This is a discussion on Compile GNU Octave with SunStudio within the comp.unix.solaris forums, part of the Solaris Operating System category; --> I'm trying to compile GNU Octave 2.9.17 with SunStudio 12 on a Ferrari 4000 with 2GB memory $ ./configure ...
| |||||||
| FAQ | Members List | Calendar | Search | Today's Posts | Mark Forums Read |
| ||||
| I'm trying to compile GNU Octave 2.9.17 with SunStudio 12 on a Ferrari 4000 with 2GB memory $ ./configure CC=cc CXX="CC" F77=f77 --disable-readline $ gmake After an hour or so I get this: [...] "../liboctave/ArrayN.h", line 91: Where: Instantiated from non-template code. Front end: Not enough space Fork for exec of /opt/SUNWspro/prod/bin/fbe failed Failure in /opt/SUNWspro/prod/bin/fbe, status = 0xfedd36d4 Fatal Error exec'ing /opt/SUNWspro/prod/bin/fbe gmake[2]: *** [pic/op-i8-i8.o] Error 1 gmake[2]: Leaving directory `/opt/local/src/octave-2.9.17/src' gmake[1]: *** [src] Error 2 gmake[1]: Leaving directory `/opt/local/src/octave-2.9.17' gmake: *** [all] Error 2 Is this due to lack of memory or something else??? |
| |||
| Thommy M. wrote: > I'm trying to compile GNU Octave 2.9.17 with SunStudio 12 on a Ferrari > 4000 with 2GB memory > > $ ./configure CC=cc CXX="CC" F77=f77 --disable-readline > $ gmake > > After an hour or so I get this: > > [...] > "../liboctave/ArrayN.h", line 91: Where: Instantiated from > non-template code. > Front end: Not enough space > Fork for exec of /opt/SUNWspro/prod/bin/fbe failed > Failure in /opt/SUNWspro/prod/bin/fbe, status = 0xfedd36d4 > Fatal Error exec'ing /opt/SUNWspro/prod/bin/fbe > gmake[2]: *** [pic/op-i8-i8.o] Error 1 > gmake[2]: Leaving directory `/opt/local/src/octave-2.9.17/src' > gmake[1]: *** [src] Error 2 > gmake[1]: Leaving directory `/opt/local/src/octave-2.9.17' > gmake: *** [all] Error 2 > > > > Is this due to lack of memory or something else??? > > > > My first guess would be lack of memory (RAM). 2 GB seems generous so that's ONLY a guess. Do you have a nice BIG swap partition configured? If not, can you add 2-4 GB swap partition or file? |
| |||
| Thommy M. wrote: > I'm trying to compile GNU Octave 2.9.17 with SunStudio 12 on a Ferrari > 4000 with 2GB memory > > $ ./configure CC=cc CXX="CC" F77=f77 --disable-readline > $ gmake > > After an hour or so I get this: > > [...] > "../liboctave/ArrayN.h", line 91: Where: Instantiated from > non-template code. > Front end: Not enough space > Fork for exec of /opt/SUNWspro/prod/bin/fbe failed > Failure in /opt/SUNWspro/prod/bin/fbe, status = 0xfedd36d4 > Fatal Error exec'ing /opt/SUNWspro/prod/bin/fbe > gmake[2]: *** [pic/op-i8-i8.o] Error 1 > gmake[2]: Leaving directory `/opt/local/src/octave-2.9.17/src' > gmake[1]: *** [src] Error 2 > gmake[1]: Leaving directory `/opt/local/src/octave-2.9.17' > gmake: *** [all] Error 2 > > > > Is this due to lack of memory or something else??? > > > > What shell are you using to compile this? I tried it on a SPARC (8 GB RAM), but whilst it configures OK, the gmake command bombs out immediatly, with some message about "/bin/sh: -e: not found" |
| |||
