* [patch] libffi testsuite standalone vs. gcc
@ 2009-06-08 19:36 Andreas Tobler
2009-06-08 19:51 ` Ralf Wildenhues
0 siblings, 1 reply; 3+ messages in thread
From: Andreas Tobler @ 2009-06-08 19:36 UTC (permalink / raw)
To: libffi-discuss, GCC Patches
[-- Attachment #1: Type: text/plain, Size: 671 bytes --]
Hi,
this is a try to keep the differences between the gcc-libffi testsuite
and the libffi-standalone minimal.
Tested in both environments.
Thoughts?
Andreas
P.S, the diff is against gcc svn 148292.
2009-06-08 Andreas Tobler <a.tobler@schweiz.org>
* configure.ac: Introduce a STANDALONE variable to be able to run the
testsuite accordingly.
* configure: Regenerate.
* testsuite/Makefile.am (AM_RUNTESTFLAGS): Pass the STANDALONE variable
to the testsuite.
* Makefile.in: Regenerate.
* include/Makefile.in: Likewise.
* testsuite/Makefile.in: Likewise.
* testsuite/lib/libffi-dg.exp (load_gcc_lib): Use the STANDALONE
variable.
(libffi-init): Likewise.
[-- Attachment #2: standalone_libffi_tests.diff --]
[-- Type: text/plain, Size: 3122 bytes --]
Index: configure.ac
===================================================================
--- configure.ac (revision 148292)
+++ configure.ac (working copy)
@@ -39,6 +39,13 @@
dnl The -no-testsuite modules omit the test subdir.
AM_CONDITIONAL(TESTSUBDIR, test -d $srcdir/testsuite)
+# Merge attention!
+# If we build libffi standalone, set STANDALONE to "yes", if we build it inside
+# gcc set it to "no". This variable is needed for the testsuite.
+
+STANDALONE="no"
+AC_SUBST(STANDALONE)
+
TARGETDIR="unknown"
case "$host" in
alpha*-*-*)
Index: testsuite/Makefile.am
===================================================================
--- testsuite/Makefile.am (revision 148292)
+++ testsuite/Makefile.am (working copy)
@@ -7,10 +7,10 @@
echo $(top_builddir)/../expect/expect ; \
else echo expect ; fi`
-RUNTEST = `if [ -f $(top_srcdir)/../dejagnu/runtest ] ; then \
+RUNTEST = "`if [ -f $(top_srcdir)/../dejagnu/runtest ] ; then \
echo $(top_srcdir)/../dejagnu/runtest ; \
- else echo runtest; fi`
+ else echo runtest; fi` $(AM_RUNTESTFLAGS)"
-AM_RUNTESTFLAGS =
+AM_RUNTESTFLAGS = STANDALONE=$(STANDALONE)
CLEANFILES = *.exe core* *.log *.sum
Index: testsuite/lib/libffi-dg.exp
===================================================================
--- testsuite/lib/libffi-dg.exp (revision 148292)
+++ testsuite/lib/libffi-dg.exp (working copy)
@@ -16,7 +16,13 @@
proc load_gcc_lib { filename } {
global srcdir
- load_file $srcdir/../../gcc/testsuite/lib/$filename
+ global STANDALONE
+
+ if {$STANDALONE == "yes"} {
+ load_file $srcdir/lib/$filename
+ } else {
+ load_file $srcdir/../../gcc/testsuite/lib/$filename
+ }
}
load_lib dg.exp
@@ -101,11 +107,17 @@
global libffi_link_flags
global tool_root_dir
global ld_library_path
+ global STANDALONE
- set blddirffi [lookfor_file [get_multilibs] libffi]
+ if {$STANDALONE == "yes"} {
+ set blddirffi [pwd]/..
