svn commit: r456674 - in head/graphics: . libQGLViewer libQGLViewer/files

[Jason Bacon]

On 12/27/17 02:27, Alexey Dokuchaev wrote:
> On Tue, Dec 26, 2017 at 02:47:05PM -0600, Jason Bacon wrote:
>> On 12/26/17 09:52, Alexey Dokuchaev wrote:
>>> On Tue, Dec 26, 2017 at 09:18:02AM -0600, Jason Bacon wrote:
>>> ...
>>>> 3. Should PORTNAME match the port directory name?
>>> Yes, it should, with a few exceptions when e.g. PKGNAMEPREFIX comes to
>>> play (examples include CPAN ports, Linux binary ports, etc.).
>> Clarifying my point #3: When I said "match" I meant in terms of
>> upper/lower case.
>>
>> Are you suggesting that PORTNAME should also be reduced to lowercase
>> to match the directory name?
> I'd say most certainly yes, unless I'm overlooking a particular edge-case
> where it could be useful, but I cannot think of right now.  Do you have
> some specific port(s) in mind?
>
> ./danfe

emulators/virtualbox-ose is an example.  DISTNAME and the main binary 
both contain caps, but the maintainers chose to lowercase both the 
portname and directory name.

math/freemat is a counterexample, where the directory name was 
lowercased, but PORTNAME contains caps to match DISTNAME.

Either of these strategies seems reasonable to me.

Mixed case in directory names is a nuisance when trying to use shell 
auto-complete, for example.

I can't think of a strong reason to favor either lowercase or mixed for 
PORTNAME, though, since pkg is case-insensitive and we'll always need to 
use -i when grepping for PORTNAME anyway, unless all PORTNAMEs are 
reduced to lower case.

I'd lean toward doing whatever simplifies a particular port.  In the 
case of lilyterm, I used a lowercase PORTNAME to avoid having to patch 
the build system, which uses share/doc/lilyterm by default.  This 
required "GH_PROJECT=LilyTerm", which I think is more palatable than 
patching.  Had upstream used share/doc/LilyTerm, I would have leaned 
toward PORTNAME=LilyTerm and dropping GH_TAGNAME.

     Jason

-- 
Earth is a beta site.