Re: CVS commit: pkgsrc/textproc

["OBATA Akio" <obache%netbsd.org@localhost>]

On Fri, 10 May 2013 21:14:18 +0900, Joerg Sonnenberger 
<joerg%britannica.bec.de@localhost> wrote:

> On Fri, May 10, 2013 at 09:08:53PM +0900, OBATA Akio wrote:
> > On Fri, 10 May 2013 21:03:11 +0900, Joerg Sonnenberger 
> > <joerg%britannica.bec.de@localhost> wrote:
> > Are you even reading what I am saying?
> > You cannot say "*Nothing*", and your personal policy is not welcomed.
>
> You asked what broke. I gave you a very precise answer ("everything not
> using GCC or its libraries"). I explained why it seemingly works at the
> moment. You are basically ignoring it as not relevant. Linking
> explicitly against libstdc++ is a bug. It is not a personal policy. It
> is no different from hard-coded -ldl and a bunch of other things. The
> configure test is broken as it doesn't fail when libstdc++ is not
> available. It is redundant as the result isn't needed for correct
> operation either. As such it is easiest to just comment it out.

Hmm,
For GCC, no need to link with libstdc++, but it is just redundant, not a bug.
For Clang, not having libstdc++ and not linked with it, but it is OK because
it is not required.

You believe there are no case other than such two case.
I don't know whether it is true or not, because I also don't know whole
development environment, but such kludge probably required for some environment,
it may be called as "a bug" or "a poor capability" of the environment.

I means that "It is redundant and not required, so it must be removed" is
"your personal policy", because you are not a upstream developer.
If you think so, please send it to upstream and start negotiation.
I usually devote much time to negotiate with upstream to change what I want.
I also have some patches not committed to pkgsrc tree even if it fixes critical
issues (dumped core, not worked as expected), because it is not accepted by 
upstream yet.

--
OBATA Akio / obache%NetBSD.org@localhost