bug#42555: closed ([PATCH core-updates] build-system/gnu: strip with --strip-unneeded)

[GNU bug Tracking System]

Your message dated Sat, 19 Sep 2020 15:49:31 +0200
with message-id <C5RE6M5VJN1J.1FA1WPBHZPQ0S@gravity>
and subject line Re: [bug#42555] [PATCH v2 core-updates] build-system/gnu: 
strip with --strip-unneeded
has caused the debbugs.gnu.org bug report #42555,
regarding [PATCH core-updates] build-system/gnu: strip with --strip-unneeded
to be marked as done.

(If you believe you have received this mail in error, please contact
help-debbugs@gnu.org.)


-- 
42555: http://debbugs.gnu.org/cgi/bugreport.cgi?bug=42555
GNU Bug Tracking System
Contact help-debbugs@gnu.org with problems
> --- Begin Message ---
>
>
>
> Subject: 
>
> [PATCH core-updates] build-system/gnu: strip with --strip-unneeded
>
>
>
>
> Date: 
>
> Mon, 27 Jul 2020 01:38:50 +0200
>
>
>
>
> Apart from debug information, one can also strip some symbols. This can
> be a significant difference, qtbase:out consists of about 5 MB of those
> symbols. As per [1], --strip-debug is included in --strip-unneeded, and
> the debug files created also contain a copy of the information removed
> by --strip-unneeded.
>
> Linux From Scratch suggests that this option shouldn't be used on static
> libraries [2], however other sources [3] indicate otherwise. Building a
> toolchain with this patch succeeds, and the result works fine for
> 'gcc -static hello-world.c'.
>
> [1]: https://stackoverflow.com/a/52555093
> [2]: http://www.linuxfromscratch.org/lfs/view/9.1/chapter05/stripping.html
> [3]: https://www.technovelty.org/linux/stripping-shared-libraries.html
>
> * guix/build/gnu-build-system.scm (strip): Use --strip-unneeded.
> ---
>  guix/build/gnu-build-system.scm | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/guix/build/gnu-build-system.scm b/guix/build/gnu-build-system.scm
> index 2e7dff2034..d3347c9518 100644
> --- a/guix/build/gnu-build-system.scm
> +++ b/guix/build/gnu-build-system.scm
> @@ -425,7 +425,7 @@ makefiles."
>                  (objcopy-command (if target
>                                       (string-append target "-objcopy")
>                                       "objcopy"))
> -                (strip-flags '("--strip-debug"
> +                (strip-flags '("--strip-unneeded"
>                                 "--enable-deterministic-archives"))
>                  (strip-directories '("lib" "lib64" "libexec"
>                                       "bin" "sbin"))
> -- 
> 2.27.0
>
>
>
>
> --- End Message ---
> --- Begin Message ---
>
>
>
> Subject: 
>
> Re: [bug#42555] [PATCH v2 core-updates] build-system/gnu: strip with --strip-unneeded
>
>
>
>
> Date: 
>
> Sat, 19 Sep 2020 15:49:31 +0200
>
>
>
>
> On Sat Sep 19, 2020 at 4:49 PM CEST, Ludovic Courtès wrote:
> > Hi Jakub,
> >
> > Jakub Kądziołka <kuba@kadziolka.net> skribis:
> >
> > > Apart from debug information, one can also strip some symbols. This can
> > > be a significant difference, the closure of gcc-toolchain@7 got reduced
> > > by 15 MB in my tests.
> >
> > Nice! Now’s a good time to get that patch in on ‘core-updates’.
>
> Thanks. Patch pushed (with a small hiccup...)
>
> > > [1]: https://stackoverflow.com/a/52555093
> >
> > Perhaps we can omit this one. In general, I think it’s safer to refer
> > to “primary sources” as Wikipedians would say.
>
> The problem with "primary sources" is that the docs on this are quite
> sparse. I think this link is much more comprehensive.
>
> > > [2]: http://www.linuxfromscratch.org/lfs/view/9.1/chapter05/stripping.html
> > > [3]: https://www.technovelty.org/linux/stripping-shared-libraries.html
> > >
> > > * guix/build/gnu-build-system.scm (strip): Use --strip-unneeded.
> > > * guix/build-system/gnu.scm (static-package, gnu-build, gnu-cross-build): 
> > > Likewise.
> >
> > LGTM!
> >
> > For the record, I tried ‘--strip-all’ before, but that was
> > problematic
> > for some reason:
> >
> > https://lists.gnu.org/archive/html/guix-devel/2018-03/msg00135.html
>
> The docs for --strip-all say "Do not copy relocation and symbol
> information", so I would expect some relocation and/or linking use case
> to get broken...
>
> Regards,
> Jakub Kądziołka
>
>
> --- End Message ---