[RFA] testsuite cleanups (part 2)

[RFA] testsuite cleanups (part 2)

[Elena Zannoni]

Here, I am observing a peculiar situation. If you run the testsuite in
a directory that is not the build directory, i.e. the gdb.blah object
dirs are recreated outside of the build tree, you will see that all
the tests which use pthreads.c will not compile. This is because
pthreads.c includes "config.h" which gets created in the build
gdb/testsuite directory at configure time. Config.h was used to
determine if there is posix thread support or not, but now we use the
method of trying to compile with various -lblah so we don't really
need this anymore.  

It took me a while to notice this failure because I have the bad habit
of looking just at the unexpected failures, glossing over the
unsupported results.

OK?

elena

[PS. I assume I used the correct autoconf, but I am nost sure: I used the
autoconf-000227 snapshot].

2003-07-29  Elena Zannoni  <ezannoni@redhat.com>
        
        * configure.in: Don't generate config.h from config.hin.
        * configure: Regenerate.
        * config.h: Remove file.
        * gdb.mi/pthreads.c: Remove include of config.h, and other ifdefs.
        * gdb.threads/pthreads.c: Ditto.


Index: configure
===================================================================
RCS file: /cvs/uberbaum/gdb/testsuite/configure,v
retrieving revision 1.9
diff -u -p -r1.9 configure
--- configure	1 May 2003 01:09:51 -0000	1.9
+++ configure	29 Jul 2003 19:10:33 -0000
@@ -34,6 +34,7 @@ program_suffix=NONE
 program_transform_name=s,x,x,
 silent=
 site=
+sitefile=
 srcdir=
 target=NONE
 verbose=
@@ -148,6 +149,7 @@ Configuration:
   --help                  print this message
   --no-create             do not create output files
   --quiet, --silent       do not print \`checking...' messages
+  --site-file=FILE        use FILE as the site file
   --version               print the version of autoconf that created configure
 Directory and file names:
   --prefix=PREFIX         install architecture-independent files in PREFIX
@@ -318,6 +320,11 @@ EOF
   -site=* | --site=* | --sit=*)
     site="$ac_optarg" ;;
 
+  -site-file | --site-file | --site-fil | --site-fi | --site-f)
+    ac_prev=sitefile ;;
+  -site-file=* | --site-file=* | --site-fil=* | --site-fi=* | --site-f=*)
+    sitefile="$ac_optarg" ;;
+
   -srcdir | --srcdir | --srcdi | --srcd | --src | --sr)
     ac_prev=srcdir ;;
   -srcdir=* | --srcdir=* | --srcdi=* | --srcd=* | --src=* | --sr=*)
@@ -483,12 +490,16 @@ fi
 srcdir=`echo "${srcdir}" | sed 's%\([^/]\)/*$%\1%'`
 
 # Prefer explicitly selected file to automatically selected ones.
-if test -z "$CONFIG_SITE"; then
-  if test "x$prefix" != xNONE; then
-    CONFIG_SITE="$prefix/share/config.site $prefix/etc/config.site"
-  else
-    CONFIG_SITE="$ac_default_prefix/share/config.site $ac_default_prefix/etc/config.site"
+if test -z "$sitefile"; then
+  if test -z "$CONFIG_SITE"; then
+    if test "x$prefix" != xNONE; then
+      CONFIG_SITE="$prefix/share/config.site $prefix/etc/config.site"
+    else
+      CONFIG_SITE="$ac_default_prefix/share/config.site $ac_default_prefix/etc/config.site"
+    fi
   fi
+else
+  CONFIG_SITE="$sitefile"
 fi
 for ac_site_file in $CONFIG_SITE; do
   if test -r "$ac_site_file"; then
@@ -528,7 +539,6 @@ fi
 
 
 
-
 ac_aux_dir=
 for ac_dir in $srcdir $srcdir/.. $srcdir/../..; do
   if test -f $ac_dir/install-sh; then
@@ -549,7 +559,7 @@ ac_config_sub=$ac_aux_dir/config.sub
 ac_configure=$ac_aux_dir/configure # This should be Cygnus configure.
 
