Already in 5.21.1 there was the same very long line,
many .o-files compiled together:
;;; Emitting code for MAXIMA-LOAD-PATHNAME-DIRECTORY.
;;; Note:
;;; Invoking external command:
;;; /home/csoliver/SAT-Algorithmen/OKplatform/ExternalSources/Installations/Gcc/4.1.2/bin/gcc "-I/home/csoliver/SAT-Algorithmen/OKplatform/ExternalSources/Installations/Ecl/10.4.1/include/" -D_GNU_SOURCE -D_FILE_OFFSET_BITS=64 -g -O2 -fPIC -I/home/csoliver/SAT-Algorithmen/OKplatform/ExternalSources/Installations/Gmp/4.1.2/5.0.1/include -Dlinux -O -w -c "/home/csoliver/SAT-Algorithmen/OKplatform/ExternalSources/builds/Maxima/ecl/maxima-5.21.1/src/binary-ecl/init-cl.c" -o "/home/csoliver/SAT-Algorithmen/OKplatform/ExternalSources/builds/Maxima/ecl/maxima-5.21.1/src/binary-ecl/init-cl.o"
;;;
;;; Finished compiling /home/csoliver/SAT-Algorithmen/OKplatform/ExternalSources/builds/Maxima/ecl/maxima-5.21.1/src/init-cl.lisp.
;;;
;;; Note:
;;; Invoking external command:
;;; /home/csoliver/SAT-Algorithmen/OKplatform/ExternalSources/Installations/Gcc/4.1.2/bin/gcc "-I/home/csoliver/SAT-Algorithmen/OKplatform/ExternalSources/Installations/Ecl/10.4.1/include/" -D_GNU_SOURCE -D_FILE_OFFSET_BITS=64 -g -O2 -fPIC -I/home/csoliver/SAT-Algorithmen/OKplatform/ExternalSources/Installations/Gmp/4.1.2/5.0.1/include -Dlinux -O -w -c "/tmp/ECLINITzvY1cp.c" -o "/tmp/ECLINITzvY1cp.o"
;;;
;;; Note:
;;; Invoking external command:
;;; /home/csoliver/SAT-Algorithmen/OKplatform/ExternalSources/Installations/Gcc/4.1.2/bin/gcc -o "/home/csoliver/SAT-Algorithmen/OKplatform/ExternalSources/builds/Maxima/ecl/maxima-5.21.1/src/binary-ecl/init-cl.fas" -L"/home/csoliver/SAT-Algorithmen/OKplatform/ExternalSources/Installations/Ecl/10.4.1/lib/" "/tmp/ECLINITzvY1cp.o" "/home/csoliver/SAT-Algorithmen/OKplatform/ExternalSources/builds/Maxima/ecl/maxima-5.21.1/src/binary-ecl/init-cl.o" -shared -Wl,-rpath=/home/csoliver/SAT-Algorithmen/OKplatform/ExternalSources/Installations/Ecl/10.4.1/lib -Wl,-rpath,/home/csoliver/SAT-Algorithmen/OKplatform/ExternalSources/Installations/Gmp/4.1.2/5.0.1/lib -Wl,-rpath=/home/csoliver/SAT-Algorithmen/OKplatform/ExternalSources/Installations/Ecl/10.4.1/lib -Wl,-rpath,/home/csoliver/SAT-Algorithmen/OKplatform/ExternalSources/Installations/Gmp/4.1.2/5.0.1/lib -L/home/csoliver/SAT-Algorithmen/OKplatform/ExternalSources/Installations/Gmp/4.1.2/5.0.1/lib -lecl -lgmp -ldl -lm
;;;
; - Loading binary file "binary-ecl/init-cl.fas"
;;; Loading "/home/csoliver/SAT-Algorithmen/OKplatform/ExternalSources/builds/Maxima/ecl/maxima-5.21.1/src/binary-ecl/init-cl.fas"
; - Providing system maxima
;;; Note:
;;; Invoking external command:
;;; /home/csoliver/SAT-Algorithmen/OKplatform/ExternalSources/Installations/Gcc/4.1.2/bin/gcc "-I/home/csoliver/SAT-Algorithmen/OKplatform/ExternalSources/Installations/Ecl/10.4.1/include/" -D_GNU_SOURCE -D_FILE_OFFSET_BITS=64 -g -O2 -fPIC -I/home/csoliver/SAT-Algorithmen/OKplatform/ExternalSources/Installations/Gmp/4.1.2/5.0.1/include -Dlinux -O -w -c "/tmp/ECLINITEsajvv.c" -o "/tmp/ECLINITEsajvv.o"
;;;
;;; Note:
;;; Invoking external command:
;;; /home/csoliver/SAT-Algorithmen/OKplatform/ExternalSources/Installations/Gcc/4.1.2/bin/gcc -o "/home/csoliver/SAT-Algorithmen/OKplatform/ExternalSo XXX here is comes now XXX
it seems that already then it was very close to the limit of what the tradiditonal Unix/Linux tools can handle,
and now, by some changed paths, it just got big enough to break it.
On Sun, Dec 26, 2010 at 01:58:38PM -0500, Raymond Toy wrote:
> On 12/26/10 1:46 PM, Oliver Kullmann wrote:
> > Further to the build-problem:
> >
> > It seems it is the very long line: on another similar machine, but
> > where longer command-lines can be handled, the build succeeds.
> >
> > So did something drastically change with 5.23.0 (in which case perhaps
> > one can avoid the long line)?
> I don't recall any "drastic" changes. It would certainly help if I knew
> which file was producing the long line. But that doesn't seem like a
> maxima problem but a ecl and/or gcc problem.
It is a well-known problem, that traditional Unix/Linux system have bounds on the number
of command-line arguments and their lengths. Don't remember the details, but all what can
be done here in general(!) is to update to a newer version of the whole tool-chain (actually I believe the
Linux kernel must support it; updating the bash won't help), which is the standard on
newer Linux distributions.
Don't remember the details, but I waisted enough time on that issue some time ago.
It's a problem of the build-system. One could side-step it, but so well, likely best is to
propose to those who experience the problem (there will be others, on older Linux/Unix's, with the Maxima-installation deep
inside the tree) to update their system.
> > We got some test failures:
> >
> > Error summary:
> > Error found in /home/csoliver/OKplatform/ExternalSources/builds/Maxima/ecl/maxima-5.23.0/tests/rtest14.mac, problem:
> > (144)
> This has been there for as long as I can remember. I've never gotten
> around to really investigating why ecl has this problem.
> > Errors found in /home/csoliver/OKplatform/ExternalSources/builds/Maxima/ecl/maxima-5.23.0/tests/rtest16.mac, problems:
> > (384 385)
> These failures are because maxima assumes cl:expt produces a
> double-float result even if one arg is a double-float and the other is
> an integer. (expt x y) is defined to be (exp (* y (log x))). If x is
> an integer, (log x) returns a single-float. If y is a double-float the
> result will be a double-float, but only have single-float precision.
>
> The maxima project agreed that this won't be fixed in maxima.
>
Maxima could generalise the test, taking Ecl into account ...
In the long run, it likely will just save time if those errors disappear (i.e.,
are hidden). And there are always those who panic if they see some error.
So well, enough time spent in front of the computer in the office today!
Oliver