svn commit: r343559 - head/net-p2p/litecoin

[Adam Weinberger]

On 10 Feb 2014, at 5:12, Baptiste Daroussin wrote:

> On Mon, Feb 10, 2014 at 09:47:17AM +0000, David Chisnall wrote:
>> On 10 Feb 2014, at 07:52, Baptiste Daroussin <bapt at FreeBSD.org> 
>> wrote:
>>
>>> I do not think that requires an approval, as for FreeBSD this change 
>>> is no-op,
>>> It is also no intrusive and just "fixes" building of that port for 
>>> Dragonfly.
>>
>> Relying on OPSYS and OSVERSION is pretty horrible and doesn't scale.  
>> Already we have problems with convoluted OSVERSION expressions that 
>> are hard to understand and really mean 'fix for PR12345 has been 
>> merged to the relevant branch'.  It would be far better to centralise 
>> these into a single file for each platform, listing features / 
>> bugfixes / needed work-arounds and start removing explicit OSVERSION 
>> checks from any other Makefiles.  If supporting Dragonfly is one of 
>> our objectives, then I'd strongly encourage the people who are 
>> working on it to engage in this kind of cleanup, rather than piling 
>> more hacks into the tree.
>>
>> David
>>
>
> If one has a nice idea to centralize those informations I'm all about 
> it, by
> nice idea I mean format and implementation.
>
> May that be OPSYS and/or OSVERSION both requires loading too many 
> times bsd.*.mk
> which is not good, looking forward for propositions.
>
> regards,
> Bapt

pkgconf. It's ISC-licenced, has no external dependencies, and would be a 
great way to store information about the system itself. Instead of using 
magic to guess which compiler to use, a compiler.pc could tell you. The 
mayhem that ensued when the threading library changed years ago, a 
thread.pc would have made it a piece of cake. You could store 
capabilities in .pc files and have a single reliable way to retrieve it.

# Adam


--
Adam Weinberger
adamw at adamw.org
http://www.adamw.org