* [PATCH v2] branch: rework the descriptions of rename and copy operations
@ 2024-02-16 12:44 Dragan Simic
2024-02-16 19:59 ` Junio C Hamano
0 siblings, 1 reply; 5+ messages in thread
From: Dragan Simic @ 2024-02-16 12:44 UTC (permalink / raw)
To: git; +Cc: code, gitster, rjusto, spectral
Move the descriptions of the <oldbranch> and <newbranch> arguments to the
descriptions of the branch rename and copy operations, where they naturally
belong. Also, improve the descriptions of these two branch operations and,
for completeness, describe the outcomes of forced operations.
Describing the arguments together with their respective operations, instead
of describing them separately in a rather unfortunate attempt to squeeze more
meaning out of fewer words, flows much better and makes the git-branch(1)
man page significantly more usable.
The subsequent improvements shall continue this approach by either dissolving
as many sentences from the "Description" section into the "Options" section,
or by having those sentences converted into some kind of more readable and
better flowing prose, as already discussed and outlined. [1][2]
[1] https://lore.kernel.org/git/xmqqttmmlahf.fsf@gitster.g/T/#u
[2] https://lore.kernel.org/git/xmqq8r4zln08.fsf@gitster.g/T/#u
Signed-off-by: Dragan Simic <dsimic@manjaro.org>
---
Notes:
This patch was originally named "branch: clarify <oldbranch> and <newbranch>
terms further", submitted and discussed in another thread, [3] but the nature
of the patch has changed, causing the patch subject to be adjusted to match.
Consequently, the version 1 is effectively version 2 of the original patch.
The version 1 of the patch includes detailed feedback from Kyle and Junio,
who suggested moving/adding the argument descriptions to their respective
commands. This resulted in more significant changes to the contents of the
git-branch(1) man page, in an attempt to make it more readable.
The version 2 includes feedback from Kristoffer and Junio, by improving the
wording of the opening sentences in the descriptions of branch rename and
copy operations, and by mentioning the additional reflog entry created while
renaming a branch, which was omitted in the v1 by mistake.
[3] https://lore.kernel.org/git/e2eb777bca8ffeec42bdd684837d28dd52cfc9c3.1707136999.git.dsimic@manjaro.org/T/#u
Documentation/git-branch.txt | 51 ++++++++++++++++--------------------
1 file changed, 23 insertions(+), 28 deletions(-)
diff --git a/Documentation/git-branch.txt b/Documentation/git-branch.txt
index 0b0844293235..d52b5e8dbacd 100644
--- a/Documentation/git-branch.txt
+++ b/Documentation/git-branch.txt
@@ -72,16 +72,6 @@ the remote-tracking branch. This behavior may be changed via the global
overridden by using the `--track` and `--no-track` options, and
changed later using `git branch --set-upstream-to`.
-With a `-m` or `-M` option, <oldbranch> will be renamed to <newbranch>.
-If <oldbranch> had a corresponding reflog, it is renamed to match
-<newbranch>, and a reflog entry is created to remember the branch
-renaming. If <newbranch> exists, -M must be used to force the rename
-to happen.
-
-The `-c` and `-C` options have the exact same semantics as `-m` and
-`-M`, except instead of the branch being renamed, it will be copied to a
-new name, along with its config and reflog.
-
With a `-d` or `-D` option, `<branchname>` will be deleted. You may
specify more than one branch for deletion. If the branch currently
has a reflog then the reflog will also be deleted.
@@ -128,18 +118,31 @@ Note that 'git branch -f <branchname> [<start-point>]', even with '-f',
refuses to change an existing branch `<branchname>` that is checked out
in another worktree linked to the same repository.
--m::
---move::
- Move/rename a branch, together with its config and reflog.
-
--M::
+-m [<oldbranch>] <newbranch>::
+--move [<oldbranch>] <newbranch>::
+ Rename an existing branch `<oldbranch>` to `<newbranch>`; if left
+ unspecified, `<oldbranch>` defaults to the current branch. The
+ configuration variables for the `<oldbranch>` branch and its reflog
+ are also renamed appropriately to be used with `<newbranch>`. In
+ addition, a reflog entry is created to remember the branch renaming.
+ Renaming fails if branch `<newbranch>` already exists, but `-M`
+ or `--move --force` can be used to overwrite the contents of the
+ existing branch `<newbranch>` while renaming.
+
+-M [<oldbranch>] <newbranch>::
Shortcut for `--move --force`.
--c::
---copy::
- Copy a branch, together with its config and reflog.
-
--C::
+-c [<oldbranch>] <newbranch>::
+--copy [<oldbranch>] <newbranch>::
+ Copy an existing branch `<oldbranch>` to `<newbranch>`; if left
+ unspecified, `<oldbranch>` defaults to the current branch. The
+ configuration variables for the `<oldbranch>` branch and its reflog
+ are also copied appropriately to be used with `<newbranch>`.