 echo $ac_n "checking build system type""... $ac_c" 1>&6
-echo "configure:553: checking build system type" >&5
+echo "configure:563: checking build system type" >&5
 
 build_alias=$build
 case "$build_alias" in
@@ -573,7 +583,7 @@ else { echo "configure: error: can not r
 fi
 
 echo $ac_n "checking host system type""... $ac_c" 1>&6
-echo "configure:577: checking host system type" >&5
+echo "configure:587: checking host system type" >&5
 
 host_alias=$host
 case "$host_alias" in
@@ -594,7 +604,7 @@ host_os=`echo $host | sed 's/^\([^-]*\)-
 echo "$ac_t""$host" 1>&6
 
 echo $ac_n "checking target system type""... $ac_c" 1>&6
-echo "configure:598: checking target system type" >&5
+echo "configure:608: checking target system type" >&5
 
 target_alias=$target
 case "$target_alias" in
@@ -709,7 +719,7 @@ fi
 
 
 echo $ac_n "checking how to run the C preprocessor""... $ac_c" 1>&6
-echo "configure:713: checking how to run the C preprocessor" >&5
+echo "configure:723: checking how to run the C preprocessor" >&5
 # On Suns, sometimes $CPP names a directory.
 if test -n "$CPP" && test -d "$CPP"; then
   CPP=
@@ -724,13 +734,13 @@ else
   # On the NeXT, cc -E runs the code through the compiler's parser,
   # not just through cpp.
   cat > conftest.$ac_ext <<EOF
-#line 728 "configure"
+#line 738 "configure"
 #include "confdefs.h"
 #include <assert.h>
 Syntax Error
 EOF
 ac_try="$ac_cpp conftest.$ac_ext >/dev/null 2>conftest.out"
-{ (eval echo configure:734: \"$ac_try\") 1>&5; (eval $ac_try) 2>&5; }
+{ (eval echo configure:744: \"$ac_try\") 1>&5; (eval $ac_try) 2>&5; }
 ac_err=`grep -v '^ *+' conftest.out | grep -v "^conftest.${ac_ext}\$"`
 if test -z "$ac_err"; then
   :
@@ -741,13 +751,13 @@ else
   rm -rf conftest*
   CPP="${CC-cc} -E -traditional-cpp"
   cat > conftest.$ac_ext <<EOF
-#line 745 "configure"
+#line 755 "configure"
 #include "confdefs.h"
 #include <assert.h>
 Syntax Error
 EOF
 ac_try="$ac_cpp conftest.$ac_ext >/dev/null 2>conftest.out"
-{ (eval echo configure:751: \"$ac_try\") 1>&5; (eval $ac_try) 2>&5; }
+{ (eval echo configure:761: \"$ac_try\") 1>&5; (eval $ac_try) 2>&5; }
 ac_err=`grep -v '^ *+' conftest.out | grep -v "^conftest.${ac_ext}\$"`
 if test -z "$ac_err"; then
   :
@@ -758,13 +768,13 @@ else
   rm -rf conftest*
   CPP="${CC-cc} -nologo -E"
   cat > conftest.$ac_ext <<EOF
-#line 762 "configure"
+#line 772 "configure"
 #include "confdefs.h"
 #include <assert.h>
 Syntax Error
 EOF
 ac_try="$ac_cpp conftest.$ac_ext >/dev/null 2>conftest.out"
-{ (eval echo configure:768: \"$ac_try\") 1>&5; (eval $ac_try) 2>&5; }
+{ (eval echo configure:778: \"$ac_try\") 1>&5; (eval $ac_try) 2>&5; }
 ac_err=`grep -v '^ *+' conftest.out | grep -v "^conftest.${ac_ext}\$"`
 if test -z "$ac_err"; then
   :
@@ -792,17 +802,17 @@ for ac_hdr in pthread.h
 do
 ac_safe=`echo "$ac_hdr" | sed 'y%./+-%__p_%'`
 echo $ac_n "checking for $ac_hdr""... $ac_c" 1>&6
-echo "configure:796: checking for $ac_hdr" >&5
+echo "configure:806: checking for $ac_hdr" >&5
 if eval "test \"`echo '$''{'ac_cv_header_$ac_safe'+set}'`\" = set"; then
   echo $ac_n "(cached) $ac_c" 1>&6
 else
   cat > conftest.$ac_ext <<EOF
-#line 801 "configure"
+#line 811 "configure"
 #include "confdefs.h"
 #include <$ac_hdr>
 EOF
 ac_try="$ac_cpp conftest.$ac_ext >/dev/null 2>conftest.out"
-{ (eval echo configure:806: \"$ac_try\") 1>&5; (eval $ac_try) 2>&5; }
+{ (eval echo configure:816: \"$ac_try\") 1>&5; (eval $ac_try) 2>&5; }
 ac_err=`grep -v '^ *+' conftest.out | grep -v "^conftest.${ac_ext}\$"`
 if test -z "$ac_err"; then
   rm -rf conftest*
@@ -830,12 +840,12 @@ done
 
