Processed: Re: Bug#971687: Bug#971668: libsane: broke ABI

2020-10-14 Thread Debian Bug Tracking System
Processing control commands:

> reopen 908681
Bug #908681 {Done: Jörg Frings-Fürst } [libsane1] libsane1: 
pointless package rename
'reopen' may be inappropriate when a bug has been closed with a version;
all fixed versions will be cleared, and you may need to re-add them.
Bug reopened
No longer marked as fixed in versions 1.0.27-3.1.

-- 
908681: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=908681
971687: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=971687
Debian Bug Tracking System
Contact ow...@bugs.debian.org with problems



Bug#971687: Bug#971668: libsane: broke ABI

2020-10-05 Thread Gianfranco Costamagna
control: severity -1 important
control: reassign -1 src:sane-backends
control: retitle -1 sane-backends: dropped unused symbols without changing 
SONAME 


Hello Vincent and Sebastian

On Mon, 5 Oct 2020 03:48:26 +0200 Vincent Lefevre  wrote:
> Control: clone -1 -2
> Control: reassign -2 libsane1 1.0.31-2
> Control: retitle -2 libsane1: broke ABI
> 
> On 2020-10-04 18:03:30 +0200, Sebastian Ramacher wrote:
> > Package: libsane
> > Version: 1.0.31-2
> > Severity: grave
> > 
> > From 1.0.31-1~experimental1:
> > 
> >* debian/libsane1.symbols:
> > - Remove 7 not longer available symbols.
> > 
> > Hence provinding libsane that depends on libsane1 with a different ABI
> > is wrong.
> 
> Not just libsane is wrong, but libsane1 (which contains the library
> itself) too (at least for programs compiled by users).
> 

Seriously?

this is the list of "dropped symbols"

- testing_append_commands_node@Base 1.0.29
- testing_known_commands_input_failed@Base 1.0.29
- testing_last_known_seq@Base 1.0.29
- testing_record_backend@Base 1.0.29
- testing_xml_doc@Base 1.0.29
- testing_xml_next_tx_node@Base 1.0.29
- testing_xml_path@Base 1.0.29


Please find a single reference of something in the archive, or outside the 
archive, that ever used
part of such (not meant to be exported) API.

The *bug* was to export them in the previous version, not to remove them, 
because meant to be internal symbols.

We had other references in the archive history, where symbols incorrectly 
exposed were dropped without
the need to change the ABI.

We discussed many times already, few people (including I guess the sponsor for 
that particular upload), and at least
3 other DDs agreeded that there was no need to change the SONAME just because 
of something that was not really
used anywhere in the world, included self-compiled stuff.

All the RFS bugs for sane-backends are public, you can find lots of discussion 
about the topic, and help
in better developing the package.

I think this is a non-issue, I would like to have some real bugs before talking 
about the Sex Of Angels... [1]

I also think its better have one single bug, instead of having two of them, for 
the very same source package.


(sorry for the Italian reference :) )
[1] https://www.englishforums.com/English/SexOfAngels/kxpgm/post.htm

just my .02$

Gianfranco

> -- 
> Vincent Lefèvre  - Web: 
> 100% accessible validated (X)HTML - Blog: 
> Work: CR INRIA - computer arithmetic / AriC project (LIP, ENS-Lyon)
> 
>