git/list[1] front-page[2] threads[3] people[4] search[5] about
 

Re: [GSOC][RFC] Heed core.bare from template config file when no command line override given, as a microproject.

From
Junio C Hamano <gitster@pobox.com>
Date
Jan 8, 2024, 17:32 UTC
Message-ID
<xmqqo7dvloiu.fsf@gitster.g>
In-Reply-To
<CY7M09XT547N.2OOTI5APX9RIX@gmail.com>
"Ghanshyam Thakkar" <shyamthakkar001@gmail.com> writes:
Show 10 quoted lines
>> Specifically, the commit that introduced the comment never wanted to
>> honor core.bare in the template.  I do not think I has core.bare in
>> mind when I wrote the comment, but I would have described it as the
>> same category as the repository format version, i.e. something you
>> would not want to copy, if I were pressed to clarify back then.
>
> Then I suppose this warrants updating the TODO comment in
> create_default_files(), which currently can be interpreted as this 
> being a unwanted behavior. And also amending the testcases which
> currently display this as knwon breakage.

I obviously agree with that, after saying that I suspect 0f7443bd comes from a misunderstanding ;-).

Show 14 quoted lines
>> If somebody wants to always create a bare repository by having
>> core.bare=true in their template and if we wanted to honor it (which
>> I am dubious of the value of, by the way), I would think the right
>> place to do so would be way before create_default_files() is called.
>> When running "git init [$DIR]", long before calling init_db(), we
>> decide if we are about to create a bare repository and either create
>> $DIR or $DIR/.git.  What is in the template, if we really wanted to
>> do so, should be read before that happens, no?
>
> That is what I proposed in my original email, after which I had a
> working solution which passed all the tests. That solution was indeed to
> check for core.bare in the template before we set GIT_DIR_ENVIRONMENT, 
> which subsequently creates either $DIR or $DIR/.git as you described 
> above. 

Yeah, if this were still in soon after 4f629539 was written, then such a change might have been a useful feature enhancement, but risk of breaking people (third-party tools) who use the same template to initialize both bare and non-bare repositories is there, so...

Thanks.
Previous: Ghanshyam ThakkarNext: Elijah Newren
Message 7 of 15 in “[GSOC][RFC] Heed core.bare from template config file when no command line override given, as a microproject.”
  1. Ghanshyam ThakkarJan 2, 2024
  2. Christian CouderJan 4, 2024
  3. Ghanshyam ThakkarJan 4, 2024
  4. Elijah NewrenJan 5, 2024
  5. Junio C HamanoJan 5, 2024
  6. Ghanshyam ThakkarJan 6, 2024
  7. Junio C HamanoJan 8, 2024
  8. Elijah NewrenJan 19, 2024
  9. setup: clarify TODO comment about ignoring core.bareGhanshyam Thakkar, Feb 29, 2024
  10. Junio C HamanoFeb 29, 2024
  11. Ghanshyam ThakkarFeb 29, 2024
  12. setup: remove unnecessary variableGhanshyam Thakkar, Mar 4, 2024
  13. Junio C HamanoMar 4, 2024
  14. Ghanshyam ThakkarMar 4, 2024
  15. Junio C HamanoMar 4, 2024

Read the whole thread, see it on lore, or plain text.

$ cat FOOTERMessages come from the public archive at lore.kernel.org/git, fetched every hour. The front page is chosen and written each morning by an AI editor and can be wrong; the threads themselves are the record. About and API. For agents: an MCP server at https://gitlist.dev/mcp, and any thread, story or person page as Markdown by adding .md to its URL (or sending Accept: text/markdown). Details in /llms.txt.