+ Copying fails if branch `<newbranch>` already exists, but `-C`
+ or `--copy --force` can be used to overwrite the contents of the
+ existing branch `<newbranch>` while copying.
+
+-C [<oldbranch>] <newbranch>::
Shortcut for `--copy --force`.
--color[=<when>]::
@@ -311,14 +314,6 @@ superproject's "origin/main", but tracks the submodule's "origin/main".
given as a branch name, a commit-id, or a tag. If this
option is omitted, the current HEAD will be used instead.
-<oldbranch>::
- The name of an existing branch. If this option is omitted,
- the name of the current branch will be used instead.
-
-<newbranch>::
- The new name for an existing branch. The same restrictions as for
- <branchname> apply.
-
--sort=<key>::
Sort based on the key given. Prefix `-` to sort in descending
order of the value. You may use the --sort=<key> option
^ permalink raw reply related [flat|nested] 5+ messages in thread* Re: [PATCH v2] branch: rework the descriptions of rename and copy operations
2024-02-16 12:44 [PATCH v2] branch: rework the descriptions of rename and copy operations Dragan Simic
@ 2024-02-16 19:59 ` Junio C Hamano
2024-02-16 21:20 ` Dragan Simic
0 siblings, 1 reply; 5+ messages in thread
From: Junio C Hamano @ 2024-02-16 19:59 UTC (permalink / raw)
To: Dragan Simic; +Cc: git, code, rjusto, spectral
Dragan Simic <dsimic@manjaro.org> writes:
> Move the descriptions of the <oldbranch> and <newbranch> arguments to the
> descriptions of the branch rename and copy operations, where they naturally
> belong. Also, improve the descriptions of these two branch operations and,
> for completeness, describe the outcomes of forced operations.
>
> Describing the arguments together with their respective operations, instead
> of describing them separately in a rather unfortunate attempt to squeeze more
> meaning out of fewer words, flows much better and makes the git-branch(1)
> man page significantly more usable.
The intention to remove non-option from the OPTIONS enumeration,
and to explain <new> and <old> used as arguments to -m and -c where
these options are described, are both very good (heh, after all,
they are parts of what I envisioned to be the way to go in the
longer term ;-).
> overridden by using the `--track` and `--no-track` options, and
> changed later using `git branch --set-upstream-to`.
>
> -With a `-m` or `-M` option, <oldbranch> will be renamed to <newbranch>.
> -If <oldbranch> had a corresponding reflog, it is renamed to match
> -<newbranch>, and a reflog entry is created to remember the branch
> -renaming. If <newbranch> exists, -M must be used to force the rename
> -to happen.
> -
> -The `-c` and `-C` options have the exact same semantics as `-m` and
> -`-M`, except instead of the branch being renamed, it will be copied to a
> -new name, along with its config and reflog.
> -
> With a `-d` or `-D` option, `<branchname>` will be deleted. You may
> specify more than one branch for deletion. If the branch currently
> has a reflog then the reflog will also be deleted.
But the halfway modification to the description section in this
patch is not an improvement. It makes some options described there
while -m and -c are completely missing now, making the section
incomplete and coverage of the operating modes of the command
uneven.
> +-m [<oldbranch>] <newbranch>::
> +--move [<oldbranch>] <newbranch>::
> + Rename an existing branch `<oldbranch>` to `<newbranch>`; if left
> + unspecified, `<oldbranch>` defaults to the current branch. The
> + configuration variables for the `<oldbranch>` branch and its reflog
> + are also renamed appropriately to be used with `<newbranch>`. In
> + addition, a reflog entry is created to remember the branch renaming.
> + Renaming fails if branch `<newbranch>` already exists, but `-M`
> + or `--move --force` can be used to overwrite the contents of the
> + existing branch `<newbranch>` while renaming.
OK. This is way more readable than the previous attempts we made.
The description of the single failure mode still worries me (see my
previous message on this). Here is my attempt:
When the command fails due to an existing '<newbranch>', you
can use `-M` (or `--move --force`) to force overwriting it.
to hint that there may be other ways for the command to fail, and
hint that `-M` may not always resolve issues, but I do not know how
successful it is. I could add
Note that `-M <old> <new>` will not resolve an error if the
reason why `-m` fails is to protect the other worktree that
checks out (or otherwise uses) <old> and <new> points at a
different commit.
but we do not necessarily want to appear to be exhaustive here, so,
I dunno.
> +-M [<oldbranch>] <newbranch>::
> Shortcut for `--move --force`.
OK.
> +--copy [<oldbranch>] <newbranch>::
> + Copy an existing branch `<oldbranch>` to `<newbranch>`; if left
> + unspecified, `<oldbranch>` defaults to the current branch. The
> + configuration variables for the `<oldbranch>` branch and its reflog
> + are also copied appropriately to be used with `<newbranch>`.
