svn commit: r365003 - head/devel/elfsh/files

[olli hauer]

On 2014-08-16 18:45, Alexey Dokuchaev wrote:
> On Fri, Aug 15, 2014 at 05:04:41PM +0000, Adam Weinberger wrote:
>> New Revision: 365003
>> URL: http://svnweb.freebsd.org/changeset/ports/365003
>> QAT: https://qat.redports.org/buildarchive/r365003/
>>
>> @@ -1,5 +1,5 @@
>> ---- vm/Makefile.orig	2003-08-21 09:23:17.000000000 +0800
>> -+++ vm/Makefile	2011-10-03 16:34:23.000000000 +0800
>> +--- vm/Makefile.orig	2003-08-20 21:23:17.000000000 -0400
>> ++++ vm/Makefile	2014-08-15 13:03:43.000000000 -0400
> 
> Our new proposed awesome "makepatch" (https://phabric.freebsd.org/D582)
> would've allowed you to make these lines TZ-agnostic so they won't be
> touched on subsequent patch file changes, thus making the diffs clearer.
> 
> ./danfe
> 

Hm fixing the timestamp, OK but the path seperator ?

There are 1182 ports with an '_' in the name and I think reading a patch name
patch-mod_foo__main.c is easier to read then
patch-mod_foo_main.c
patch-mod_foo-main.c
patch-mod_foo--main.c
patch-mod_foo+main.c


Sample ports that will get really bad patch names with '-' or '+' or a
single '_'as new path separtor

o japanese/timidity++-slang
o lang/libstdc++_stldoc
o textproc/libxml++26-reference
o databases/mysql-connector-c++
o devel/bncsutil-ghost++


And I can extent the list with a few dozens more.

So I vote to keep the current makepatch notation, discussing the removal of
the timestamps would be better with a separate patch, so an partial agreement
does not implicit the full patch ;)


-- 
olli