Opened 12 years ago
Last modified 12 years ago
#9722 closed defect
PARI/GP build error on Fedora 13 — at Version 50
Reported by: | mpatel | Owned by: | tbd |
---|---|---|---|
Priority: | blocker | Milestone: | sage-4.5.3 |
Component: | packages: standard | Keywords: | FLTK, DSO, libstdc++, plotfltk, FL/Fl.H |
Cc: | ggrafendorfer, leif, jdemeyer, jsp | Merged in: | |
Authors: | Leif Leonhardy | Reviewers: | Jeroen Demeyer |
Report Upstream: | N/A | Work issues: | |
Branch: | Commit: | ||
Dependencies: | Stopgaps: |
Description (last modified by )
New spkg now available:
This package includes
which should apply in succession to the existing pari-2.3.5.p2.spkg
.
Reported by Georg Grafendorfer on sage-release:
sage does not built on my machine, AMD Phenom(tm) II X4 925, Fedora 13, here are the relevant lines of the log file: Shall we try to build pari 2.3.5 (released) now (y/n)? [n] Ok. Type "make install" when you are ready Bye ! Building and install PARI make[2]: Entering directory `/scratch/sage-4.5.3.alpha0/spkg/build/ pari-2.3.5.p2/src' Making gp in Olinux-x86_64 make[3]: Entering directory `/scratch/sage-4.5.3.alpha0/spkg/build/ pari-2.3.5.p2/src/Olinux-x86_64' File ../src/funclist not changed. ../config/genkernel ../src/kernel/x86_64/asm0.h > parilvl0.h cat ../src/kernel/gmp/tune.h ../src/kernel/gmp/int.h ../src/kernel/ none/level1.h > parilvl1.h cat parilvl0.h parilvl1.h > pariinl.h gcc -c -O1 -Wall -fno-strict-aliasing -fomit-frame-pointer -I. - I../src/headers -I../src/language -I/scratch/sage-4.5.3.alpha0/local/ include -o gp.o ../src/gp/gp.c cd ../src/desc && /usr/bin/perl merge_822 ../functions/*/* > def-linux- x86_64-6659.tmp mv ../src/desc/def-linux-x86_64-6659.tmp ../src/desc/pari.desc cd ../src/desc && /usr/bin/perl gen_proto gp pari.desc > gp_init-linux- x86_64-6659.tmp mv ../src/desc/gp_init-linux-x86_64-6659.tmp ../src/gp/gp_init.h gcc -c -O1 -Wall -fno-strict-aliasing -fomit-frame-pointer -I. - I../src/headers -I../src/graph -o gp_init.o ../src/gp/gp_init.c gcc -c -O1 -Wall -fno-strict-aliasing -fomit-frame-pointer -I. - I../src/headers -I../src/language -I/scratch/sage-4.5.3.alpha0/local/ include -o gp_rl.o ../src/gp/gp_rl.c cd ../src/desc && /usr/bin/perl gen_proto highlevel pari.desc > highlvl-linux-x86_64-6659.tmp mv ../src/desc/highlvl-linux-x86_64-6659.tmp ../src/gp/highlvl.h gcc -c -O1 -Wall -fno-strict-aliasing -fomit-frame-pointer -I. - I../src/headers -DDL_DFLT_NAME=NULL -o highlvl.o ../src/gp/highlvl.c gcc -c -O1 -Wall -fno-strict-aliasing -fomit-frame-pointer -I. - I../src/headers -o whatnow.o ../src/gp/whatnow.c gcc -c -O1 -Wall -fno-strict-aliasing -fomit-frame-pointer -I. - I../src/headers -I../src/graph -o plotport.o ../src/graph/plotport.c g++ -c -O1 -Wall -fno-strict-aliasing -fomit-frame-pointer -I. -I../ src/headers -I"/usr/local"/include -o plotfltk.o ../src/graph/ plotfltk.c ../src/graph/plotfltk.c:27:19: error: FL/Fl.H: No such file or directory ../src/graph/plotfltk.c:28:26: error: FL/Fl_Window.H: No such file or