> + Copying fails if branch `<newbranch>` already exists, but `-C`
> + or `--copy --force` can be used to overwrite the contents of the
> + existing branch `<newbranch>` while copying.
Exactly the same comment on "other failure modes" applies here.
> -<oldbranch>::
> - The name of an existing branch. If this option is omitted,
> - the name of the current branch will be used instead.
> -
> -<newbranch>::
> - The new name for an existing branch. The same restrictions as for
> - <branchname> apply.
> -
Removals of these lines are very pleasing ;-).
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: [PATCH v2] branch: rework the descriptions of rename and copy operations
2024-02-16 19:59 ` Junio C Hamano
@ 2024-02-16 21:20 ` Dragan Simic
2024-02-16 21:45 ` Junio C Hamano
0 siblings, 1 reply; 5+ messages in thread
From: Dragan Simic @ 2024-02-16 21:20 UTC (permalink / raw)
To: Junio C Hamano; +Cc: git, code, rjusto, spectral
Hello Junio,
On 2024-02-16 20:59, Junio C Hamano wrote:
> Dragan Simic <dsimic@manjaro.org> writes:
>
>> Move the descriptions of the <oldbranch> and <newbranch> arguments to
>> the
>> descriptions of the branch rename and copy operations, where they
>> naturally
>> belong. Also, improve the descriptions of these two branch operations
>> and,
>> for completeness, describe the outcomes of forced operations.
>>
>> Describing the arguments together with their respective operations,
>> instead
>> of describing them separately in a rather unfortunate attempt to
>> squeeze more
>> meaning out of fewer words, flows much better and makes the
>> git-branch(1)
>> man page significantly more usable.
>
> The intention to remove non-option from the OPTIONS enumeration,
> and to explain <new> and <old> used as arguments to -m and -c where
> these options are described, are both very good (heh, after all,
> they are parts of what I envisioned to be the way to go in the
> longer term ;-).
Yes, that's what I plan to work on after this patch is, hopefully,
accepted (see also below). My initial hope was that we'd define
the general outline for the completely reworked git-branch(1) even
further with this patch, which should in turn make the future work
more efficient. I think we're on a good way. :)
>> overridden by using the `--track` and `--no-track` options, and
>> changed later using `git branch --set-upstream-to`.
>>
>> -With a `-m` or `-M` option, <oldbranch> will be renamed to
>> <newbranch>.
>> -If <oldbranch> had a corresponding reflog, it is renamed to match
>> -<newbranch>, and a reflog entry is created to remember the branch
>> -renaming. If <newbranch> exists, -M must be used to force the rename
>> -to happen.
>> -
>> -The `-c` and `-C` options have the exact same semantics as `-m` and
>> -`-M`, except instead of the branch being renamed, it will be copied
>> to a
>> -new name, along with its config and reflog.
>> -
>> With a `-d` or `-D` option, `<branchname>` will be deleted. You may
>> specify more than one branch for deletion. If the branch currently
>> has a reflog then the reflog will also be deleted.
>
> But the halfway modification to the description section in this
> patch is not an improvement. It makes some options described there
> while -m and -c are completely missing now, making the section
> incomplete and coverage of the operating modes of the command
> uneven.
If I got it right, you'd prefer this patch not to be accepted
separately, but as part of the future series that would rework the
entire git-branch(1) man page? I'm fine with that as well.
>> +-m [<oldbranch>] <newbranch>::
>> +--move [<oldbranch>] <newbranch>::
>> + Rename an existing branch `<oldbranch>` to `<newbranch>`; if left
>> + unspecified, `<oldbranch>` defaults to the current branch. The
>> + configuration variables for the `<oldbranch>` branch and its reflog
>> + are also renamed appropriately to be used with `<newbranch>`. In
>> + addition, a reflog entry is created to remember the branch renaming.
>> + Renaming fails if branch `<newbranch>` already exists, but `-M`
>> + or `--move --force` can be used to overwrite the contents of the
>> + existing branch `<newbranch>` while renaming.
>
> OK. This is way more readable than the previous attempts we made.
>
> The description of the single failure mode still worries me (see my
> previous message on this). Here is my attempt:
>
> When the command fails due to an existing '<newbranch>', you
> can use `-M` (or `--move --force`) to force overwriting it.
>
> to hint that there may be other ways for the command to fail, and
> hint that `-M` may not always resolve issues, but I do not know how
> successful it is. I could add
Makes sense. It's intentionally a bit vague, but should work fine.
I'd just replace "the command" with "renaming", and avoid addressing
the reader directly.
> Note that `-M <old> <new>` will not resolve an error if the
> reason why `-m` fails is to protect the other worktree that
> checks out (or otherwise uses) <old> and <new> points at a
> different commit.
>
> but we do not necessarily want to appear to be exhaustive here, so,
> I dunno.
Huh-uh... I'm not sure that such an exhaustive explanation would
make it more clear to the majority of users. Perhaps it's better
to remain a bit vague, at least for now, and omit such details.
>> +-M [<oldbranch>] <newbranch>::
>> Shortcut for `--move --force`.
>
> OK.
>
>> +--copy [<oldbranch>] <newbranch>::
>> + Copy an existing branch `<oldbranch>` to `<newbranch>`; if left
>> + unspecified, `<oldbranch>` defaults to the current branch. The
>> + configuration variables for the `<oldbranch>` branch and its reflog
>> + are also copied appropriately to be used with `<newbranch>`.
>> + Copying fails if branch `<newbranch>` already exists, but `-C`
>> + or `--copy --force` can be used to overwrite the contents of the
>> + existing branch `<newbranch>` while copying.
>
> Exactly the same comment on "other failure modes" applies here.
Noted.
>> -<oldbranch>::
>> - The name of an existing branch. If this option is omitted,
>> - the name of the current branch will be used instead.
>> -
>> -<newbranch>::
>> - The new name for an existing branch. The same restrictions as for
>> - <branchname> apply.
>> -
>
> Removals of these lines are very pleasing ;-).
Oh yes, it's like a clear embodiment of making the current mess
a little bit smaller. :)
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: [PATCH v2] branch: rework the descriptions of rename and copy operations
2024-02-16 21:20 ` Dragan Simic
@ 2024-02-16 21:45 ` Junio C Hamano
2024-02-16 21:58 ` Dragan Simic
0 siblings, 1 reply; 5+ messages in thread
From: Junio C Hamano @ 2024-02-16 21:45 UTC (permalink / raw)
To: Dragan Simic; +Cc: git, code, rjusto, spectral
Dragan Simic <dsimic@manjaro.org> writes:
>> But the halfway modification to the description section in this
>> patch is not an improvement. It makes some options described there
>> while -m and -c are completely missing now, making the section
>> incomplete and coverage of the operating modes of the command
>> uneven.
>
> If I got it right, you'd prefer this patch not to be accepted
> separately, but as part of the future series that would rework the
> entire git-branch(1) man page? I'm fine with that as well.
Not necessarily. If you wanted to this this in multiple steps, we
can first whip the OPTIONS part into a good shape, and then fix the
DESCRIPTION part.
What we want to avoid (not limited to this topic) is to say "this
temporarily makes things worse here, but trust me it will eventually
become perfect". Removing only -m/-c from the description section
makes the description section worse than before the patch---we'd be
better off leaving the original as-is if we are not revamping the
entire section.
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: [PATCH v2] branch: rework the descriptions of rename and copy operations
2024-02-16 21:45 ` Junio C Hamano
@ 2024-02-16 21:58 ` Dragan Simic
0 siblings, 0 replies; 5+ messages in thread
From: Dragan Simic @ 2024-02-16 21:58 UTC (permalink / raw)
To: Junio C Hamano; +Cc: git, code, rjusto, spectral
On 2024-02-16 22:45, Junio C Hamano wrote:
> Dragan Simic <dsimic@manjaro.org> writes:
>
>>> But the halfway modification to the description section in this
>>> patch is not an improvement. It makes some options described there
>>> while -m and -c are completely missing now, making the section
>>> incomplete and coverage of the operating modes of the command
>>> uneven.
>>
>> If I got it right, you'd prefer this patch not to be accepted
>> separately, but as part of the future series that would rework the
>> entire git-branch(1) man page? I'm fine with that as well.
>
> Not necessarily. If you wanted to this this in multiple steps, we
> can first whip the OPTIONS part into a good shape, and then fix the
> DESCRIPTION part.
I'll think a bit more about it, to see what might be our best choice
moving forward.
> What we want to avoid (not limited to this topic) is to say "this
> temporarily makes things worse here, but trust me it will eventually
> become perfect". Removing only -m/-c from the description section
> makes the description section worse than before the patch---we'd be
> better off leaving the original as-is if we are not revamping the
> entire section.
The way you wrote this brought a smile to my face. :) I agree, making
things a bit worse while promising perfection later is rarely justified.
Perhaps only when some nasty bug has to be fixed ASAP.
^ permalink raw reply [flat|nested] 5+ messages in thread
end of thread, other threads:[~2024-02-16 21:58 UTC | newest]
Thread overview: 5+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2024-02-16 12:44 [PATCH v2] branch: rework the descriptions of rename and copy operations Dragan Simic
2024-02-16 19:59 ` Junio C Hamano
2024-02-16 21:20 ` Dragan Simic
2024-02-16 21:45 ` Junio C Hamano
2024-02-16 21:58 ` Dragan Simic
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).