linux-hwmon.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: "Ondřej Lysoněk" <olysonek@redhat.com>
To: Jean Delvare <jdelvare@suse.de>
Cc: linux-hwmon@vger.kernel.org
Subject: Re: libsensors soname bump
Date: Mon, 17 Dec 2018 11:35:46 +0100	[thread overview]
Message-ID: <3799b201-38f9-d74f-273d-f819736dcfb6@redhat.com> (raw)
In-Reply-To: <514c8587-ccf3-f7af-b88d-744add6f1da2@redhat.com>

On 17. 12. 18 10:46, Ondřej Lysoněk wrote:
> Hi Jean,
> 
> On 16. 12. 18 12:43, Jean Delvare wrote:
>> Hi Ondřej,
>>
>> You have recently released lm-sensors 3.5.0 with a new soname for
>> libsensors:
>>
>> -LIBMAINVER := 4
>> -LIBMINORVER := 4.0
>> +LIBMAINVER := 5
>> +LIBMINORVER := 0.0
>>
>> -#define SENSORS_API_VERSION            0x440
>> +#define SENSORS_API_VERSION            0x500
>>
>> This is declaring the new library as incompatible with the previous
>> version, meaning that distributions will have to build and ship both
>> libsensors4 and libsensors5 for a long time until all applications have
>> been updated and rebuilt to link with the new library. This is a
>> significant effort for the whole community and should only be done when
>> necessary.
> 
> I thought there was an ABI change, which would warrant a soname bump. Or
> am I wrong in thinking that? I was mistaken however and I'm sorry about
> that, there was no ABI change.
> 
> I don't see why distributions would have to ship two versions of the
> library. The *name* of the library didn't change, it's still libsensors
> (not libsensors4 or libsensors5).
> [~/git/lm-sensors]$ make
> ...
> [~/git/lm-sensors]$ ls lib/ | grep libsensors.so
> libsensors.so
> libsensors.so.5
> libsensors.so.5.0.0
> 
> So all distros need to do is rebuild dependent packages. No changes to
> other packages should be required.
> 
> Am I missing something?

I mean, I would love to revert the soname change, however doing so now
seems like a bad thing to do - people may have already adopted
lm_sensors 3.5.0. So I'd like to avoid reverting the change unless there
is a good justification to do so.

Ondřej Lysoněk

  reply	other threads:[~2018-12-17 10:35 UTC|newest]

Thread overview: 10+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2018-12-16 11:43 libsensors soname bump Jean Delvare
2018-12-17  9:46 ` Ondřej Lysoněk
2018-12-17 10:35   ` Ondřej Lysoněk [this message]
2018-12-17 10:59     ` Jean Delvare
2018-12-17 11:27       ` Aurelien Jarno
2018-12-17 11:48         ` Ondřej Lysoněk
2018-12-17 12:06         ` Ondřej Lysoněk
2018-12-17 10:48   ` Jean Delvare
2018-12-18 17:06 ` Ondřej Lysoněk
2018-12-19 15:10   ` Jean Delvare

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=3799b201-38f9-d74f-273d-f819736dcfb6@redhat.com \
    --to=olysonek@redhat.com \
    --cc=jdelvare@suse.de \
    --cc=linux-hwmon@vger.kernel.org \
    /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).