 
 echo $ac_n "checking for Cygwin environment""... $ac_c" 1>&6
-echo "configure:834: checking for Cygwin environment" >&5
+echo "configure:844: checking for Cygwin environment" >&5
 if eval "test \"`echo '$''{'ac_cv_cygwin'+set}'`\" = set"; then
   echo $ac_n "(cached) $ac_c" 1>&6
 else
   cat > conftest.$ac_ext <<EOF
-#line 839 "configure"
+#line 849 "configure"
 #include "confdefs.h"
 
 int main() {
@@ -846,7 +856,7 @@ int main() {
 return __CYGWIN__;
 ; return 0; }
 EOF
-if { (eval echo configure:850: \"$ac_compile\") 1>&5; (eval $ac_compile) 2>&5; }; then
+if { (eval echo configure:860: \"$ac_compile\") 1>&5; (eval $ac_compile) 2>&5; }; then
   rm -rf conftest*
   ac_cv_cygwin=yes
 else
@@ -863,19 +873,19 @@ echo "$ac_t""$ac_cv_cygwin" 1>&6
 CYGWIN=
 test "$ac_cv_cygwin" = yes && CYGWIN=yes
 echo $ac_n "checking for mingw32 environment""... $ac_c" 1>&6
-echo "configure:867: checking for mingw32 environment" >&5
+echo "configure:877: checking for mingw32 environment" >&5
 if eval "test \"`echo '$''{'ac_cv_mingw32'+set}'`\" = set"; then
   echo $ac_n "(cached) $ac_c" 1>&6
 else
   cat > conftest.$ac_ext <<EOF
-#line 872 "configure"
+#line 882 "configure"
 #include "confdefs.h"
 
 int main() {
 return __MINGW32__;
 ; return 0; }
 EOF
-if { (eval echo configure:879: \"$ac_compile\") 1>&5; (eval $ac_compile) 2>&5; }; then
+if { (eval echo configure:889: \"$ac_compile\") 1>&5; (eval $ac_compile) 2>&5; }; then
   rm -rf conftest*
   ac_cv_mingw32=yes
 else
@@ -894,7 +904,7 @@ test "$ac_cv_mingw32" = yes && MINGW32=y
 
 
 echo $ac_n "checking for executable suffix""... $ac_c" 1>&6
-echo "configure:898: checking for executable suffix" >&5
+echo "configure:908: checking for executable suffix" >&5
 if eval "test \"`echo '$''{'ac_cv_exeext'+set}'`\" = set"; then
   echo $ac_n "(cached) $ac_c" 1>&6
 else
@@ -904,10 +914,10 @@ else
   rm -f conftest*
   echo 'int main () { return 0; }' > conftest.$ac_ext
   ac_cv_exeext=
-  if { (eval echo configure:908: \"$ac_link\") 1>&5; (eval $ac_link) 2>&5; }; then
+  if { (eval echo configure:918: \"$ac_link\") 1>&5; (eval $ac_link) 2>&5; }; then
     for file in conftest.*; do
       case $file in
-      *.c | *.o | *.obj) ;;
+      *.c | *.o | *.obj | *.ilk | *.pdb) ;;
       *) ac_cv_exeext=`echo $file | sed -e s/conftest//` ;;
       esac
     done
@@ -990,7 +1000,19 @@ fi
 