directory ../src/graph/plotfltk.c:29:24: error: FL/fl_draw.H: No such file or directory ../src/graph/plotfltk.c:31: error: expected class-name before '{' token ../src/graph/plotfltk.c:45: error: 'Fl_Color' does not name a type ../src/graph/plotfltk.c:48: error: 'Fl_Color' does not name a type ../src/graph/plotfltk.c: In constructor 'Plotter::Plotter(long int*, long int*, long int*, long int, const char*)': ../src/graph/plotfltk.c:56: error: class 'Plotter' does not have any field named 'Fl_Window' ../src/graph/plotfltk.c:60: error: 'color' was not declared in this scope ../src/graph/plotfltk.c:60: error: 'FL_WHITE' was not declared in this scope ../src/graph/plotfltk.c:61: error: 'FL_BLACK' was not declared in this scope ../src/graph/plotfltk.c:62: error: 'FL_BLUE' was not declared in this scope ../src/graph/plotfltk.c:63: error: 'rgb_color' was not declared in this scope ../src/graph/plotfltk.c:64: error: 'FL_RED' was not declared in this scope ../src/graph/plotfltk.c:65: error: 'FL_GREEN' was not declared in this scope ../src/graph/plotfltk.c: In function 'void DrawPoint(void*, long int, long int)': ../src/graph/plotfltk.c:73: error: 'fl_point' was not declared in this scope ../src/graph/plotfltk.c: In function 'void DrawLine(void*, long int, long int, long int, long int)': ../src/graph/plotfltk.c:79: error: 'fl_line' was not declared in this scope ../src/graph/plotfltk.c: In function 'void DrawRectangle(void*, long int, long int, long int, long int)': ../src/graph/plotfltk.c:85: error: 'fl_rect' was not declared in this scope ../src/graph/plotfltk.c: In function 'void DrawPoints(void*, long int, plot_points*)': ../src/graph/plotfltk.c:92: error: 'fl_point' was not declared in this scope ../src/graph/plotfltk.c: In function 'void SetForeground(void*, long int)': ../src/graph/plotfltk.c:97: error: 'Fl_Color' was not declared in this scope ../src/graph/plotfltk.c:97: error: 'color' was not declared in this scope ../src/graph/plotfltk.c:97: error: expected primary-expression before ')' token ../src/graph/plotfltk.c:97: error: expected ';' before 'data' ../src/graph/plotfltk.c:98: error: 'fl_color' was not declared in this scope ../src/graph/plotfltk.c: In function 'void DrawLines(void*, long int, plot_points*)': ../src/graph/plotfltk.c:105: error: 'fl_line' was not declared in this scope ../src/graph/plotfltk.c: In function 'void DrawString(void*, long int, long int, char*, long int)': ../src/graph/plotfltk.c:111: error: 'fl_draw' was not declared in this scope ../src/graph/plotfltk.c: In member function 'void Plotter::draw()': ../src/graph/plotfltk.c:118: error: 'class Plotter' has no member named 'w' ../src/graph/plotfltk.c:119: error: 'class Plotter' has no member named 'h' ../src/graph/plotfltk.c:121: error: 'FL_COURIER' was not declared in this scope ../src/graph/plotfltk.c:121: error: 'fl_font' was not declared in this scope ../src/graph/plotfltk.c:122: error: 'FL_WHITE' was not declared in this scope ../src/graph/plotfltk.c:122: error: 'fl_color' was