| Message ID | 1335557783-13575-1-git-send-email-peter.portante@redhat.com |
|---|---|
| State | New |
| Headers | show |
On 04/27/2012 02:16 PM, Peter Portante wrote: > The package config check for gthreads might have already placed > a -lrt switch in LIBS earlier. > > Refactored the code from the pthread switch removal, from commit > e3c56761b465a4253871c32b06ebbc2d8b3fc3e1, to make it work for the > more general case. > > +add_to_libs() { > + found=no > + for lib_entry in $LIBS; do > + if test "$lib_entry" = "$1"; then > + found=yes > + break > + fi > + done > + if test "$found" = "no"; then > + LIBS="$1 $LIBS" > + fi > +} Shorter (and probably faster) to write this as: add_to_libs() { case " $LIBS " in " $1 ") ;; *) LIBS="$1 $LIBS" ;; esac }
On 27 April 2012 22:55, Eric Blake <eblake@redhat.com> wrote: > Shorter (and probably faster) to write this as: > > add_to_libs() { > case " $LIBS " in > " $1 ") ;; > *) LIBS="$1 $LIBS" ;; > esac > } ...doesn't that need to be *" $1 "*) ;; ? -- PMM
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Am 27.04.2012 23:55, schrieb Eric Blake: > On 04/27/2012 02:16 PM, Peter Portante wrote: >> The package config check for gthreads might have already placed a >> -lrt switch in LIBS earlier. >> >> Refactored the code from the pthread switch removal, from commit >> e3c56761b465a4253871c32b06ebbc2d8b3fc3e1, to make it work for >> the more general case. >> > >> +add_to_libs() { + found=no + for lib_entry in $LIBS; do + >> if test "$lib_entry" = "$1"; then + found=yes + break + >> fi + done + if test "$found" = "no"; then + LIBS="$1 $LIBS" >> + fi +} > > Shorter (and probably faster) to write this as: > > add_to_libs() { case " $LIBS " in " $1 ") ;; *) LIBS="$1 $LIBS" ;; > esac } Er, no. That would only catch the initial duplication case -lrt -lrt but not -lbla -lrt -lrt or -lrt -lbla -lrt, wouldn't it? But then again Peter's function wouldn't find duplicates when $1 contains a space either... tricky. Andreas - -- SUSE LINUX Products GmbH, Maxfeldstr. 5, 90409 Nürnberg, Germany GF: Jeff Hawn, Jennifer Guild, Felix Imendörffer; HRB 16746 AG Nürnberg -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.18 (GNU/Linux) iQIcBAEBAgAGBQJPmxuZAAoJEPou0S0+fgE/ycMP/iW+aPgv810H/RUAzc90pQf0 pjRh+R1dFKyukIcwGlYT7WiuAjLdNTM7HcUbOL1Gfads6ZvxEHonYBpNxTOi424U kHpfhWU0hfZxv78Rv8gvD4VpSlIo6x4nwrmFvRKqskiOjPCf/yQ/Qde1MVJMmRAR x1Mjmy1Wnm71smA8sbdSCqPM3VTWzqZKcXIkSw1kymz3g43OIM98QFNPz/EkwmQb +q7pos8GwjanwQDBt4OQW5sYQDLpWMKA2na4tRYU/nxL6jXBi2BLXqm+8lbu1bhE MHOulQZ90CELSiYyKgZeLNQCNrH4F/B4XoIwBgWhqV/ND1zWQndSDlWxWYVZ2DHY awqs5KPOFybVWFfwFLaZ4cWefOPRJPNSHNufVdgVrjbn7UgXYZ/n97TAaHbqNX5/ HqtR/N4XSn4SYblsRXbkzNg6EcTdolLUuE/RSey46+NlfiCjWnr047SzJjoHmfyq YRFf3VK7N8Aa46y8szQOvVCOBOv117U8uqzfzVEGn4zfmkYi6No+1O640SmsYGyY Iii8EpBDCwyGzWw8o96R/BeJBufpNHg3vyUcWXAK5MEftHf/5aLhEmOAhGBz+TGU jHKmhiXUSLFH0xVGOkwq0wNXKkMsHuk9hIYb6k687P/AV6Iu8SkEtG7/4HFV6RV/ btycq/KaQ75a7TJj/XFW =+qrm -----END PGP SIGNATURE-----
On 04/27/2012 04:16 PM, Peter Maydell wrote: > On 27 April 2012 22:55, Eric Blake <eblake@redhat.com> wrote: >> Shorter (and probably faster) to write this as: >> >> add_to_libs() { >> case " $LIBS " in >> " $1 ") ;; >> *) LIBS="$1 $LIBS" ;; >> esac >> } > > ...doesn't that need to be > *" $1 "*) ;; Yep; that's what I get for typing without testing.