 trap 'rm -f $CONFIG_STATUS conftest*; exit 1' 1 2 15
 
-DEFS=-DHAVE_CONFIG_H
+# Transform confdefs.h into DEFS.
+# Protect against shell expansion while executing Makefile rules.
+# Protect against Makefile macro expansion.
+cat > conftest.defs <<\EOF
+s%#define \([A-Za-z_][A-Za-z0-9_]*\) *\(.*\)%-D\1=\2%g
+s%[ 	`~#$^&*(){}\\|;'"<>?]%\\&%g
+s%\[%\\&%g
+s%\]%\\&%g
+s%\$%$$%g
+EOF
+DEFS=`sed -f conftest.defs confdefs.h | tr '\012' ' '`
+rm -f conftest.defs
+
 
 # Without the "./", some shells look in PATH for config.status.
 : ${CONFIG_STATUS=./config.status}
@@ -1029,7 +1051,7 @@ ac_given_srcdir=$srcdir
 
 trap 'rm -fr `echo "Makefile gdb.arch/Makefile gdb.asm/Makefile gdb.base/Makefile \
   gdb.c++/Makefile gdb.disasm/Makefile gdb.java/Makefile gdb.mi/Makefile \
-  gdb.objc/Makefile gdb.threads/Makefile gdb.trace/Makefile config.h:config.hin" | sed "s/:[^ ]*//g"` conftest*; exit 1' 1 2 15
+  gdb.objc/Makefile gdb.threads/Makefile gdb.trace/Makefile" | sed "s/:[^ ]*//g"` conftest*; exit 1' 1 2 15
 EOF
 cat >> $CONFIG_STATUS <<EOF
 