| On Nov 17, 12:09 pm, Dave <nowh...@nowhere.com> wrote: > Thommy M. wrote: > > I'm trying to compile GNU Octave 2.9.17 with SunStudio 12 on a Ferrari > > 4000 with 2GB memory > > $ ./configure CC=cc CXX="CC" F77=f77 --disable-readline > > $ gmake > > After an hour or so I get this: > > [...] > > "../liboctave/ArrayN.h", line 91: Where: Instantiated from > > non-template code. > > Front end: Not enough space > > Fork for exec of /opt/SUNWspro/prod/bin/fbe failed > > Failure in /opt/SUNWspro/prod/bin/fbe, status = 0xfedd36d4 > > Fatal Error exec'ing /opt/SUNWspro/prod/bin/fbe > > gmake[2]: *** [pic/op-i8-i8.o] Error 1 > > gmake[2]: Leaving directory `/opt/local/src/octave-2.9.17/src' > > gmake[1]: *** [src] Error 2 > > gmake[1]: Leaving directory `/opt/local/src/octave-2.9.17' > > gmake: *** [all] Error 2 > > Is this due to lack of memory or something else??? > > What shell are you using to compile this? > > I tried it on a SPARC (8 GB RAM), but whilst it configures OK, the gmake > command bombs out immediatly, with some message about "/bin/sh: -e: not > found" I dont get this at all - what happens for me on a 2 GB Blade SPARC is mostly errors like this: CC -c -KPIC -I. -I.. -I../liboctave -I../src -I../libcruft/misc - DHAVE_CONFIG_H -g ./OPERATORS/op-i8-i8.cc -o pic/op-i8-i8.o "./pr-output.cc", line 1605: Error: "static" is not allowed here. 1 Error(s) detected. The lines are: template <> static inline void Use g++ and no complaints. I dont know how to fix this "error" and as its all over the place so it would take a LOT of fixing. Maybe use gcc/g++ instead : < |
| |||
| Thommy M. wrote: > Dave wrote: >> Thommy M. wrote: >> >>>> I tried exactly the same as those posted by the original poster: >>> Well, that was me... ;-) >> I realised that later! >> >>> Might be that I'm on SXCE b76 x86 and we have different gmake versions >> I was using gmake 3.80, but I tried the same version as you, which I got >> from Blastwave, and it still bombs out. > > > > Mine is standard OpenSolaris. Might be different from blastwave... Possibly, but I am surprised it is to fussy, if that is the case. Nothing else I know if has this problem. I tried tried building Octave on another Sun - this time an Ultra 60, with an older version of Solaris 10 main-webserver # cat /etc/release Solaris 10 3/05 s10_74L2a SPARC Copyright 2005 Sun Microsystems, Inc. All Rights Reserved. Use is subject to license terms. Assembled 22 January 2005 (the other machine is a Blade 2000 with the latest release of Solaris 10). It will not build on the Ultra 60 either. Same problem. main-webserver # gmake ************************************************** ********* * * To compile Octave, you will need a recent versions of * the following software: * * g++ (3.2.x or a more recent version) * * flex (2.5.4 or a more recent version) -- required if * you need to recreate lex.cc from lex.l * * bison (1.31 or a more recent version) -- required if * you need to recreate parse.cc from parse.y * * gperf (3.0.1 or a more recent version) -- required if * you need to recreate oct-gperf.h from octave.gperf * * Now would be a good time to read INSTALL.OCTAVE if * you have not done so already. * ************************************************** ********* gmake -f octMakefile all gmake[1]: -e: Command not found gmake[1]: Entering directory `/tmp/octave-2.9.17' gmake[1]: -e: Command not found making octave-bug from octave-bug.in /bin/sh: -e: not found /bin/sh: -e: not found gmake[1]: *** [octave-bug] Error 1 gmake[1]: Leaving directory `/tmp/octave-2.9.17' gmake: *** [all] Error 2 |
| |||
| Thommy M. wrote: > Mine is standard OpenSolaris. Might be different from blastwave... > > > Anyway, as there are older versions of Octave both on blastwave and > sunfreeware it might be worth asking the maintainers how they compiled > their versions... > I tried to rebuild the configure scritp, but automake complains kestrel /tmp/octave-2.9.17 # automake configure.in: no proper invocation of AM_INIT_AUTOMAKE was found. configure.in: You should verify that configure.in invokes AM_INIT_AUTOMAKE, configure.in: that aclocal.m4 is present in the top-level directory, configure.in: and that aclocal.m4 was recently regenerated (using aclocal). automake: no `Makefile.am' found for any configure output There is no line with AM_INIT_AUTOMAKE in configure.in, despite the manual http://sources.redhat.com/automake/a....html#Examples saying there should be! |
