svn commit: r323795 - in head: audio/mp32ogg audio/p5-Shout databases/p5-DBD-InterBase databases/p5-DBIx-Class-TimeStamp databases/p5-DBIx-MySQLSequence devel/cvslines devel/cvsplot devel/cvsstat d...

[Andrej Zverev]

On Sat, Jul 27, 2013 at 9:20 PM, Alexey Dokuchaev <danfe at freebsd.org> wrote:
> On Sat, Jul 27, 2013 at 09:10:47PM +0400, Andrej Zverev wrote:
>> On Sat, Jul 27, 2013 at 8:25 PM, Alexey Dokuchaev <danfe at freebsd.org> wrote:
>> > On Sat, Jul 27, 2013 at 04:06:25PM +0000, Andrej Zverev wrote:
>> >> New Revision: 323795
>> >> URL: http://svnweb.freebsd.org/changeset/ports/323795
>> >>
>> >> @@ -32,7 +29,7 @@ NO_BUILD=   yes
>> >>  PLIST_FILES= bin/mp32ogg
>> >>
>> >>  post-patch:
>> >> -     @${PERL} -pi -e "s,%%LOCALBASE%%,${LOCALBASE},g" ${WRKSRC}/mp32ogg
>> >> +     @${REINPLACE_CMD} -i '' -e 's|%%LOCALBASE%%|${LOCALBASE}|' [...]
>> >
>> > Andrej, can you clarify why explicit -i '' is needed here and quite a few
>> > other ports in the commit?  Is it because of otherwise .bak files will be
>> > installed?  Just looks weird to see these args all over the place...
>>
>> Yes, thats a reason. I don't see (personally) why REPLACE ARGS default
>> with -i.bak, but overriding port Makefile everytime really annoying :(
>
> Understood, thanks for clarification.
>
> ./danfe
>
> P.S.  Personally I like having those .bak files around, because it makes
> debugging (via diff'ing) so much easier.  Not related to the commit, but
> just in case someone(tm) would have an idea to suppress them by default.
> Please don't.  Thank yo

I agree, but I'm just changing tool (in this task). With the .bak by
default I would have to do a full test to make sure there are no
remaining leftovers. It would take a couple of weeks. Next time I'll
try do my best as you suggested.