cvs commit: ports/devel/mpatrol Makefile ports/devel/mpatrol/files patch-..::..::src::mptrace.c

[Maxim Sobolev]

Vasil Dimov wrote:
> On Thu, Nov 02, 2006 at 10:50:26PM -0800, Maxim Sobolev wrote:
>> Vasil Dimov wrote:
>>> On Thu, Nov 02, 2006 at 10:18:50PM -0800, Maxim Sobolev wrote:
>>>> Vasil Dimov wrote:
>>>>> On Fri, Nov 03, 2006 at 03:51:40AM +0000, Maxim Sobolev wrote:
>>>>>> sobomax     2006-11-03 03:51:40 UTC
>>>>> [...]
>>>>>> Added files:
>>>>>>   devel/mpatrol/files  patch-..::..::src::mptrace.c 
>>>>> Aargh :-/
>>>>> I thought we were supposed to use Tools/scripts/splitpatch.pl for new
>>>>> patches.
>>>> Why so? We have Tools/scripts/patchtool.py for the few years now.
>>> The note I made was about the convention, not the exact tool used. You
>>> should read:
>>>
>>> "I thought we were supposed to use the convention used by
>>> Tools/scripts/splitpatch.pl for new patches."
>> Who those "we" are and why exactly they should use that "convention"? Is 
>> it official or what?
> 
> Those "we" are the ones creating new patches for FreeBSD ports and placing
> them in files/ subdirectories.
> 
> _They_ should use that convention because it is better to have some
> convention than no convention at all.
> 
> Please read
> http://www.freebsd.org/doc/en/books/porters-handbook/slow-patch.html
> especially:
> 
> "Each patch you wish to apply should be saved into a file named patch-*
> where * indicates the pathname of the file that is patched, such as
> patch-Imakefile or patch-src-config.h
> ...
> Please only use characters [-+._a-zA-Z0-9] for naming your patches."
> 
> I dunno what do you mean by "official" and if you consider the
> "Porters Handbook" being "official".
> 
> Anyway you can name your patches
> patch-@}>-->---ILoveYouMary-src_file.c if you like but you will raise
> usless discussions like the current one.

Heh, those who don't know the history are doomed to repeat its mistakes.

The Porters Handbook unfortunately is out of sync with reality. Long 
time ago, when the trees were big and the sky was blue, we did have the 
strict rule saying that any patch should be named patch-[a-z][a-z]. It 
was OK until the ports collection had reached the point when keeping 
individual history of each patch-XY became important enough, which you 
can't do with names like patch-XY. At that point "vote" was called to 
allow "free" names for patchfiles (but consistent within each port), 
provided that those names don't break any existing packaging tools. The 
result of the vote was positive. Several tools were created to assist 
with routine patch management tasts based on this. That's the status quo 
for the moment.

If somebody has sneaked its own definition into Porters Handbook I can't 
care less. It is not official unless majority of active porters agree 
upon it.

-Maxim