git.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: "Ævar Arnfjörð Bjarmason" <avarab@gmail.com>
To: Elia Pinto <gitter.spiros@gmail.com>
Cc: git@vger.kernel.org
Subject: Re: [PATCH] git-curl-compat.h: addition of all symbols defined by curl
Date: Tue, 15 Mar 2022 20:30:21 +0100	[thread overview]
Message-ID: <220315.86pmmndmre.gmgdl@evledraar.gmail.com> (raw)
In-Reply-To: <20220315131638.924669-1-gitter.spiros@gmail.com>


On Tue, Mar 15 2022, Elia Pinto wrote:

> This file was produced from a modified version of symbols.pl
> (https://github.com/curl/curl/blob/master/docs/libcurl/symbols.pl) and
> by manually adding the previous comments describing the dates of release
> of some curl versions not currently reported in the symbols-in-versions.
>
> To do this the symbols are listed in the order defined in the file
> symbols-in-versions rather than as they were previously inserted based
> on release dates.
>
> Most of these symbols are not used by git today. However, inserting
> them all starting from an automatic tool makes it largely unnecessary
> to update this file and therefore reduces the possibility
> of introducing possible errors in the future.
> [...]
>  1 file changed, 2875 insertions(+), 69 deletions(-)

*gulp* :)

> +#define LIBCURL_HAS(x) \
> +  (defined(x ## _FIRST) && (x ## _FIRST <= LIBCURL_VERSION_NUM) && \
> +   (!defined(x ## _LAST) || ( x ## _LAST >= LIBCURL_VERSION_NUM)))

As the recent author of this file I think there's certainly space to
improve it, e.g. perhaps this macro & the change Junio suggests
downthread, but...

> +#define CURLALTSVC_H1_FIRST 0x074001 /* Added in 7.64.1 */
> +#define GIT_CURL_HAVE_CURLALTSVC_H1 \

[snip ~3k lines lines]

I'd really prefer for us not to go down this route. If you look at the
patches I did where I removed the version comparisons in favor of these
GIT_* constants the improvement was in readability & discovery. I.e. to
see at a glance where this project was on curl compatibility &
compatibility shims.

If we add a new version-dependant curl feature the time saving is really
not going to be in adding a trivial GIT_CURL_* line to the file. It's
always going to be the compatibility shim itself that takes *much*
longer.

Right now you can easily discover where we need those shims with "git
grep GIT_?CURL" (the "?" needing fixing, as noted downthread), after
this...


      parent reply	other threads:[~2022-03-15 19:44 UTC|newest]

Thread overview: 4+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-03-15 13:16 [PATCH] git-curl-compat.h: addition of all symbols defined by curl Elia Pinto
2022-03-15 15:47 ` Junio C Hamano
2022-03-16 10:33   ` Elia Pinto
2022-03-15 19:30 ` Ævar Arnfjörð Bjarmason [this message]

Reply instructions:

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:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=220315.86pmmndmre.gmgdl@evledraar.gmail.com \
    --to=avarab@gmail.com \
    --cc=git@vger.kernel.org \
    --cc=gitter.spiros@gmail.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* 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).