From: Junio C Hamano <gitster@pobox.com>
To: Lars Kellogg-Stedman <lars@redhat.com>
Cc: git@vger.kernel.org
Subject: Re: [PATCH] add support for specifying an SSL cipher list
Date: Thu, 07 May 2015 09:33:01 -0700 [thread overview]
Message-ID: <xmqqvbg4qreq.fsf@gitster.dls.corp.google.com> (raw)
In-Reply-To: <20150507160413.GB16334@redhat.com> (Lars Kellogg-Stedman's message of "Thu, 7 May 2015 12:04:13 -0400")
Lars Kellogg-Stedman <lars@redhat.com> writes:
> [Apologies for the dupe; this should have been cc'd to the list]
>
>> It is not clear to me what definition of "override" this sentence
>> uses.
>
> I was using it in what I thought was the common sense of "git will use
> the value in the environment variable if it exists rather than any
> value in the git configuration". I apologize if this wasn't clear;
> can you suggest how I might rephrase that?
I was hinting that the usual "override" that needs to specify
the list to be used exactly would not be very useful, in that
people often want to say one of the three things:
- allow this to be used in addition to what you usually use; or
- what you usually use is fine, but never use this one as it was
recently discovered to be insecure; or
- I have something nonstandard configured but ignore that
configuration for this invocation only and reset to the default
behaviour.
If you are changing the behaviour in your reroll, I suspect you
wouldn't be doing the common "override". If you are going to do the
'reset on empty', then 'You can set the environment variable to an
empty string to reset to the default cipher list used by libcURL.'
may be a natural way to describe it.
I briefly wondered if lack of the other two ("allow this too",
"forbid this") might become an issue not just for the environment,
but also for the configuration variable. It is probably not a huge
issue because you can say "http.<url>.sslCipherList" to limit the
scope of the affected site [*1*].
CURLOPT_SSL_CIPHER_LIST appeared in cURL 7.9 which is relatively
ancient, so it should be safe to use (please write that down in your
commit log message).
Thanks.
[Footnote]
*1* And it is a bad idea to address "allow this too" and "forbid
this" at our level---the semantics of CURLOPT_SSL_CIPHER_LIST
given by libcURL itself depends on the crypto backend (when
using OpenSSL and GnuTLS, you can say !, +, - to tweak; when
using NSS, you can only say "use these and nothing else").
next prev parent reply other threads:[~2015-05-07 16:33 UTC|newest]
Thread overview: 20+ messages / expand[flat|nested] mbox.gz Atom feed top
2015-05-07 14:16 [PATCH] add support for specifying an SSL cipher list Lars Kellogg-Stedman
2015-05-07 15:53 ` Junio C Hamano
2015-05-07 16:04 ` Lars Kellogg-Stedman
2015-05-07 16:33 ` Junio C Hamano [this message]
2015-05-07 16:58 ` Lars Kellogg-Stedman
2015-05-07 16:08 ` [PATCH v2] http: " Lars Kellogg-Stedman
2015-05-07 16:42 ` [PATCH] " Tay Ray Chuan
2015-05-07 16:57 ` Lars Kellogg-Stedman
2015-05-07 18:17 ` [PATCH v3] http: " Lars Kellogg-Stedman
2015-05-07 18:41 ` Eric Sunshine
2015-05-07 18:48 ` Lars Kellogg-Stedman
2015-05-07 18:54 ` Eric Sunshine
2015-05-07 20:51 ` Junio C Hamano
2015-05-08 3:44 ` [PATCH v4] " Lars Kellogg-Stedman
2015-05-08 3:53 ` Eric Sunshine
2015-05-08 12:15 ` SZEDER Gábor
2015-05-08 15:59 ` Junio C Hamano
2015-05-08 13:22 ` [PATCH v5] " Lars Kellogg-Stedman
2015-05-14 19:25 ` Lars Kellogg-Stedman
2015-05-14 19:39 ` Eric Sunshine
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=xmqqvbg4qreq.fsf@gitster.dls.corp.google.com \
--to=gitster@pobox.com \
--cc=git@vger.kernel.org \
--cc=lars@redhat.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).