[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Bug#89433: I want to adopt osh



Oohara Yuuma <oohara@libra.interq.or.jp> cum veritate scripsit:

> debian/rules says:
> | # to compile with debugging information:
> | # $ debuild -e DEB_BUILD_OPTIONS="debug,nostrip"
> | # (this won't work:
> | #  DEB_BUILD_OPTIONS="debug,nostrip" && debuild)
> Note the word "won't".

That won't work, because it is syntactically incorrect,
and also this is not a place to document how to write 
bash command lone.

> > o Why are you commenting out static char *rcsid ?
> > o Adding extra {} in if statements does not seem to help anything
> > o What is the point of:
> > +/*
> >  fatal(mes)
> > +*/
> > +int fatal(mes)
> > and other changes, with commented-out changes?
> To avild gcc -Wall warnings.

Can you just make the patch smaller with

-fatal(mes)
+int fatal(mes)

which looks much better.


> > o Why are you building the binary with -O3 -fomit-frame-pointer?
> I like optimization, but I don't know it well.  Can you tell me
> a better option?

Read the gcc manpage.

> * -O : the old version (1.7-8.1) uses it
> * -O2 : the Policy (version 3.5.6.0, section 11.1) suggests it
>   "Generally the following compilation parameters should be used:"
>   (by the way, it includes -Wall)
> * -O3 : the Policy also suggests it for "Certain binaries (such as
>   computationally-intensive programs)"

Usually Anything more than -O2 has problems on some
architectures, and it is not advisable to use anything more than
-O2 unless you know what you are doing.

> > o are you sending the mods to the upstream ? / have you
> >   contacted the upstream ?
> Not yet.

Please do.

Reducing the -Wall warnings, and other things will
benefit everyone.


regards,
	junichi

-- 
dancer@debian.org : Junichi Uekawa   http://www.netfort.gr.jp/~dancer
GPG Fingerprint : 17D6 120E 4455 1832 9423  7447 3059 BF92 CD37 56F4



Reply to: