On Tuesday 05 December 2017 12:14 AM, Junio C Hamano wrote:
Kaartic Sivaraam <kaartic.sivar...@gmail.com> writes:

Stepping back a bit, the mild suspicion above says

     $ git checkout HEAD^0
     ... do things ...
     $ git checkout -b temp
     ... do more things ...
     $ git checkout -B @{-1}

that creates a new branch whose name is 40-hex of a commit that
happens to be where we started the whole dance *is* a bug.  No sane
user expects that to happen, and the last step "checkout -B @{-1}"
should result in an error instead [*1*].

I was wondering if "git check-ref-format --branch @{-1}", when used
in place of "checkout -B @{-1}" in the above sequence,

I guess you mean '... "git checkout -B $(git check-ref-format --branch
@{-1}", when used in place of "git checkout -B @{-1}" ...' ?

No you guessed wrong.  I was (and am) wondering if the last step in
the following sequence should fail.
     $ git checkout HEAD^0
     ... do things ...
     $ git checkout -b temp
     ... do more things ...
     $ git check-ref-format --branch @{-1}


Ok. Now I get what you say.


And I am leaning towards saying that it is a bug that it does not
fail; @{-1} is a detached HEAD and not a concrete branch name in
this case,

It seems your thought is similar to the following thought that I expressed in [1],

-- 8< --

I thought this the other way round. Rather than letting the callers
error out when @{-N} didn't expand to a branch name, I thought we
should not be expanding @{-N} syntax for "check-ref-format --branch" at
all to make a "stronger guarantee" that the result is "always" a valid
branch name. Then I thought it might be too restrictive and didn't
mention it. So, I dunno.

-- >8 --



so "check-ref-format --branch" should at least notice
and say that it is a request that may lead to a nonsense next step
(which is to create a branch with that 40-hex name).


Makes sense, this should at least be noted in the Documentation. Is that what you had in mind too or do you expect 'check-ref-format' to do something else too?


[1]: https://public-inbox.org/git/1511880237.10193.5.ca...@gmail.com/

---
Kaartic

Reply via email to