Re: CVS commit: pkgsrc/devel/kyua-cli

[John Marino <netbsd%marino.st@localhost>]

On 6/6/2012 16:14, Julio Merino wrote:
> On Wed, Jun 6, 2012 at 8:25 AM, John Marino<netbsd%marino.st@localhost>  wrote:
> > These for example:
> > utils/process/children.cpp:130: error: 'S_IRUSR' was not declared in this scope
> > utils/process/children.cpp:130: error: 'S_IWUSR' was not declared in this scope
> > utils/process/children.cpp:130: error: 'S_IRGRP' was not declared in this scope
> > utils/process/children.cpp:130: error: 'S_IROTH' was not declared in this scope
>
> Ah, I see.  Well, according to the documentation of chmod(2) in
> various systems, sys/stat.h is a prerequisite for this call.
> Therefore, the code was bogus in all such systems and not only
> DragonFly... which means that the applied patches with conditional
> includes are incomplete.

I imagine many of my "DragonFly" patches apply to other systems, but I'm 
not testing the patches on them.  So I rely on the package maintainer to 
remove the condition if that's the right thing.  You're right; adding 
these errors to the commit message would have improved it.


> > You can imagine more than one package has this error without the sys/stat.h 
> > header included for DragonFly.
> > previous build log:
>
> I could imagine if I tried (only now that you have shown what was
> broken), but I should not be doing that.  The commit logs are there to
> explain why things are done, so please try to be a bit more
> descriptive.  Also, pkglint warns about your patches because they do
> not contain a description at the top.  Hmm, writing the explanation in
> there would be better in this case, I presume.

Okay, that's not fair.  > 95% of the patches have no comment, and as far 
as I know, pkgdiff doesn't even generate comments.  This feature isn't 
used much and frankly pkglint shouldn't check patch comments by default.


> Anyway.  Thanks for the details.  I'll include this fix in the next release.

Okay, sorry my commit wasn't conducive to a quicker understanding but at 
least the fact it was made led to discovering a problem.

John