Re: Wishes to the new G-Wrap maintainer?

[Rob Browning]

address@hidden (Linas Vepstas) writes:

> On Mon, Jul 12, 2004 at 08:54:22PM +0200, Andreas Rottmann was heard to 
> remark:
>> OK, but I think as long as you are consistent, it's IMHO no problem to
>> break API on minor version numbers :-p

[...]

> I really just don't get it.  What's wrong with just incrementing
> the major version number every time you do this? So what if we get
> to version 59.1 ?  What's wrong with that, as compared to a version 
> 5.9.1 that's backwards-incompatbible with version 5.8.0, which 
> will mess everyone up?

Hmm.  I'm cautious about commenting on a discussion I haven't seen the
whole of, but to whatever extent it's relelvant, as a general rule,
I'm pretty much categorically opposed to breaking binary backward
compatibility in any way, with minor releases of public libraries.
There's no real reason for it (avoiding it is easy), and it causes
untold problems with users, distributions, packaging, etc.

Of course, I suspect there may just be some kind of misunderstanding
here since, as an example, it'll be very difficult to maintain the
Debian packages reasonably if the upstream doesn't bump the major
number (change the soname) every time they break backward binary
compatibility.

HTH
-- 
Rob Browning
rlb @defaultvalue.org and @debian.org; previously @cs.utexas.edu
GPG starting 2002-11-03 = 14DD 432F AE39 534D B592  F9A0 25C8 D377 8C7E 73A4