[colord] ABI break in colord 1.0.0?

[Christopher James Halse Rogers]

On Fri, 7 Jun, 2013 at 6:44 PM, Richard Hughes <hughsient at gmail.com> 
wrote:
> On 7 June 2013 05:53, Christopher James Halse Rogers
> <chris at cooperteam.net> wrote:
>>>  I didn't see the point. Nothing was using those methods outside of
>>>  colord as they were so new and the functionality didn't yet work
>>>  properly.
>>> 
>>  Does ‘I'll need to patch it to use a Debian-specific SONAME to 
>> satisfy
>>  policy’ count as a sufficiently good justification? ☺
>> 
> Okay, I've added the compat ABI and API on color-1-0 in
> ac8fe2d92c95a1c39eebedabf85743ffbf551410 so you don't need to bump the
> soname on that branch. I can roll a new release today if you want if
> 
‘3’

Thanks very much.

>  I suppose I should really revert my revert and
> actually bump soname for master. As an aside, you really shouldn't be
> shipping master at the moment as other API breaks will happen before
> the first tarball release.
> 

I wasn't planning to ship it. I just noticed the commit when pulling

>  I'd also appreciate if you don't bump the
> soname in Debian packages, as otherwise you are going to cause
> yourself all kind of pain in the future and confuse people building
> against the library.
> 
It would be annoying, but not particularly dangerous. It would make 
binaries not portable, but not much besides that.