@@ -1176,113 +1198,6 @@ s%@top_srcdir@%$top_srcdir%g
 " $ac_file_inputs | (eval "$ac_sed_cmds") > $ac_file
 fi; done
 rm -f conftest.s*
-
-# These sed commands are passed to sed as "A NAME B NAME C VALUE D", where
-# NAME is the cpp macro being defined and VALUE is the value it is being given.
-#
-# ac_d sets the value in "#define NAME VALUE" lines.
-ac_dA='s%^\([ 	]*\)#\([ 	]*define[ 	][ 	]*\)'
-ac_dB='\([ 	][ 	]*\)[^ 	]*%\1#\2'
-ac_dC='\3'
-ac_dD='%g'
-# ac_u turns "#undef NAME" with trailing blanks into "#define NAME VALUE".
-ac_uA='s%^\([ 	]*\)#\([ 	]*\)undef\([ 	][ 	]*\)'
-ac_uB='\([ 	]\)%\1#\2define\3'
-ac_uC=' '
-ac_uD='\4%g'
-# ac_e turns "#undef NAME" without trailing blanks into "#define NAME VALUE".
-ac_eA='s%^\([ 	]*\)#\([ 	]*\)undef\([ 	][ 	]*\)'
-ac_eB='$%\1#\2define\3'
-ac_eC=' '
-ac_eD='%g'
-
-if test "${CONFIG_HEADERS+set}" != set; then
-EOF
-cat >> $CONFIG_STATUS <<EOF
-  CONFIG_HEADERS="config.h:config.hin"
-EOF
-cat >> $CONFIG_STATUS <<\EOF
-fi
-for ac_file in .. $CONFIG_HEADERS; do if test "x$ac_file" != x..; then
-  # Support "outfile[:infile[:infile...]]", defaulting infile="outfile.in".
-  case "$ac_file" in
-  *:*) ac_file_in=`echo "$ac_file"|sed 's%[^:]*:%%'`
-       ac_file=`echo "$ac_file"|sed 's%:.*%%'` ;;
-  *) ac_file_in="${ac_file}.in" ;;
-  esac
-
-  echo creating $ac_file
-
-  rm -f conftest.frag conftest.in conftest.out
-  ac_file_inputs=`echo $ac_file_in|sed -e "s%^%$ac_given_srcdir/%" -e "s%:% $ac_given_srcdir/%g"`
-  cat $ac_file_inputs > conftest.in
-
-EOF
-
-# Transform confdefs.h into a sed script conftest.vals that substitutes
-# the proper values into config.h.in to produce config.h.  And first:
-# Protect against being on the right side of a sed subst in config.status.
-# Protect against being in an unquoted here document in config.status.
-rm -f conftest.vals
-cat > conftest.hdr <<\EOF
-s/[\\&%]/\\&/g
-s%[\\$`]%\\&%g
-s%#define \([A-Za-z_][A-Za-z0-9_]*\) *\(.*\)%${ac_dA}\1${ac_dB}\1${ac_dC}\2${ac_dD}%gp
-s%ac_d%ac_u%gp
-s%ac_u%ac_e%gp
-EOF
-sed -n -f conftest.hdr confdefs.h > conftest.vals
-rm -f conftest.hdr
-
-# This sed command replaces #undef with comments.  This is necessary, for
-# example, in the case of _POSIX_SOURCE, which is predefined and required
-# on some systems where configure will not decide to define it.
-cat >> conftest.vals <<\EOF
-s%^[ 	]*#[ 	]*undef[ 	][ 	]*[a-zA-Z_][a-zA-Z_0-9]*%/* & */%
-EOF
-
-# Break up conftest.vals because some shells have a limit on
-# the size of here documents, and old seds have small limits too.
-
-rm -f conftest.tail
-while :
-do
-  ac_lines=`grep -c . conftest.vals`
-  # grep -c gives empty output for an empty file on some AIX systems.
-  if test -z "$ac_lines" || test "$ac_lines" -eq 0; then break; fi
-  # Write a limited-size here document to conftest.frag.
-  echo '  cat > conftest.frag <<CEOF' >> $CONFIG_STATUS
-  sed ${ac_max_here_lines}q conftest.vals >> $CONFIG_STATUS
-  echo 'CEOF
-  sed -f conftest.frag conftest.in > conftest.out
-  rm -f conftest.in
-  mv conftest.out conftest.in
-' >> $CONFIG_STATUS
-  sed 1,${ac_max_here_lines}d conftest.vals > conftest.tail
-  rm -f conftest.vals
-  mv conftest.tail conftest.vals
-done
-rm -f conftest.vals
-
-cat >> $CONFIG_STATUS <<\EOF
-  rm -f conftest.frag conftest.h
-  echo "/* $ac_file.  Generated automatically by configure.  */" > conftest.h
-  cat conftest.in >> conftest.h
-  rm -f conftest.in
-  if cmp -s $ac_file conftest.h 2>/dev/null; then
-    echo "$ac_file is unchanged"
-    rm -f conftest.h
-  else
-    # Remove last slash and all that follows it.  Not all systems have dirname.
-      ac_dir=`echo $ac_file|sed 's%/[^/][^/]*$%%'`
-      if test "$ac_dir" != "$ac_file" && test "$ac_dir" != .; then
-      # The file is in a subdirectory.
-      test ! -d "$ac_dir" && mkdir "$ac_dir"
-    fi
-    rm -f $ac_file
-    mv conftest.h $ac_file
-  fi
-fi; done
 
 EOF
 cat >> $CONFIG_STATUS <<EOF
Index: configure.in
===================================================================
RCS file: /cvs/uberbaum/gdb/testsuite/configure.in,v
retrieving revision 1.10
diff -u -p -r1.10 configure.in
--- configure.in	1 May 2003 01:09:51 -0000	1.10
+++ configure.in	29 Jul 2003 19:10:33 -0000
@@ -21,7 +21,6 @@
 
 AC_PREREQ(2.13)
 AC_INIT(gdb.base)
-AC_CONFIG_HEADER(config.h:config.hin)
 
