Re: [PATCH 2/2] git-credential-netrc: accept gpg option
- From
Luis Marsano <luis.marsano@gmail.com>
- Date
- May 14, 2018, 10:20 UTC
- Message-ID
- <20180514062024.5819@lmm-notebook.localdomain>
- In-Reply-To
- <xmqq36yzkf3v.fsf@gitster-ct.c.googlers.com>
Junio C Hamano <gitster@pobox.com> wrote:
Show 8 quoted lines
> Luis Marsano <luis.marsano@gmail.com> writes: > > > git-credential-netrc was hardcoded to decrypt with 'gpg' regardless of the gpg.program option > > this now uses the gpg command option if set, else, the gpg.program option set in the git repository or global configuration, else defaults to 'gpg' > > for git-credential-netrc > > These lines are way overlong. Wrap at around 72-78 cols, perhaps. > Complete each sentence with a full-stop.
Thanks, corrected this in the updated patch https://public-inbox.org/git/20180512091728.4931-3-luis.marsano@gmail.com/.
Show 15 quoted lines
> > - use Git.pm for repository and global option queries > > - add -g|--gpg command option & document it in command usage > > - test repository & command options > > - support unicode > > There are other changes that are not explained/justified here, I > think. > > - Instead of ALLCAPS as a placeholder for a command line argument in > the help text, use <placeholder>, because doing so is better due > to such and such reasons. > > I think it is good to consistently do so, but it is unclear why > ALLCAPS is bad and <placeholder> is better. That needs to be > explained.
Not necessarily bad, but the reason was to conform with Documentation/CodingGuidelines. The updated commit message now explains this.
Show 6 quoted lines
> - Replace three-dots in the help text with U+2026 to punish those > who are still using unicode-inapable terminal in this century. > > I do not think this part of the patch is a good idea at all, but > perhaps I misunderstood the reason behind this change you had in > mind (as you did not explain it in the proposed log message).
The original intent for this was semantics & accessibility: screenreaders more reliably interpret … than ... as ellipsis, and I imagine other assistive technology would, too. However, after research, I've learned there are better supported ways to go about it, so I'm retracting that change.
The updated patch https://public-inbox.org/git/20180512091728.4931-3-luis.marsano@gmail.com/ should reflect all corrections. Thank you for the feedback, and please let me know of further issues.