From: Stavros Ntentos <firstname.lastname@example.org>
Cc: email@example.com, firstname.lastname@example.org, email@example.com,
firstname.lastname@example.org, email@example.com, firstname.lastname@example.org,
Subject: Re: [PATCH v2 1/1] pathspec: warn for a no-glob entry that contains `**`
Date: Fri, 26 Mar 2021 17:43:50 +0200 [thread overview]
Message-ID: <email@example.com> (raw)
I think I have managed to address everyone's questions in this thread.
Hopefully everything will be addressed by this, and the patch that will soon follow:
> You may notice that I would call both of those latter two "globbing",
> but only one of them is triggered by the ":(glob)" magic. :)
And that's why I argued a DWIM was warranted here
(https://firstname.lastname@example.org/; it's more clear
in Junio's quote, but you could of course read the original).
I would want to be considered as an above-average git user, and I still was
oblivious to the fact that `**` required a `:(glob)` from the command line.
Especially since `.gitignore` files are treated differently
(i.e. don't require `:(glob)`)
I cannot / don't want to argue to "do it" or not, as I think my experience
is not substantial enough to navigate such argument, and go from concept
That being said, if there was a cli.pathspec.wildmatch flag,
I would've had it set to true in my global gitconfig.
Ofc I could set `GIT_GLOB_PATHSPECS=1`, but I am not that happy to
force it by setting it in a shell rc, instead of where it belongs.
> I am not sure everything after the "the sky:" you wrote is what you
> meant to write. Without being marked with a "glob" magic, a
> wildcard character in a pattern matches even a slash, so these two
> git ls-files 'Documentation**v2.txt'
> git ls-files 'Documentation*v2.txt'
> give the identical result and there is nothing about "surrounded by
> slashes" involved in it.
It seems you are right - in `fnmatch` mode, `:!*.gitignore` would've
served me right (and avoided the whole thread).
If you think that it's just my (i.e. few people's) lacking of understanding
`fnmatch` glob, then we can drop this.
However, given the disparity between `.gitignore` syntax and pathspec given
from the command line (from my limited POV meaningless/confusing), and your
(plural) arguments of backwards compatibility, I think we can draw the line
at an advice been acceptable.
Unless I see other points (like the warn vs advice), or pure C code-review
points, I am getting the impression that this thread will not move forward.
As I don't know how reviews usually happen here, and lacking some other medium
of discussion, I would appreciate (at some point) an explicit go/no-go
decision - to save everyone's time.
Being unfamiliar to the project, and being who I am, I prefer explicit vs implicit points.
Navigating an unknown process from top to bottom is pressure enough to me :-D
> seems to tell me that the "zero-or-more directories" magic happens
> only when /**/ form is used, not when two asterisks are placed next
> to each other in a random context.
Not exactly: ** needs to either start or end with a slash (or both) for its
wildmatch behavior. I can try to make the code more explicit, but of course
the code (and tests) will increase - and then maybe disproportionately to
the otherwise intended-to-be small change (since DWIM is not wanted).
next prev parent reply other threads:[~2021-03-26 15:44 UTC|newest]
Thread overview: 43+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-02-26 20:44 Pathspec does not accept / does not warn for specific pathspecs Σταύρος Ντέντος
2021-02-26 23:27 ` Junio C Hamano
2021-03-25 10:22 ` [PATCH v1 0/1] pathspec: warn for a no-glob entry that contains `**` Σταύρος Ντέντος
2021-03-25 10:22 ` [PATCH v1 1/1] " Σταύρος Ντέντος
2021-03-25 11:00 ` Bagas Sanjaya
2021-03-25 11:04 ` Bagas Sanjaya
2021-03-25 16:09 ` Stavros Ntentos
2021-03-25 20:09 ` Junio C Hamano
2021-03-25 23:36 ` [PATCH v2 0/1] " Σταύρος Ντέντος
2021-03-25 23:36 ` [PATCH v2 1/1] " Σταύρος Ντέντος
2021-03-26 0:41 ` Junio C Hamano
2021-03-26 1:32 ` Eric Sunshine
2021-03-26 3:02 ` Junio C Hamano
2021-03-26 5:13 ` Jeff King
2021-03-26 15:43 ` Stavros Ntentos [this message]
2021-03-26 15:48 ` [PATCH v3 1/2] " Stavros Ntentos
2021-03-26 15:48 ` [PATCH v3 2/2] pathspec: convert no-glob warn to advice Stavros Ntentos
2021-03-26 2:40 ` [RFC PATCH v1 0/1] pathspec: warn: long and short forms are incompatible Σταύρος Ντέντος
2021-03-26 2:40 ` [RFC PATCH v1 1/2] " Σταύρος Ντέντος
2021-03-26 5:28 ` Jeff King
2021-03-26 16:16 ` Stavros Ntentos
2021-03-27 9:41 ` Jeff King
2021-03-27 18:36 ` Junio C Hamano
2021-03-28 15:44 ` Stavros Ntentos
2021-03-26 2:40 ` [RFC PATCH v1 2/2] fixup! " Σταύρος Ντέντος
2021-03-26 8:14 ` Bagas Sanjaya
2021-03-26 15:55 ` Stavros Ntentos
2021-03-28 15:26 ` [RFC PATCH v1 3/3] squash! " Stavros Ntentos
2021-03-26 2:44 ` [RFC PATCH v1 0/1] " Σταύρος Ντέντος
2021-03-26 2:44 ` [RFC PATCH v1] " Σταύρος Ντέντος
2021-04-03 12:26 ` [PATCH v3] pathspec: advice: " Stavros Ntentos
2021-04-04 7:19 ` Junio C Hamano
2021-04-11 15:07 ` Σταύρος Ντέντος
2021-04-11 19:10 ` Junio C Hamano
2021-04-11 20:53 ` Σταύρος Ντέντος
2021-03-28 15:45 ` [PATCH v2] " Stavros Ntentos
2021-03-28 19:12 ` Junio C Hamano
2021-03-30 17:37 ` Junio C Hamano
2021-03-30 19:05 ` Stavros Ntentos
2021-03-30 19:17 ` Stavros Ntentos
2021-03-30 20:36 ` Junio C Hamano
2021-04-03 12:48 ` [PATCH v3] " Stavros Ntentos
2021-04-03 12:51 ` [PATCH v4] " Stavros Ntentos
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).