On 04/27/2012 04:20 PM, Andreas Färber wrote: > Am 27.04.2012 23:55, schrieb Eric Blake: >> On 04/27/2012 02:16 PM, Peter Portante wrote: >>> The package config check for gthreads might have already placed a >>> -lrt switch in LIBS earlier. >>> >>> Refactored the code from the pthread switch removal, from commit >>> e3c56761b465a4253871c32b06ebbc2d8b3fc3e1, to make it work for >>> the more general case. >>> > >>> +add_to_libs() { + found=no + for lib_entry in $LIBS; do + >>> if test "$lib_entry" = "$1"; then + found=yes + break + >>> fi + done + if test "$found" = "no"; then + LIBS="$1 $LIBS" >>> + fi +} > >> Shorter (and probably faster) to write this as: > >> add_to_libs() { case " $LIBS " in " $1 ") ;; *) LIBS="$1 $LIBS" ;; >> esac } > > Er, no. That would only catch the initial duplication case -lrt -lrt > but not -lbla -lrt -lrt or -lrt -lbla -lrt, wouldn't it? > > But then again Peter's function wouldn't find duplicates when $1 > contains a space either... tricky. If you want to filter duplicates out of $1, then: add_to_libs() { for lib in $1; do case " $LIBS " in *" $lib "*) ;; *) LIBS="$lib $LIBS" ;; esac done } Hopefully, we don't encounter any libs with circular dependencies, where a lib must be listed twice to be effective.
diff --git a/configure b/configure index 15bbc73..4c4ecfb 100755 --- a/configure +++ b/configure @@ -81,6 +81,19 @@ path_of() { return 1 } +add_to_libs() { + found=no + for lib_entry in $LIBS; do + if test "$lib_entry" = "$1"; then + found=yes + break + fi + done + if test "$found" = "no"; then + LIBS="$1 $LIBS" + fi +} + # default parameters source_path=`dirname "$0"` cpu="" @@ -2083,16 +2096,7 @@ else for pthread_lib in $PTHREADLIBS_LIST; do if compile_prog "" "$pthread_lib" ; then pthread=yes - found=no - for lib_entry in $LIBS; do - if test "$lib_entry" = "$pthread_lib"; then - found=yes - break - fi - done - if test "$found" = "no"; then - LIBS="$pthread_lib $LIBS" - fi + add_to_libs "$pthread_lib" break fi done @@ -2581,7 +2585,7 @@ EOF if compile_prog "" "" ; then : elif compile_prog "" "-lrt" ; then - LIBS="-lrt $LIBS" + add_to_libs "-lrt" fi if test "$darwin" != "yes" -a "$mingw32" != "yes" -a "$solaris" != yes -a \
The package config check for gthreads might have already placed a -lrt switch in LIBS earlier. Refactored the code from the pthread switch removal, from commit e3c56761b465a4253871c32b06ebbc2d8b3fc3e1, to make it work for the more general case. Signed-off-by: Peter Portante <peter.portante@redhat.com> --- configure | 26 +++++++++++++++----------- 1 files changed, 15 insertions(+), 11 deletions(-)