not declared in this scope ../src/graph/plotfltk.c:123: error: 'class Plotter' has no member named 'w' ../src/graph/plotfltk.c:123: error: 'class Plotter' has no member named 'h' ../src/graph/plotfltk.c:123: error: 'fl_rectf' was not declared in this scope ../src/graph/plotfltk.c:133: error: 'color' was not declared in this scope ../src/graph/plotfltk.c: In member function 'int Plotter::handle(int)': ../src/graph/plotfltk.c:140: error: 'FL_PUSH' was not declared in this scope ../src/graph/plotfltk.c:141: error: 'Fl' has not been declared ../src/graph/plotfltk.c:155: error: 'class Plotter' has no member named 'x' ../src/graph/plotfltk.c:156: error: 'class Plotter' has no member named 'y' ../src/graph/plotfltk.c:157: error: 'class Plotter' has no member named 'w' ../src/graph/plotfltk.c:158: error: 'class Plotter' has no member named 'h' ../src/graph/plotfltk.c:159: error: 'class Plotter' has no member named 'fullscreen' ../src/graph/plotfltk.c:163: error: 'class Plotter' has no member named 'fullscreen_off' ../src/graph/plotfltk.c:164: error: 'class Plotter' has no member named 'size_range' ../src/graph/plotfltk.c: In function 'void rectdraw0(long int*, long int*, long int*, long int)': ../src/graph/plotfltk.c:188: error: 'Fl' has not been declared ../src/graph/plotfltk.c:188: error: 'FL_DOUBLE' was not declared in this scope ../src/graph/plotfltk.c:188: error: 'FL_INDEX' was not declared in this scope ../src/graph/plotfltk.c:190: error: 'class Plotter' has no member named 'size_range' ../src/graph/plotfltk.c:191: error: 'class Plotter' has no member named 'box' ../src/graph/plotfltk.c:191: error: 'FL_FLAT_BOX' was not declared in this scope ../src/graph/plotfltk.c:192: error: 'class Plotter' has no member named 'end' ../src/graph/plotfltk.c:193: error: 'class Plotter' has no member named 'show' ../src/graph/plotfltk.c:194: error: 'Fl' has not been declared make[3]: *** [plotfltk.o] Error 1 make[3]: Leaving directory `/scratch/sage-4.5.3.alpha0/spkg/build/ pari-2.3.5.p2/src/Olinux-x86_64' make[2]: *** [gp] Error 2 make[2]: Leaving directory `/scratch/sage-4.5.3.alpha0/spkg/build/ pari-2.3.5.p2/src' Error building GP
There's now a dedicated thread.
Change History (57)
comment:1 follow-up: ↓ 2 Changed 12 years ago by
comment:2 in reply to: ↑ 1 ; follow-up: ↓ 3 Changed 12 years ago by
Replying to mpatel:
Georg, if I've correctly guessed your trac name, could you please update the Sage trac account name-real name map?
done, Georg
comment:3 in reply to: ↑ 2 ; follow-up: ↓ 4 Changed 12 years ago by
As suggested by leif i have made the admins here (eth zürich) install fltk-devel (fltk is installed anyway), still does not work, the log message looks different now:
Shall we try to build pari 2.3.5 (released) now (y/n)? [n] Ok. Type "make install" when you are ready Bye ! Building and install PARI make[2]: Entering directory `/scratch/sage-4.5.3.alpha0/spkg/build/pari-2.3.5.p2/src' Making gp in Olinux-x86_64 make[3]: Entering directory `/scratch/sage-4.5.3.alpha0/spkg/build/pari-2.3.5.p2/src/Olinux-x86_64' File ../src/funclist not changed. ../config/genkernel ../src/kernel/x86_64/asm0.h > parilvl0.h cat ../src/kernel/gmp/tune.h ../src/kernel/gmp/int.h ../src/kernel/none/level1.h > parilvl1.h cat parilvl0.h parilvl1.h > pariinl.h gcc -c -O1 -Wall -fno-strict-aliasing -fomit-frame-pointer -I. -I../src/headers -I../src/language -I/scratch/sage-4.5.3.alpha0/local/include -o gp.o ../src/gp/gp.c cd ../src/desc && /usr/bin/perl merge_822 ../functions/*/* > def-linux-x86_64-379.tmp mv ../src/desc/def-linux-x86_64-379.tmp ../src/desc/pari.desc cd ../src/desc && /usr/bin/perl gen_proto gp pari.desc > gp_init-linux-x86_64-379.tmp mv ../src/desc/gp_init-linux-x86_64-379.tmp ../src/gp/gp_init.h gcc -c -O1 -Wall -fno-strict-aliasing -fomit-frame-pointer -I. -I../src/headers -I../src/graph -o gp_init.o ../src/gp/gp_init.c gcc -c -O1 -Wall -fno-strict-aliasing -fomit-frame-pointer -I. -I../src/headers -I../src/language -I/scratch/sage-4.5.3.alpha0/local/include -o gp_rl.o ../src/gp/gp_rl.c cd ../src/desc && /usr/bin/perl gen_proto highlevel pari.desc > highlvl-linux-x86_64-379.tmp mv ../src/desc/highlvl-linux-x86_64-379.tmp ../src/gp/highlvl.h gcc -c -O1 -Wall -fno-strict-aliasing -fomit-frame-pointer -I. -I../src/headers -DDL_DFLT_NAME=NULL -o highlvl.o ../src/gp/highlvl.c gcc -c -O1 -Wall -fno-strict-aliasing -fomit-frame-pointer -I. -I../src/headers -o whatnow.o ../src/gp/whatnow.c gcc -c -O1 -Wall -fno-strict-aliasing -fomit-frame-pointer -I. -I../src/headers -I../src/graph -o plotport.o ../src/graph/plotport.c g++ -c -O1 -Wall -fno-strict-aliasing -fomit-frame-pointer -I. -I../src/headers -I"/usr/local"/include -o plotfltk.o ../src/graph/plotfltk.c cat ../src/kernel/gmp/mp.c ../src/kernel/none/cmp.c ../src/kernel/none/gcdll.c ../src/kernel/none/ratlift.c ../src/kernel/none/invmod.c ../src/kernel/gmp/gcd.c ../src/kernel/none/mp_indep.c ../src/kernel/none/add.c > mp.c gcc -c -O1 -Wall -fno-strict-aliasing -fomit-frame-pointer -fPIC -I. -I../src/headers -I/scratch/sage-4.5.3.alpha0/local/include -o mp.o mp.c gcc -c -O1 -Wall -fno-strict-aliasing -fomit-frame-pointer -I. -I../src/headers -fPIC -o mpinl.o ../src/kernel/none/mpinl.c gcc -c -O1 -Wall -fno-strict-aliasing -fomit-frame-pointer -I. -I../src/headers -fPIC -o Flx.o ../src/basemath/Flx.c gcc -c -O1 -Wall -fno-strict-aliasing -fomit-frame-pointer -I. -I../src/headers -fPIC -o Qfb.o ../src/basemath/Qfb.c gcc -c -O1 -Wall -fno-strict-aliasing -fomit-frame-pointer -I. -I../src/headers -fPIC -o RgX.o ../src/basemath/RgX.c gcc -c -O1 -Wall -fno-strict-aliasing -fomit-frame-pointer -I. -I../src/headers -fPIC -o alglin1.o ../src/basemath/alglin1.c gcc -c -O1 -Wall -fno-strict-aliasing -fomit-frame-pointer -I. -I../src/headers -fPIC -o alglin2.o ../src/basemath/alglin2.c gcc -c -O1 -Wall -fno-strict-aliasing -fomit-frame-pointer -I. -I../src/headers -fPIC -o arith1.o ../src/basemath/arith1.c gcc -c -O1 -Wall -fno-strict-aliasing -fomit-frame-pointer -I. -I../src/headers -fPIC -o