work on all other changes
# ... repeat above five steps until one commit remains ...
$ edit/build/test remaining parts
$ git commit foo -m 'Remaining parts'
----------------------------------------------------------------
Quoted 13 lines
behind-the-scenes, git stash saves your working tree as a commit and your index as another commit. "git-stash apply" and "git-stash pop" only apply
your stashed-index if you do "git-stash-apply --index". The default is to
only apply your stashed-work-tree. You can create new branches from your stashes with "git-stash branch". You may find that much better to deal with
for managing your work. Stashes aren't really intended to be the primary way to manage your work, but instead are a supplement. Branches are a better main tool for managing work. You can create a branch from your stash
and when the branch is ready you can merge it into your other branch.
Thanks for explanation, I understand there is not much pressure on improving it.
But it still does not explain how leaving code in index (with --keep-index) while
still stashing it might be helpful?
Quoted 7 lines
I would understand the use of --index-only (I gave an example of use case),
or even --workdir-only, but not --keep-index. If I'm missing something please
correct me.
Quoted 2 lines
Keep in mind that a key to his workflow is "git-add --patch" method (which you did not include in your example).
I did - in my first mail.
Quoted 5 lines
The workflow of the manpage assumes
you will have to resolve conflicts in such a case. If you look at the "Pulling into a dirty tree" workflow of the same manpage its obvious that this workflow will almost always get conflicts, but they don't explicitly include conflict resolution in the example.
ok, for this workflow, when you stash away your changes to do a pull, you can't do anything to avoid conflicts.
I mean this case:
Quoted 21 lines
You could resolve the conflicts
for "Testing partial commits" in this way:
Neal's Version of Testing partial commits::
----------------------------------------------------------------
# ... hack hack hack ...
$ git add --patch foo            # add just first part to the index
$ git stash save --keep-index    # save all other changes to the stash
$ edit/build/test first part
$ git commit -m 'First part'     # commit fully tested change
$ git stash pop                  # prepare to work on all other changes
   (conflict resolution)
   $ edit conflicted file        # keep new version of the hunk already
                                 # committed in prior iteration
   $ git reset HEAD -- foo       # clear conflict from index
   $ git reset HEAD -- foo       # (again) reset index to match HEAD
# ... repeat above five (or eight) steps until one commit remains ...
$ edit/build/test remaining parts
$ git commit foo -m 'Remaining parts'
----------------------------------------------------------------
Quoted 1 line
Maybe there's a better way to do this. Hope this helps.
How about following workflow?
    # ... hack hack hack ...
    $ git add --patch foo
    $ git stash save --keep-index
    $ build/test first part         # note I have removed the "edit" part
There are two possibilities now:
1. You're happy with the result - you commit your changes:
    $ git commit -m 'First part'
    $ git stash pop
Important: there will be no conflicts in this case, and committed changes won't be reverted/applied/conflicted, because you have not edited them!
I wasn't fully aware of this before (looks like I'm still new to git).
2. You're not happy with the result - maybe something does not build and you need to stage more changes, or maybe fix is not working yet.
You should first pop your stashed changes! This way you'll avoid conflicts, you'll be able to stage/modify other changes. So the next steps are like this:
    $ git stash pop
    $ edit/add -p
    $ git stash save --keep-index
and now you're back to the "build/test" part.

Quoted 8 lines

Quoted 5 lines

(conflict resolution)

   $ edit conflicted file        # keep new version of the hunk already
                                 # committed in prior iteration
   $ git reset HEAD -- foo       # clear conflict from index
   $ git reset HEAD -- foo       # (again) reset index to match HEAD

I'm not really sure why/how this works :/

example:
(1) I've added this code:
if me then
echo me
endif
if you then
echo you
endif
if them then
echo them
endif

(2) via git-add --patch I only add the first "if" to the index:
if me then
echo me
endif

(3) git stash --keep-index:
puts all 3 "if" statements in stash, but keeps only the first "if" in my working-tree and index.

(4) I test the first "If" and decide to modify it:
if me or mine then
echo "all mine"
endif

(5) I add and commit this change.

(6) git stash pop: generates conflict on the first "if".

(7) I edit the file and keep the new version of the first "if" and delete the old version of the first "if":
if me or mine then
echo "all mine"
endif
if you then
echo you
endif
if them then
echo them
endif

(8) git status still shows this file as conflicted, because I haven't "added" it yet. However, I don't want to add the whole thing, but just a patch of the second "if". I'm not ready for the 3rd "if" yet. I can't do git-add --patch on a conflicted item. So I clear the conflict with "git reset HEAD --foo".

(9) the first git reset only cleared the conflict -- it did not actually reset the foo in index to match the foo in HEAD. therefore, git status still shows foo as being modified. therefore, i do another git-reset to reset the index foo to match the HEAD foo.

(10) now i can do git-add --patch and add the second "if":
if me or mine then
echo "all mine"
endif
if you then
echo you
endif

(repeat steps 3 thru 10)

If you are new to git then maybe some clarification about the index will be helpful. In a nutshell, the index is the staging area in-between your working-tree and the object-store. Changes in your working-tree first go to the index and then to the commit (object store). The index holds what is going to be committed. It also is the go-between in the opposite direction. In order to get an object from the object-store to your work-tree it goes from the object-store to your index and then to your worktree. when you checkout out a commit from the object-store it goes into your index and then from the index into the working tree. The main thing to remember is that you have to put stuff in the index before you can commit it. Another thing to remember is that in order to put stuff from the object-store into your working-tree, git first has to put in in the index. the index is the middle-man between the work-tree and object-store, AND between the object-store and the work-tree.

v/r,
neal

Junio C Hamano

2016-06-15 22:50 UTC #6 lore raw

In reply to Piotr Krukowiecki

Quoted 9 lines · Piotr Krukowiecki

How about following workflow?

   # ... hack hack hack ...
   $ git add --patch foo
   $ git stash save --keep-index
   $ build/test first part         # note I have removed the "edit" part

There are two possibilities now:

Quoted 8 lines

1. You're happy with the result - you commit your changes:
   $ git commit -m 'First part'
   $ git stash pop
Important: there will be no conflicts in this case, and committed changes won't be reverted/applied/conflicted, because you have not edited them!

Exactly; I won't take _any_ credit for --keep-index (I wasn't involved deeply in "git stash"), but the above matches my understanding of its primary intended use case. For this to work well, the stash should record the index that is going to be committed and the work tree you had before stashing.

Quoted 8 lines

2. You're not happy with the result - maybe something does not build and you need to stage more changes, or maybe fix is not working yet.
You should first pop your stashed changes! This way you'll avoid conflicts, you'll be able to stage/modify other changes. So the next steps are like this:

Quoted 5 lines

   $ git stash pop
   $ edit/add -p
   $ git stash save --keep-index
and now you're back to the "build/test" part.

Yes.