Thank you, Jan-Henrik. I forked the latest master and it appears that it does not build, unlike the 5.8.1 distribution load that is delivered in a tgz file. I did bootstrap the freshly checked out code with two minor warnings about "underquoted definition of AM_PATH_GTK". Then, I ran the "configure" script, which finished with no warnings or errors. Then, when I ran make, it failed in the bison-generated code, I believe. See the build log below. Is it a known issue?
$ ./configure --prefix=/home/smaslyak/tmp/monit-fix-bin --without-pam --without-ssl --enable-optimized --without-largefiles ... $ make -j4 /sea/local/bin/flex -i -osrc/lex.yy.c src/l.l bison -y -dvt -o src/y.tab.c src/p.y bison -y -dvt -o src/y.tab.c src/p.y echo "#include <config.h>" > src/.y.tab.c echo "#include <config.h>" > src/.y.tab.c cat src/y.tab.c >> src/.y.tab.c cat src/y.tab.c >> src/.y.tab.c mv src/.y.tab.c src/y.tab.c mv src/.y.tab.c src/y.tab.c mv src/y.tab.h src/tokens.h mv: cannot stat `src/.y.tab.c': No such file or directory make: [src/y.tab.c] Error 1 (ignored) mv src/y.tab.h src/tokens.h mv: cannot stat `src/y.tab.h': No such file or directory make: [src/y.tab.c] Error 1 (ignored) ... ... gcc -DHAVE_CONFIG_H -I. -I./src -DLINUX -DSYSCONFDIR="\"/home/smaslyak/tmp/monit-fix-bin/etc\"" -I./src -I./src/device -I./src/http -I./src/process -I./src/protocols -I./libmonit/src -Wno-pointer-sign -O2 -O3 -Wall -Wunused -Wno-unused-label -funsigned-char -D_GNU_SOURCE -std=c99 -D _REENTRANT -c -o src/monit.o src/monit.c src/p.y:126: error: redefinition of 'struct IHavePrecedence' src/p.y:132: error: redefinition of 'struct myrate' src/p.y:155: error: redefinition of 'cfg_errflag' src/p.y:155: error: previous definition of 'cfg_errflag' was here src/p.y:156: error: redefinition of 'tail' src/p.y:156: error: previous definition of 'tail' was here src/p.y:157: error: redefinition of 'current' src/p.y:157: error: previous definition of 'current' was here src/p.y:158: error: redefinition of 'urlrequest' src/p.y:158: error: previous definition of 'urlrequest' was here ... ... On Fri, May 30, 2014 at 5:30 AM, Jan-Henrik Haukeland <[email protected]> wrote: > This switch has been in the build code for a long time and I think it was > required as we were using 64-bits API interface. It is unclear if it can be > removed and build Monit for x86 without code changes. We only use and test > Solaris on 64-bits. If you want to change this and *can test and verify > that Monit works on Solaris 32-bits*, sure, we'll be happy to accept a Pull > Request. > > On 29 May 2014, at 00:01, Sergey Maslyakov <[email protected]> wrote: > > > Any opinions on this? Somebody must have put the "-m64" switch there on > purpose. > > > > If vendors do not object, I could open an issue to remove this switch > for Solaris builds and to let the compiler or the user decide whether to > force 64-bit mode or not. I think the flag makes it more restrictive than > it should and results in build failures with 32-bit compilers on > Solaris/X86. > > > > Please let me know if this worth doing and I will self-assign the ticket > and contribute the fix. But if it will be rejected, I'd like to know this > before I start. > > > > > > Thank you, > > /Sergey > > > -- > To unsubscribe: > https://lists.nongnu.org/mailman/listinfo/monit-general >
-- To unsubscribe: https://lists.nongnu.org/mailman/listinfo/monit-general