| |||
| Dave wrote: > Thommy M. wrote: > >> Mine is standard OpenSolaris. Might be different from blastwave... >> >> >> Anyway, as there are older versions of Octave both on blastwave and >> sunfreeware it might be worth asking the maintainers how they compiled >> their versions... >> > > > I tried to rebuild the configure scritp, but automake complains > > > > kestrel /tmp/octave-2.9.17 # automake > configure.in: no proper invocation of AM_INIT_AUTOMAKE was found. > configure.in: You should verify that configure.in invokes AM_INIT_AUTOMAKE, > configure.in: that aclocal.m4 is present in the top-level directory, > configure.in: and that aclocal.m4 was recently regenerated (using aclocal). > automake: no `Makefile.am' found for any configure output > > > There is no line with AM_INIT_AUTOMAKE in configure.in, despite the manual > > http://sources.redhat.com/automake/a....html#Examples > > saying there should be! Are you surprised? |
| |||
| Thommy M. wrote: >> There is no line with AM_INIT_AUTOMAKE in configure.in, despite the manual >> >> http://sources.redhat.com/automake/a....html#Examples >> >> saying there should be! > > Are you surprised? I think you know the answer to that one! It seems like Octave needs a bit of work from the developers. I think the problem for far too many open-source projects is people test on linux, and nothing much else. |
| |||
| Thommy M. wrote: > Dave wrote: >> Thommy M. wrote: >> >>> Mine is standard OpenSolaris. Might be different from blastwave... >>> >>> >>> Anyway, as there are older versions of Octave both on blastwave and >>> sunfreeware it might be worth asking the maintainers how they compiled >>> their versions... >>> >> >> I tried to rebuild the configure scritp, but automake complains >> >> >> >> kestrel /tmp/octave-2.9.17 # automake >> configure.in: no proper invocation of AM_INIT_AUTOMAKE was found. >> configure.in: You should verify that configure.in invokes AM_INIT_AUTOMAKE, >> configure.in: that aclocal.m4 is present in the top-level directory, >> configure.in: and that aclocal.m4 was recently regenerated (using aclocal). >> automake: no `Makefile.am' found for any configure output >> >> >> There is no line with AM_INIT_AUTOMAKE in configure.in, despite the manual >> >> http://sources.redhat.com/automake/a....html#Examples >> >> saying there should be! > > Are you surprised? I used the sdt. make, GCC, and built it on a rather large array (about 200Gb free) with no issues once I had all of the dependencies installed. (System was an Ultra 80, 4Gb RAM, 4 x 450MHz processors). Ed |
| ||||
| Dave wrote: > Thommy M. wrote: > >>> There is no line with AM_INIT_AUTOMAKE in configure.in, despite the >>> manual >>> >>> http://sources.redhat.com/automake/a....html#Examples >>> >>> saying there should be! >> >> Are you surprised? > > I think you know the answer to that one! ;-) > It seems like Octave needs a bit of work from the developers. I think > the problem for far too many open-source projects is people test on > linux, and nothing much else. True. Linux has made as much bad as good for many GNU tools/programs. Remember the days when SunOS was the development platform of choice for most GNU projects... |