arith2.o ../src/basemath/arith2.c gcc -c -O1 -Wall -fno-strict-aliasing -fomit-frame-pointer -I. -I../src/headers -fPIC -o base1.o ../src/basemath/base1.c gcc -c -O1 -Wall -fno-strict-aliasing -fomit-frame-pointer -I. -I../src/headers -fPIC -o base2.o ../src/basemath/base2.c gcc -c -O1 -Wall -fno-strict-aliasing -fomit-frame-pointer -I. -I../src/headers -fPIC -o base3.o ../src/basemath/base3.c gcc -c -O1 -Wall -fno-strict-aliasing -fomit-frame-pointer -I. -I../src/headers -fPIC -o base4.o ../src/basemath/base4.c gcc -c -O1 -Wall -fno-strict-aliasing -fomit-frame-pointer -I. -I../src/headers -fPIC -o base5.o ../src/basemath/base5.c gcc -c -O1 -Wall -fno-strict-aliasing -fomit-frame-pointer -I. -I../src/headers -fPIC -o bibli1.o ../src/basemath/bibli1.c gcc -c -O1 -Wall -fno-strict-aliasing -fomit-frame-pointer -I. -I../src/headers -fPIC -o bibli2.o ../src/basemath/bibli2.c gcc -c -O1 -Wall -fno-strict-aliasing -fomit-frame-pointer -I. -I../src/headers -fPIC -o buch1.o ../src/basemath/buch1.c gcc -c -O1 -Wall -fno-strict-aliasing -fomit-frame-pointer -I. -I../src/headers -fPIC -o buch2.o ../src/basemath/buch2.c gcc -c -O1 -Wall -fno-strict-aliasing -fomit-frame-pointer -I. -I../src/headers -fPIC -o buch3.o ../src/basemath/buch3.c gcc -c -O1 -Wall -fno-strict-aliasing -fomit-frame-pointer -I. -I../src/headers -fPIC -o buch4.o ../src/basemath/buch4.c gcc -c -O1 -Wall -fno-strict-aliasing -fomit-frame-pointer -I. -I../src/headers -fPIC -o galconj.o ../src/basemath/galconj.c gcc -c -O1 -Wall -fno-strict-aliasing -fomit-frame-pointer -I. -I../src/headers -fPIC -o gen1.o ../src/basemath/gen1.c gcc -c -O1 -Wall -fno-strict-aliasing -fomit-frame-pointer -I. -I../src/headers -fPIC -o gen2.o ../src/basemath/gen2.c gcc -c -O1 -Wall -fno-strict-aliasing -fomit-frame-pointer -I. -I../src/headers -fPIC -o gen3.o ../src/basemath/gen3.c gcc -c -O1 -Wall -fno-strict-aliasing -fomit-frame-pointer -I. -I../src/headers -fPIC -o ifactor1.o ../src/basemath/ifactor1.c gcc -c -O1 -Wall -fno-strict-aliasing -fomit-frame-pointer -I. -I../src/headers -fPIC -o perm.o ../src/basemath/perm.c gcc -c -O1 -Wall -fno-strict-aliasing -fomit-frame-pointer -I. -I../src/headers -fPIC -o polarit1.o ../src/basemath/polarit1.c gcc -c -O1 -Wall -fno-strict-aliasing -fomit-frame-pointer -I. -I../src/headers -fPIC -o polarit2.o ../src/basemath/polarit2.c gcc -c -O1 -Wall -fno-strict-aliasing -fomit-frame-pointer -I. -I../src/headers -fPIC -o polarit3.o ../src/basemath/polarit3.c gcc -c -O1 -Wall -fno-strict-aliasing -fomit-frame-pointer -I. -I../src/headers -fPIC -o rootpol.o ../src/basemath/rootpol.c gcc -c -O1 -Wall -fno-strict-aliasing -fomit-frame-pointer -I. -I../src/headers -fPIC -o subcyclo.o ../src/basemath/subcyclo.c gcc -c -O1 -Wall -fno-strict-aliasing -fomit-frame-pointer -I. -I../src/headers -fPIC -o subgroup.o ../src/basemath/subgroup.c gcc -c -O1 -Wall -fno-strict-aliasing -fomit-frame-pointer -I. -I../src/headers -fPIC -o trans1.o ../src/basemath/trans1.c gcc -c -O1 -Wall -fno-strict-aliasing -fomit-frame-pointer -I. -I../src/headers -fPIC -o trans2.o ../src/basemath/trans2.c gcc -c -O1 -Wall -fno-strict-aliasing -fomit-frame-pointer -I. -I../src/headers -fPIC -o trans3.o ../src/basemath/trans3.c cd ../src/desc && /usr/bin/perl gen_member pari.desc > members-linux-x86_64-379.tmp mv ../src/desc/members-linux-x86_64-379.tmp ../src/language/members.h gcc -c -O1 -Wall -fno-strict-aliasing -fomit-frame-pointer -I. -I../src/headers -fPIC -o anal.o ../src/language/anal.c gcc -c -O1 -Wall -fno-strict-aliasing -fomit-frame-pointer -I. -I../src/headers -fPIC -o compat.o ../src/language/compat.c gcc -c -O1 -Wall -fno-strict-aliasing -fomit-frame-pointer -I. -I../src/headers -fPIC -o default.o ../src/language/default.c gcc -c -O1 -Wall -fno-strict-aliasing -fomit-frame-pointer -I. -I../src/headers -fPIC -o errmsg.o ../src/language/errmsg.c gcc -c -O1 -Wall -fno-strict-aliasing -fomit-frame-pointer -I. -I../src/headers -fPIC -o es.o ../src/language/es.c cd ../src/desc && /usr/bin/perl gen_proto basic pari.desc > init-linux-x86_64-379.tmp mv ../src/desc/init-linux-x86_64-379.tmp ../src/language/init.h gcc -c -O1 -Wall -fno-strict-aliasing -fomit-frame-pointer -I. -I../src/headers -fPIC -o init.o ../src/language/init.c gcc -c -O1 -Wall -fno-strict-aliasing -fomit-frame-pointer -I. -I../src/headers -fPIC -o intnum.o ../src/language/intnum.c gcc -c -O1 -Wall -fno-strict-aliasing -fomit-frame-pointer -I. -I../src/headers -fPIC -o members.o ../src/language/members.c gcc -c -O1 -Wall -fno-strict-aliasing -fomit-frame-pointer -I. -I../src/headers -fPIC -o sumiter.o ../src/language/sumiter.c gcc -c -O1 -Wall -fno-strict-aliasing -fomit-frame-pointer -I. -I../src/headers -fPIC -o aprcl.o ../src/modules/aprcl.c gcc -c -O1 -Wall -fno-strict-aliasing -fomit-frame-pointer -I. -I../src/headers -fPIC -o elldata.o ../src/modules/elldata.c gcc -c -O1 -Wall -fno-strict-aliasing -fomit-frame-pointer -I. -I../src/headers -fPIC -o elliptic.o ../src/modules/elliptic.c gcc -c -O1 -Wall -fno-strict-aliasing -fomit-frame-pointer -I. -I../src/headers -fPIC -o galois.o ../src/modules/galois.c gcc -c -O1 -Wall -fno-strict-aliasing -fomit-frame-pointer -I. -I../src/headers -fPIC -o groupid.o ../src/modules/groupid.c gcc -c -O1 -Wall -fno-strict-aliasing -fomit-frame-pointer -I. -I../src/headers -fPIC -o kummer.o ../src/modules/kummer.c gcc -c -O1 -Wall -fno-strict-aliasing -fomit-frame-pointer -I. -I../src/headers -fPIC -o mpqs.o ../src/modules/mpqs.c gcc -c -O1 -Wall -fno-strict-aliasing -fomit-frame-pointer -I. -I../src/headers -fPIC -o nffactor.o ../src/modules/nffactor.c gcc -c -O1 -Wall -fno-strict-aliasing -fomit-frame-pointer -I. -I../src/headers -fPIC -o part.o ../src/modules/part.c gcc -c -O1 -Wall -fno-strict-aliasing -fomit-frame-pointer -I. -I../src/headers -fPIC -o stark.o ../src/modules/stark.c gcc -c -O1 -Wall -fno-strict-aliasing -fomit-frame-pointer -I. -I../src/headers -fPIC -o subfield.o ../src/modules/subfield.c gcc -c -O1 -Wall -fno-strict-aliasing -fomit-frame-pointer -I. -I../src/headers -fPIC -o thue.o ../src/modules/thue.c rm -f libpari-gmp.so.2.3.5 gcc -o libpari-gmp.so.2.3.5 -shared -O1 -Wall -fno-strict-aliasing -fomit-frame-pointer -fPIC -Wl,-shared,-soname=libpari-gmp.so.2 mp.o mpinl.o Flx.o Qfb.o RgX.o alglin1.o alglin2.o arith1.o arith2.o base1.o base2.o base3.o base4.o base5.o bibli1.o bibli2.o buch1.o buch2.o buch3.o buch4.o galconj.o gen1.o gen2.o gen3.o ifactor1.o perm.o polarit1.o polarit2.o polarit3.o rootpol.o subcyclo.o subgroup.o trans1.o trans2.o trans3.o anal.o compat.o default.o errmsg.o es.o init.o intnum.o members.o sumiter.o aprcl.o elldata.o elliptic.o galois.o groupid.o kummer.o mpqs.o nffactor.o part.o stark.o subfield.o thue.o -lc -ldl -lm -L/scratch/sage-4.5.3.alpha0/local/lib -lgmp if test "libpari-gmp.so.2.3.5" != "libpari.so"; then rm -f libpari.so; ln -s libpari-gmp.so.2.3.5 libpari.so; fi if test "libpari-gmp.so.2.3.5" != "libpari-gmp.so.2"; then rm -f libpari-gmp.so.2; ln -s libpari-gmp.so.2.3.5 libpari-gmp.so.2; fi rm -f gp-dyn gcc -o gp-dyn -O1 -Wall -fno-strict-aliasing -fomit-frame-pointer -Wl,--export-dynamic gp.o gp_init.o gp_rl.o highlvl.o whatnow.o plotport.o plotfltk.o -L"/scratch/sage-4.5.3.alpha0/spkg/build/pari-2.3.5.p2/src/Olinux-x86_64" -L/scratch/sage-4.5.3.alpha0/local/lib -lreadline -L/scratch/sage-4.5.3.alpha0/local/lib/ -ltermcap -L/scratch/sage-4.5.3.alpha0/local/lib -lpari -L"/usr/local"/lib -lfltk -ldl -lm -L/scratch/sage-4.5.3.alpha0/local/lib -lgmp /usr/bin/ld: plotfltk.o: undefined reference to symbol 'gxx_personality_v0@@CXXABI_1.3' /usr/bin/ld: note: 'gxx_personality_v0@@CXXABI_1.3' is defined in DSO /usr/lib64libstdc++.so.6 so try adding it to the linker command line /usr/lib64libstdc++.so.6: could not read symbols: Invalid operation collect2: ld returned 1 exit status make[3]: * [gp-dyn] Error 1 make[3]: Leaving directory `/scratch/sage-4.5.3.alpha0/spkg/build/pari-2.3.5.p2/src/Olinux-x86_64' make[2]: * [gp] Error 2 make[2]: Leaving directory `/scratch/sage-4.5.3.alpha0/spkg/build/pari-2.3.5.p2/src' Error building GP
real 0m45.782s user 0m41.047s sys 0m3.976s sage: An error occurred while installing pari-2.3.5.p2 Please email sage-devel http://groups.google.com/group/sage-devel explaining the problem and send the relevant part of of /scratch/sage-4.5.3.alpha0/install.log. Describe your computer, operating system, etc. If you want to try to fix the problem yourself, *don't* just cd to /scratch/sage-4.5.3.alpha0/spkg/build/pari-2.3.5.p2 and type 'make check' or whatever is appropriate. Instead, the following commands setup all environment variables correctly and load a subshell for you to debug the error: (cd '/scratch/sage-4.5.3.alpha0/spkg/build/pari-2.3.5.p2' && '/scratch/sage-4.5.3.alpha0/sage' -sh) When you are done debugging, you can type "exit" to leave the subshell. make[1]: * [installed/pari-2.3.5.p2] Error 1 make[1]: Leaving directory `/scratch/sage-4.5.3.alpha0/spkg'
real 35m5.899s user 24m57.842s sys 8m21.169s Error building Sage.