 AC_CANONICAL_BUILD
 AC_CANONICAL_HOST
Index: gdb.mi/pthreads.c
===================================================================
RCS file: /cvs/uberbaum/gdb/testsuite/gdb.mi/pthreads.c,v
retrieving revision 1.2
diff -u -p -r1.2 pthreads.c
--- gdb.mi/pthreads.c	24 Sep 2002 19:36:06 -0000	1.2
+++ gdb.mi/pthreads.c	29 Jul 2003 19:10:37 -0000
@@ -1,26 +1,4 @@
 #include <stdio.h>
-
-#include "config.h"
-
-#ifndef HAVE_PTHREAD_H
-
-/* Don't even try to compile.  In fact, cause a syntax error that we can
-   look for as a compiler error message and know that we have no pthread
-   support.  In that case we can just suppress the test completely. */
-
-#error "no posix threads support"
-
-#else
-
-/* OK.  We have the right header.  If we try to compile this and fail, then
-   there is something wrong and the user should know about it so the testsuite
-   should issue an ERROR result.. */
-
-#ifdef __linux__
-#define  _MIT_POSIX_THREADS 1	/* GNU/Linux (or at least RedHat 4.0)
-                                   needs this */
-#endif
-
 #include <pthread.h>
 
 /* Under OSF 2.0 & 3.0 and HPUX 10, the second arg of pthread_create
@@ -77,4 +55,3 @@ main (int argc, char *argv[])
   return 0;
 }
 
-#endif /* ifndef HAVE_PTHREAD_H */
Index: gdb.threads/pthreads.c
===================================================================
RCS file: /cvs/uberbaum/gdb/testsuite/gdb.threads/pthreads.c,v
retrieving revision 1.3
diff -u -p -r1.3 pthreads.c
--- gdb.threads/pthreads.c	24 Feb 2002 22:56:08 -0000	1.3
+++ gdb.threads/pthreads.c	29 Jul 2003 19:10:37 -0000
@@ -1,26 +1,4 @@
 #include <stdio.h>
-
-#include "config.h"
-
-#ifndef HAVE_PTHREAD_H
-
-/* Don't even try to compile.  In fact, cause a syntax error that we can
-   look for as a compiler error message and know that we have no pthread
-   support.  In that case we can just suppress the test completely. */
-
-#error "no posix threads support"
-
-#else
-
-/* OK.  We have the right header.  If we try to compile this and fail, then
-   there is something wrong and the user should know about it so the testsuite
-   should issue an ERROR result.. */
-
-#ifdef __linux__
-#define  _MIT_POSIX_THREADS 1	/* GNU/Linux (or at least RedHat 4.0)
-                                   needs this */
-#endif
-
 #include <pthread.h>
 
 /* Under OSF 2.0 & 3.0 and HPUX 10, the second arg of pthread_create
@@ -168,4 +146,3 @@ main(argc, argv)
   exit(0);
 }
 
-#endif	/* ifndef HAVE_PTHREAD_H */

Re: [RFA] testsuite cleanups (part 2)

[Michael Snyder]

Elena Zannoni wrote:

>Here, I am observing a peculiar situation. If you run the testsuite in
>a directory that is not the build directory, i.e. the gdb.blah object
>dirs are recreated outside of the build tree, you will see that all
>the tests which use pthreads.c will not compile. This is because
>pthreads.c includes "config.h" which gets created in the build
>gdb/testsuite directory at configure time. Config.h was used to
>determine if there is posix thread support or not, but now we use the
>method of trying to compile with various -lblah so we don't really
>need this anymore.  
>
>It took me a while to notice this failure because I have the bad habit
>of looking just at the unexpected failures, glossing over the
>unsupported results.
>
>OK?
>
>elena
>
>[PS. I assume I used the correct autoconf, but I am nost sure: I used the
>autoconf-000227 snapshot].
>

In view of the large diffs in configure, I suspect this is indeed the wrong
version of autoconf. The rest of the change is approved. If you don't
easily solve the autoconf issue, you might consider leaving configure
alone -- it's just an extra test...

