Mismatch [1.5] [1.7]: lzip-1.8-1

[Dave Korn]

Charles Wilson wrote:
> Dave Korn wrote:
>>   If you build and package exactly the same sources in exactly the same way
>> under 1.5 and 1.7, I would by default expect them to be released with the same
>> version numbers.  Nothing is actually going to conflict, is it?
> 
> But the configure checks will "discover" different things on the two
> platforms, so the resultant binaries will differ by more than (possibly)
> using entry points in cygwin1.dll not available in the other version.

  Absolutely, that's precisely my reason for building the same package
separately under 1.5 and 1.7.

> When I see a bug report "Problem with ncurses-5.7-11" I know right away
> that the OP is using cygwin-1.7; sure, I could clarify if I needed to,
> but one less step in the problem-analysis phase is always nice.

  Well, when speaking for myself, I'd rather occasionally have to ask someone
which cygwin version they were using than roll a respin.  (And the whole issue
may become moot, depending on a separate decision but I'll post about that
another time because I'm just off AFK for a bit.)

>>   Note that this is not just a theoretical concern for me, I've already rolled
>> a gcc4-4.3.4-2 that just needs some testing, but now it seems I may have to
>> respin it (or perhaps drop the 1.5 version I was planning, but that's a topic
>> for another thread.)
> 
> Well, first: did you actually *build* it twice?  Or only once (on
> cygwin-1.5) and are relying on union-fs to DTRT?  If the latter, then
> you're good to go.  If the former...well, ick,

  Yes, twice, and indeed as you and I both expected, the configure test for
__wrap_Znaj only succeeds on 1.7, leading to different specs in the generated
compilers, to name but one minor difference.  But unless I've misunderstood
union-fs there's no problem uploading two different versions and putting an
explicit one in the release-2 tree will just override the shadowing, won't it?

    cheers,
      DaveK