public inbox for newlib@sourceware.org
 help / color / mirror / Atom feed
From: Joel Sherrill <joel@rtems.org>
To: Mike Frysinger <vapier@gentoo.org>
Cc: Newlib <newlib@sourceware.org>
Subject: Re: [PATCH] newlib: libm: workaround ar duplicate member behavior
Date: Mon, 21 Feb 2022 14:51:05 -0600	[thread overview]
Message-ID: <CAF9ehCUdDQk20r_6EcPkUGfMxECun=qZo5fpxA5Q5W=iRUe36A@mail.gmail.com> (raw)
In-Reply-To: <20220221204327.2945-1-vapier@gentoo.org>

Could this be de-duped with the gnu make sort command? That would avoid the
copy.

Alternatively, I see some awk one liners on the web to uniq the set. Or
just invoke uniq.

--joel

On Mon, Feb 21, 2022, 2:43 PM Mike Frysinger <vapier@gentoo.org> wrote:

> GNU ar has undocumented behavior where it doesn't dedupe its inputs if
> they're all on the same command line, so we have to dedupe ourselves.
> ---
>  newlib/Makefile.am | 16 ++++++++++++++++
>  newlib/Makefile.in | 15 ++++++++++-----
>  2 files changed, 26 insertions(+), 5 deletions(-)
>
> diff --git a/newlib/Makefile.am b/newlib/Makefile.am
> index 80256952da3d..5bbe266cc4d5 100644
> --- a/newlib/Makefile.am
> +++ b/newlib/Makefile.am
> @@ -124,6 +124,22 @@ libm_a_CCASFLAGS = $(AM_CCASFLAGS)
> $(libm_a_CCASFLAGS_$(subst /,_,$(@D))) $(libm
>  libm_a_CPPFLAGS = $(AM_CPPFLAGS) -I$(srcdir)/libm/common
> $(libm_a_CPPFLAGS_$(subst /,_,$(@D))) $(libm_a_CPPFLAGS_$(subst
> /,_,$(@D)_$(<F)))
>  $(libm_a_OBJECTS): stmp-targ-include
>
> +## GNU ar has undocumented behavior when specifying the same name
> multiple times
> +## in a single invocation, so we have to dedupe ourselves by copying
> objects to
> +## a temporary directory.  We start with using ln to create fast
> hardlinks.  If
> +## that fails, GNU cp has annoying behavior where you can't copy multiple
> files
> +## to the same target filename, so have to run cp for each input in a
> loop.
> +## https://sourceware.org/PR28917
> +libm.a: $(libm_a_OBJECTS) $(libm_a_DEPENDENCIES)
> +       $(AM_V_at)rm -rf $@ $@.tmp && mkdir $@.tmp
> +       $(AM_V_AR)if ! ln $(libm_a_OBJECTS) $@.tmp/ 2>/dev/null; then \
> +          rm -f $@.tmp/*; \
> +          for o in $(libm_a_OBJECTS); do cp $$o $@.tmp/ || exit $$?;
> done; \
> +          fi; \
> +          $(AR) $(ARFLAGS) $@ $@.tmp/*.o
> +       $(AM_V_at)rm -rf $@.tmp
> +       $(AM_V_at)$(RANLIB) $@
> +
>  if HAVE_MULTISUBDIR
>  $(BUILD_MULTISUBDIR):
>         $(MKDIR_P) $@
> diff --git a/newlib/Makefile.in b/newlib/Makefile.in
> index 2b60ef1ca96c..84f3e0cbde37 100644
> --- a/newlib/Makefile.in
> +++ b/newlib/Makefile.in
> @@ -3248,11 +3248,6 @@ libm/machine/x86_64/libm_a-fetestexcept.$(OBJEXT):
> \
>  libm/machine/x86_64/libm_a-feupdateenv.$(OBJEXT):  \
>         libm/machine/x86_64/$(am__dirstamp)
>
> -libm.a: $(libm_a_OBJECTS) $(libm_a_DEPENDENCIES)
> $(EXTRA_libm_a_DEPENDENCIES)
> -       $(AM_V_at)-rm -f libm.a
> -       $(AM_V_AR)$(libm_a_AR) libm.a $(libm_a_OBJECTS) $(libm_a_LIBADD)
> -       $(AM_V_at)$(RANLIB) libm.a
> -
>  mostlyclean-compile:
>         -rm -f *.$(OBJEXT)
>         -rm -f libm/common/*.$(OBJEXT)
> @@ -8650,6 +8645,16 @@ libg.a: libc.a
>         $(AM_V_GEN)ln libc.a libg.a >/dev/null 2>/dev/null || cp libc.a
> libg.a
>  $(libm_a_OBJECTS): stmp-targ-include
>
> +libm.a: $(libm_a_OBJECTS) $(libm_a_DEPENDENCIES)
> +       $(AM_V_at)rm -rf $@ $@.tmp && mkdir $@.tmp
> +       $(AM_V_AR)if ! ln $(libm_a_OBJECTS) $@.tmp/ 2>/dev/null; then \
> +          rm -f $@.tmp/*; \
> +          for o in $(libm_a_OBJECTS); do cp $$o $@.tmp/ || exit $$?;
> done; \
> +          fi; \
> +          $(AR) $(ARFLAGS) $@ $@.tmp/*.o
> +       $(AM_V_at)rm -rf $@.tmp
> +       $(AM_V_at)$(RANLIB) $@
> +
>  @HAVE_MULTISUBDIR_TRUE@$(BUILD_MULTISUBDIR):
>  @HAVE_MULTISUBDIR_TRUE@        $(MKDIR_P) $@
>
> --
> 2.34.1
>
>

  reply	other threads:[~2022-02-21 20:51 UTC|newest]

Thread overview: 25+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-02-12 20:34 [PATCH] newlib: libm: merge build up a directory Mike Frysinger
2022-02-16  8:50 ` [HEADSUP] " Corinna Vinschen
2022-02-16  9:40   ` Sebastian Huber
2022-02-16 10:48     ` Corinna Vinschen
2022-02-17  4:38   ` Mike Frysinger
2022-02-17  4:42 ` [PATCH v2] " Mike Frysinger
2022-02-17 12:08   ` Corinna Vinschen
2022-02-21 11:20   ` Corinna Vinschen
2022-02-21 18:00     ` Mike Frysinger
2022-02-21 18:04       ` Jon Turney
2022-02-21 18:30         ` Mike Frysinger
2022-02-21 19:12           ` Jon Turney
2022-02-21 19:24             ` Corinna Vinschen
2022-02-21 20:30               ` Mike Frysinger
2022-02-21 20:31         ` Mike Frysinger
2022-02-21 18:28       ` Mike Frysinger
2022-02-21 20:43     ` [PATCH] newlib: libm: workaround ar duplicate member behavior Mike Frysinger
2022-02-21 20:51       ` Joel Sherrill [this message]
2022-02-21 22:12         ` Mike Frysinger
2022-02-21 22:14           ` Joel Sherrill
2022-02-22  0:21       ` [PATCH v2] " Mike Frysinger
2022-02-22 11:31         ` Corinna Vinschen
2022-02-22 12:34           ` Jon Turney
2022-02-22 17:17             ` Mike Frysinger
2022-02-23  8:56               ` Corinna Vinschen

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to='CAF9ehCUdDQk20r_6EcPkUGfMxECun=qZo5fpxA5Q5W=iRUe36A@mail.gmail.com' \
    --to=joel@rtems.org \
    --cc=newlib@sourceware.org \
    --cc=vapier@gentoo.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).