Re: [RFA] testsuite cleanups (part 2)

[Andrew Cagney]

> In view of the large diffs in configure, I suspect this is indeed the wrong
> version of autoconf.

No, Elena's autoconf version is correct - its adding back sitefile stuff 
accidently deleted by an earlier re-generate.  It's just that no one 
noticed.  Happens all the time :-/

Andrew

(oh for autoconf 2.5x)

Re: [RFA] testsuite cleanups (part 2)

[Elena Zannoni]

Michael Snyder writes:
 > Elena Zannoni wrote:
 > 
 > >Here, I am observing a peculiar situation. If you run the testsuite in
 > >a directory that is not the build directory, i.e. the gdb.blah object
 > >dirs are recreated outside of the build tree, you will see that all
 > >the tests which use pthreads.c will not compile. This is because
 > >pthreads.c includes "config.h" which gets created in the build
 > >gdb/testsuite directory at configure time. Config.h was used to
 > >determine if there is posix thread support or not, but now we use the
 > >method of trying to compile with various -lblah so we don't really
 > >need this anymore.  
 > >
 > >It took me a while to notice this failure because I have the bad habit
 > >of looking just at the unexpected failures, glossing over the
 > >unsupported results.
 > >
 > >OK?
 > >
 > >elena
 > >
 > >[PS. I assume I used the correct autoconf, but I am nost sure: I used the
 > >autoconf-000227 snapshot].
 > >
 > 
 > In view of the large diffs in configure, I suspect this is indeed the wrong
 > version of autoconf. The rest of the change is approved. If you don't
 > easily solve the autoconf issue, you might consider leaving configure
 > alone -- it's just an extra test...
 > 
 > 

Or the previous configure was generated with the incorrect version of
autoconf. :-)  I know for sure that the autoconf-000227.tar is the right
one for the gdb/configure, and, I would estrapolate, it is the right one
for gdb/testsuite/configure as well.

elena

Re: [RFA] testsuite cleanups (part 2)

[Elena Zannoni]

Elena Zannoni writes:
 > 
 > 
 > Here, I am observing a peculiar situation. If you run the testsuite in
 > a directory that is not the build directory, i.e. the gdb.blah object
 > dirs are recreated outside of the build tree, you will see that all
 > the tests which use pthreads.c will not compile. This is because
 > pthreads.c includes "config.h" which gets created in the build
 > gdb/testsuite directory at configure time. Config.h was used to
 > determine if there is posix thread support or not, but now we use the
 > method of trying to compile with various -lblah so we don't really
 > need this anymore.  
 > 
 > It took me a while to notice this failure because I have the bad habit
 > of looking just at the unexpected failures, glossing over the
 > unsupported results.
 > 
 > OK?
 > 
 > elena
 > 
 > [PS. I assume I used the correct autoconf, but I am nost sure: I used the
 > autoconf-000227 snapshot].
 > 
 > 2003-07-29  Elena Zannoni  <ezannoni@redhat.com>
 >         
 >         * configure.in: Don't generate config.h from config.hin.
 >         * configure: Regenerate.
 >         * config.h: Remove file.
 >         * gdb.mi/pthreads.c: Remove include of config.h, and other ifdefs.
 >         * gdb.threads/pthreads.c: Ditto.


committed mainline&branch

elena

Re: [RFA] testsuite cleanups (part 2)

[Michael Elizabeth Chastain]

eza> Config.h was used to determine if there is posix thread support or not,
eza> but now we use the method of trying to compile with various -lblah so we
eza> don't really need this anymore.  

Ah, I've been looking forward to this!  No need to create and use a
config.h at all.  If the test program doesn't compile, then key off that
to give an UNSUPPORTED.  That works great for me with tls.exp and
tls-shared.exp and gcc 2.95.3.

Michael C

Re: [RFA] testsuite cleanups (part 2)

[Michael Elizabeth Chastain]

eza> [PS. I assume I used the correct autoconf, but I am nost sure: I used the
eza> autoconf-000227 snapshot].

ms> In view of the large diffs in configure, I suspect this is indeed the
ms> wrong version of autoconf. The rest of the change is approved. If you
ms> don't easily solve the autoconf issue, you might consider leaving
ms> configure alone -- it's just an extra test...

