curl / Mailing Lists / curl-library / Single Mail



From: Daniel Stenberg <>
Date: Sun, 27 Aug 2017 19:22:57 +0200 (CEST)

On Sat, 26 Aug 2017, Zakrzewski, Jakub wrote:

Hey, welcome to history hour in which I'll explain some archaic stuff from
this project's past... =)

> First, the comment suggests that the configure script can make a decision if
> the SONAME should be bumped. I haven't found any code supporting that
> statement.

It used to be able to, but I removed that logic not too long ago. Mostly
because I don't think it was ever used. In fact, I don't think the manual
bumping has ever been used either...

> Second: unless I'm that stupid (entirely possible), the bump here does not
> adhere to the guide (the comment). The current version will produce
>, the bumped one instead of expected

I'd agree. I suppose that too shows that it hasn't been used.

> The third thing is the message at the configure summary:
>> SONAME bump: yes - WARNING: this library will be built with the SONAME
>> number bumped due to (a detected) ABI breakage.
>> See lib/README.curl_off_t for details on this.
> ?I haven't found the mentioned file lib/README.curl_off_t.

That file was moved into the document a while ago, so the comment
is clearly wrong.

The logic tried to detect systems where the previous logic and the current
logic would detect different 'curl_off_t' sizes so that the libs would end up
with differen ABIs. I think we can now safely say that that logic is now water
under the bridge and there won't be any automatic bump and most certainly no
manual bump either...

Received on 2017-08-27