Date: Wed, 11 Apr 2001 21:18:07 +0300 From: Maxim Sobolev <sobomax@FreeBSD.org> To: Mikhail Teterin <mi@aldan.algebra.com> Cc: cvs-committers@FreeBSD.org, cvs-all@FreeBSD.org Subject: Re: cvs commit: ports/graphics/lcms Makefile distinfo pkg-descrpkg-plist Message-ID: <3AD49FDE.CD028423@FreeBSD.org> References: <200104111712.f3BHCgA20443@aldan.algebra.com>
next in thread | previous in thread | raw e-mail | index | archive | help
Mikhail Teterin wrote:
> On 11 Apr, Maxim Sobolev wrote:
> >> It tests if the compiled library passes its own tests after it is built.
>
> [...]
>
> > I can't agree with you. I certainly don't want any superfluous tests
> > run on my machine,
>
> They are not superfluous. They test what can easily be broken by overly
> agressive compiler options or by the compiler bugs (difference?).
>
> And they just run once -- at build time. They don't affect run-time
> performance.
>
> > just consuming my time.
>
> If one's time is so precious, one should use the prebuilt binaries.
> Testing is part of any engineering/assembly process... If you don't want
> to spend time testing -- get the thing off the shelf, prepackaged.
I can't because I have to be sure that I have most up-to-date configuration,
which is impossible with packages. It is why rebuild my ports regularly and
objecting to putting time-consuming tests here and there (and no, I'm not
stupid enough to use -O100).
> > Perhaps you are overdesigning things. If you really think that it is a
> > good thing, please provide a method to disable/enable those tests, or
> > even better - separate rule and a post-install note:
> > test:
> > blabla
> >
> > post-install:
> > ${ECHO_MSG} "If you want to ensure that the foo library ...
>
> This is one way of doing it, but, I think, this sort of tests should
> just run automaticly. They are fast and reassuring. BTW, the testcms
> routine is being built by the software's own ``install.gcc'' script.
You think so, but I don't. Want a recipe to make us both happy? See below.
> > In general please don't make any unusual assumptions about user's
> > preferences. BTW, the same applies for the automatic -j2 in your
> > several ports - there is no way to disable it, while in low memory
> > conditions it may pessimise performance due to additional swaping.
>
> Right. And not using them pessimises the sufficient-memory
> configurations. One way or another, someone will lose. I believe,
> however, that with my way, the majority wins, while things _still work_
> for the minority.
How do you know what is the best for a majority? Did you make a pool or what?
Or you use false assumption "what's good for me should be good for all
others..."?
Actually, there is a way to ensure that all would win by providing an *option*.
-Maxim
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe cvs-all" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?3AD49FDE.CD028423>