A little sniffing indicates that the last commit to 'configure.in'
used autoconf 2.13 to generate 'configure'.  That's where the diff
is coming from.

So either the blessed version has changed, or the last commit to
'configure.in' was done with the wrong autoconf.

I agree it might be useful to split the patch, if we can't get a
quick resolution to "which autoconf is blessed".

Michael C

Re: [RFA] testsuite cleanups (part 2)

[Elena Zannoni]

Michael Elizabeth Chastain writes:
 > eza> [PS. I assume I used the correct autoconf, but I am nost sure: I used the
 > eza> autoconf-000227 snapshot].
 > 
 > ms> In view of the large diffs in configure, I suspect this is indeed the
 > ms> wrong version of autoconf. The rest of the change is approved. If you
 > ms> don't easily solve the autoconf issue, you might consider leaving
 > ms> configure alone -- it's just an extra test...
 > 
 > A little sniffing indicates that the last commit to 'configure.in'
 > used autoconf 2.13 to generate 'configure'.  That's where the diff
 > is coming from.

I think that 000227 also brands itself as 2.13. I think this may be
where the confusion comes from, the configure file tells you
"configure generated by autoconf version 2.13" in both cases.

 > 
 > So either the blessed version has changed, or the last commit to
 > 'configure.in' was done with the wrong autoconf.
 > 
 > I agree it might be useful to split the patch, if we can't get a
 > quick resolution to "which autoconf is blessed".
 > 

Ok, this issue seems to be resolved now. Let me know when I can commit,
or not :-).

 > Michael C

elena

Re: [RFA] testsuite cleanups (part 2)

[Michael Snyder]

Elena Zannoni wrote:

>Michael Elizabeth Chastain writes:
> > eza> [PS. I assume I used the correct autoconf, but I am nost sure: I used the
> > eza> autoconf-000227 snapshot].
> > 
> > ms> In view of the large diffs in configure, I suspect this is indeed the
> > ms> wrong version of autoconf. The rest of the change is approved. If you
> > ms> don't easily solve the autoconf issue, you might consider leaving
> > ms> configure alone -- it's just an extra test...
> > 
> > A little sniffing indicates that the last commit to 'configure.in'
> > used autoconf 2.13 to generate 'configure'.  That's where the diff
> > is coming from.
>
>I think that 000227 also brands itself as 2.13. I think this may be
>where the confusion comes from, the configure file tells you
>"configure generated by autoconf version 2.13" in both cases.
>
> > 
> > So either the blessed version has changed, or the last commit to
> > 'configure.in' was done with the wrong autoconf.
> > 
> > I agree it might be useful to split the patch, if we can't get a
> > quick resolution to "which autoconf is blessed".
> > 
>
>Ok, this issue seems to be resolved now. Let me know when I can commit,
>or not :-).
>

Go. ;-)

Re: [RFA] testsuite cleanups (part 2)

[Michael Elizabeth Chastain]

eza> I think that 000227 also brands itself as 2.13.

I just checked and this is correct, autoconf-000227 writes
"autoconf 2.13" into the header of its output.  Ouch!

eza> Ok, this issue seems to be resolved now.
eza> Let me know when I can commit, or not :-).

Spin, spin, spin ... gotta spin another hour or two.  Sigh.

Michael C

Re: [RFA] testsuite cleanups (part 2)

[Michael Elizabeth Chastain]

Proofread, nothing objectionable found.
Tested, no regressions or improvements.
Recommended for approval.

Michael C

===

2003-07-29  Elena Zannoni  <ezannoni@redhat.com>
        
        * configure.in: Don't generate config.h from config.hin.
        * configure: Regenerate.
        * config.h: Remove file.
        * gdb.mi/pthreads.c: Remove include of config.h, and other ifdefs.
        * gdb.threads/pthreads.c: Ditto.