Re: "loco" after ottava

[Jean Abou Samra]

Le 03/12/2022 à 14:46, Thomas Morley a écrit :
> I didn't know about -S
> Thanks.
>
> Alas, this command extended as
> $ git log -S "check-internal-types" --oneline --author="Han-Wen"
> prints for me:
> $ git log -S "check-internal-types" --oneline --author="Han-Wen"
> 264570e68e Documentation: remove documentation for internal options
> e435a88122 Remove en/contributor/build-notes.texi
> 8ad3efa86f output-distance: make self test pass again:
> 51117d43f1 Don't Graham's compile logs.
> cf7f4f614b (tag: release/2.11.33-1) Make test generation generic.
> 9291834601 replace SCM_ASSERT_TYPE with LY_ASSERT_TYPE and friends
> acc8482ad0 set LILYPOND_BOOK_LILYPOND_FLAGS explicitly for test.
> Removes PDF for tests.
> b7da29a7aa * scm/framework-ps.scm (output-preview-framework): use new name.
> 29ebcd1876 * scm/framework-ps.scm (output-preview-framework): use new name.
>
> I.e. this one is not found:
>
> > commit a1dead391ed016a8aefd1be7fb62946e094b4535
> > Author: Han-Wen Nienhuys <hanwenn@gmail.com>
> > Date:   Sat Feb 26 12:19:32 2022 +0100
> >
> >       Add internal/external distinction to -d flags
> whereas
>
> > commit 264570e68ea13b06c1b9dd88c491948979926aa1
> > Author: Han-Wen Nienhuys <hanwenn@gmail.com>
> > Date:   Sun Apr 3 12:45:51 2022 +0200
> >
> >       Documentation: remove documentation for internal options
> is indeed found.



-S only shows commits that changed the number of occurrences
of the pattern, which is useful to find out when something
was added or deleted. To find all the times a line containing
the pattern was changed, use -G instead.



> Anyway, "Add internal/external distinction to -d flags" is
> https://gitlab.com/lilypond/lilypond/-/merge_requests/1231
>
>
> That said, I researched a bit more about "check-internal-types".
> and found we have:
>
> lilypond --help
> lilypond -dhelp
> and actually
> lilypond -dhelp-internal
>
> The latter is undocumented as intended by one of the mentioned
> patches, but shows "check-internal-types"
>
> I agree with separating internal/external options and not explicitley
> documenting internal ones.
> But not to document at all how to reach the internal ones is too much, imho.


You can see it in "lilypond -dhelp".



> I'd vote for documenting help-internal in "Application Usage".  With a
> short hint, something at the lines of:
>
> help-internal bool
>    If bool is #t, display options for internal usage. Default: #f.
>
> Opinions?



If the purpose of internal options is to be hidden
from the user, why add them back in the manual?

What do you find -dcheck-internal-types useful for
as a user? *If* it has functionality useful for
non-developers, that functionality possibly shouldn't
be internal in the first place.

Best,
Jean

OpenPGP_signature
Description: OpenPGP digital signature