+ } else {
+ set blddirffi [lookfor_file [get_multilibs] libffi]
+ set blddircxx [lookfor_file [get_multilibs] libstdc++-v3]
+ verbose "libstdc++ $blddircxx"
+ }
+
verbose "libffi $blddirffi"
- set blddircxx [lookfor_file [get_multilibs] libstdc++-v3]
- verbose "libstdc++ $blddircxx"
set gccdir [lookfor_file $tool_root_dir gcc/libgcc.a]
if {$gccdir != ""} {
@@ -132,8 +144,10 @@
}
# add the library path for libffi.
append ld_library_path ":${blddirffi}/.libs"
- # add the library path for libstdc++ as well.
- append ld_library_path ":${blddircxx}/src/.libs"
+ if {$STANDALONE == "no"} {
+ # add the library path for libstdc++ as well.
+ append ld_library_path ":${blddircxx}/src/.libs"
+ }
verbose "ld_library_path: $ld_library_path"
@@ -146,7 +160,9 @@
if { $libffi_dir != "" } {
set libffi_dir [file dirname ${libffi_dir}]
set libffi_link_flags "-L${libffi_dir}/.libs"
- lappend libffi_link_flags "-L${blddircxx}/src/.libs"
+ if {$STANDALONE == "no"} {
+ lappend libffi_link_flags "-L${blddircxx}/src/.libs"
+ }
}
set_ld_library_path_env_vars
^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: [patch] libffi testsuite standalone vs. gcc
2009-06-08 19:36 [patch] libffi testsuite standalone vs. gcc Andreas Tobler
@ 2009-06-08 19:51 ` Ralf Wildenhues
2009-06-11 8:24 ` Andrew Haley
0 siblings, 1 reply; 3+ messages in thread
From: Ralf Wildenhues @ 2009-06-08 19:51 UTC (permalink / raw)
To: Andreas Tobler; +Cc: libffi-discuss, GCC Patches
Hello Andreas,
* Andreas Tobler wrote on Mon, Jun 08, 2009 at 09:36:25PM CEST:
> this is a try to keep the differences between the gcc-libffi testsuite
> and the libffi-standalone minimal.
> Thoughts?
> * configure.ac: Introduce a STANDALONE variable to be able to run the
> testsuite accordingly.
Yes, unless standalone libffi routine uses a bootstrap script that sets
STANDALONE, you should default to the stand-alone version, and the
in-GCC version should key off some value passed from the toplevel
makefile.
However, I don't think you can easily postpone the decision of
stand-alone-ness until configure time for all differences so easily,
for autoconf-time differences you'd use a m4 macro not a shell variable.
Cheers,
Ralf
^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: [patch] libffi testsuite standalone vs. gcc
2009-06-08 19:51 ` Ralf Wildenhues
@ 2009-06-11 8:24 ` Andrew Haley
0 siblings, 0 replies; 3+ messages in thread
From: Andrew Haley @ 2009-06-11 8:24 UTC (permalink / raw)
To: Ralf Wildenhues, Andreas Tobler, libffi-discuss, GCC Patches
Ralf Wildenhues wrote:
> Hello Andreas,
>
> * Andreas Tobler wrote on Mon, Jun 08, 2009 at 09:36:25PM CEST:
>> this is a try to keep the differences between the gcc-libffi testsuite
>> and the libffi-standalone minimal.
>
>> Thoughts?
>
>> * configure.ac: Introduce a STANDALONE variable to be able to run the
>> testsuite accordingly.
>
> Yes, unless standalone libffi routine uses a bootstrap script that sets
> STANDALONE, you should default to the stand-alone version, and the
> in-GCC version should key off some value passed from the toplevel
> makefile.
>
> However, I don't think you can easily postpone the decision of
> stand-alone-ness until configure time for all differences so easily,
> for autoconf-time differences you'd use a m4 macro not a shell variable.
OK, so do you have a better solution? Patches welcome.
Andrew.
^ permalink raw reply [flat|nested] 3+ messages in thread
end of thread, other threads:[~2009-06-11 8:24 UTC | newest]
Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2009-06-08 19:36 [patch] libffi testsuite standalone vs. gcc Andreas Tobler
2009-06-08 19:51 ` Ralf Wildenhues
2009-06-11 8:24 ` Andrew Haley
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).