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

Re: [PATCH v2 5/7] reflog: improve and update documentation

From
Michael Haggerty <mhagger@alum.mit.edu>
Date
Mar 3, 2015, 11:35 UTC
Message-ID
<54F59C7C.5040705@alum.mit.edu>
In-Reply-To
<xmqqfv9n9htg.fsf@gitster.dls.corp.google.com>
On 03/02/2015 11:04 PM, Junio C Hamano wrote:
Show 15 quoted lines
> Michael Haggerty <mhagger@alum.mit.edu> writes:
> 
>> +Reference logs, or "reflogs", record when the tips of branches and
>> +other references were updated in the local repository. Reflogs are
>> +useful in various Git commands, to specify the old value of a
>> +reference. For example, `HEAD@{2}` means "where HEAD used to be two
>> +moves ago", `master@{one.week.ago}` means "where master used to point
>> +to one week ago", and so on. See linkgit:gitrevisions[7] for more
>> +details.
> 
> Looks very good, especially the part that mentions "in the local
> repository".  It seems to be a common novice misunderstanding what
> `master@{one.week.ago}` means, and it might be beneficial to be more
> verbose by saying "where master used to point to one week ago in
> this local repository".
Yes, that's good. I will change it.
Show 14 quoted lines
>> +The "expire" subcommand prunes older reflog entries. Entries older
>> +than `expire` time, or entries older than `expire-unreachable` time
>> +and not reachable from the current tip, are removed from the reflog.
>> +This is typically not used directly by end users -- instead, see
>> +linkgit:git-gc[1].
>> +
>> +The "delete" subcommand deletes single entries from the reflog. Its
>> +argument must be an _exact_ entry (e.g. "`git reflog delete
>> +master@{2}`").
> 
> Just like "expire", "delete" should be accompanied by the same
> "typically not".  I do not think it is even worth mentioning that it
> exists merely as an implementation detail for likgit:git-stash[1]
> and for no other reason.
OK, will change.
Show 18 quoted lines
>> +Options for `expire` and/or `delete`
>> +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> 
> I think this started from a hope that these two share many, but
> looking at the result I notice the shared ones are a tiny and
> trivial minority.  It probably makes sense to retitle this section
> "Options for expire" (and remove "For the expire command only"), and
> add an "Options for delete" section immediately after it that looks
> like:
> 
> 	Options for `delete`
>         ~~~~~~~~~~~~~~~~~~~~
> 
> 	--updateref::
>         --rewrite::
>         --dry-run::
>         	The `delete` command takes these options whose
>                 meanings are the same as those for `expire`.

Either way is a little bit awkward, because these two subcommands share 4 out of 8 of their options. But since "delete" is really quite useless to end users, I stuck its options in a separate section as you suggested, but inline in a paragraph, where they won't bother anybody.

Show 24 quoted lines
>> diff --git a/builtin/reflog.c b/builtin/reflog.c
>> index 49c64f9..dd68a72 100644
>> --- a/builtin/reflog.c
>> +++ b/builtin/reflog.c
>> @@ -13,9 +13,9 @@
>>   */
>>  
>>  static const char reflog_expire_usage[] =
>> -"git reflog expire [--verbose] [--dry-run] [--stale-fix] [--expire=<time>] [--expire-unreachable=<time>] [--all] <refs>...";
>> +"git reflog expire [--expire=<time>] [--expire-unreachable=<time>] [--rewrite] [--updateref] [--stale-fix] [--dry-run | -n] [--verbose] [--all] <refs>...";
>>  static const char reflog_delete_usage[] =
>> -"git reflog delete [--verbose] [--dry-run] [--rewrite] [--updateref] <refs>...";
>> +"git reflog delete [--rewrite] [--updateref] [--dry-run | -n] [--verbose] <refs>...";
>>  
>>  static unsigned long default_reflog_expire;
>>  static unsigned long default_reflog_expire_unreachable;
> 
> Thanks for being complete, but I sense that it may be time we
> switched to parse-options here, which gives us the help string for
> free.  Perhaps throw in a comment line before this hunk
> 
> 	/* NEEDSWORK: switch to parse-options */
> 
> or something to leave hint for other people?
OK.
Thanks for your review! A reroll will be coming soon.
Michael
-- 
Michael Haggerty
mhagger@alum.mit.edu
Previous: Junio C HamanoNext: Michael Haggerty
Message 10 of 13 in “Fix some problems with reflog expiration”
  1. 0/7 Fix some problems with reflog expirationMichael Haggerty, Mar 2, 2015
  2. 1/7 write_ref_sha1(): remove check for lock == NULLMichael Haggerty, Mar 2, 2015
  3. 2/7 write_ref_sha1(): Move write elision test to callersMichael Haggerty, Mar 2, 2015
  4. 3/7 lock_ref_sha1_basic(): do not set force_write for missing referencesMichael Haggerty, Mar 2, 2015
  5. 4/7 struct ref_lock: delete the force_write memberMichael Haggerty, Mar 2, 2015
  6. Junio C HamanoMar 2, 2015
  7. Michael HaggertyMar 3, 2015
  8. 5/7 reflog: improve and update documentationMichael Haggerty, Mar 2, 2015
  9. Junio C HamanoMar 2, 2015
  10. Michael HaggertyMar 3, 2015
  11. 6/7 reflog_expire(): ignore --updateref for symbolic referencesMichael Haggerty, Mar 2, 2015
  12. 7/7 reflog_expire(): never update a reference to null_sha1Michael Haggerty, Mar 2, 2015
  13. Junio C HamanoMar 2, 2015

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.