Georg
comment:4 in reply to: ↑ 3 ; follow-up: ↓ 5 Changed 12 years ago by
- Cc jdemeyer added
Replying to ggrafendorfer:
As suggested by leif i have made the admins here (eth zürich) install fltk-devel (fltk is installed anyway), still does not work, the log message looks different now:
... gcc -o gp-dyn -O1 -Wall -fno-strict-aliasing -fomit-frame-pointer -Wl,--export-dynamic gp.o gp_init.o gp_rl.o highlvl.o whatnow.o plotport.o plotfltk.o -L"/scratch/sage-4.5.3.alpha0/spkg/build/pari-2.3.5.p2/src/Olinux-x86_64" -L/scratch/sage-4.5.3.alpha0/local/lib -lreadline -L/scratch/sage-4.5.3.alpha0/local/lib/ -ltermcap -L/scratch/sage-4.5.3.alpha0/local/lib -lpari -L"/usr/local"/lib -lfltk -ldl -lm -L/scratch/sage-4.5.3.alpha0/local/lib -lgmp /usr/bin/ld: plotfltk.o: undefined reference to symbol '__gxx_personality_v0@@CXXABI_1.3' /usr/bin/ld: note: '__gxx_personality_v0@@CXXABI_1.3' is defined in DSO /usr/lib64//libstdc++.so.6 so try adding it to the linker command line /usr/lib64//libstdc++.so.6: could not read symbols: Invalid operation collect2: ld returned 1 exit status ...
That's just the usual Fedora 13 error. ;-)
The simplest solution would be to uninstall FLTK completely, but since you seem to not have admin rights, we should perhaps fix this in the PARI package.
There's already a new PARI Sage package underway, which is afaik intended to get merged into Sage 4.6, but I don't believe you want to wait for that.
So we have to decide if we want to provide a new intermediate PARI 2.3.5.p3 package here, to get merged into Sage 4.5.3. (I think we can solve this by a 9-character patch.)
Mitesh, your opinion?
Jeroen, perhaps you know if this is already fixed upstream/in svn 12577. (There are various ways to fix this.)
comment:5 in reply to: ↑ 4 ; follow-up: ↓ 6 Changed 12 years ago by
There's already a new PARI Sage package underway, which is afaik intended to get merged into Sage 4.6, but I don't believe you want to wait for that.
I'm not doing numerical calculations right now, but I'm trying to spread sage here (eth zürich), which is a bit difficult if it does not built ;-), some of us are using it for teaching, the semester starts in middle of September, would be great to fix it until then,
btw, this bug exists since version after 4.3.4, which was the last which built correctly, I reported it two times before, but was ignored (once a workaround was given)),
Georg
comment:6 in reply to: ↑ 5 ; follow-up: ↓ 7 Changed 12 years ago by
Replying to ggrafendorfer:
btw, this bug exists since version after 4.3.4, which was the last which built correctly,
Hmmm, I don't think this error showed up before Fedora 13 was released... ;-)
I reported it two times before, but was ignored (once a workaround was given)),
Where?
Anyway, I think I have a trivial patch/new spkg ready.
Are you familiar with patching spkgs? (Then you could apply it yourself immediately - or make sure it works on Fedora 13; I've only tested it on Ubuntu, where this is not an issue. It does not fix the first error, i.e. PARI attempting to use FLTK despite no headers being present, since I currently cannot reproduce that. I could perhaps fix that later though.)
comment:7 in reply to: ↑ 6 ; follow-up: ↓ 9 Changed 12 years ago by
Replying to leif:
Replying to ggrafendorfer:
btw, this bug exists since version after 4.3.4, which was the last which built correctly,
Hmmm, I don't think this error showed up before Fedora 13 was released... ;-)
We had Fedora 11 before, where the same error occured, my second report is here:
Anyway, I think I have a trivial patch/new spkg ready.
Are you familiar with patching spkgs?
no, is it difficult, where is the new spkgs?
It does not fix the first error, i.e. PARI attempting to use FLTK despite no headers being present, since I currently cannot reproduce that. I could perhaps fix that later though.)
Since all the machines are the same, it should also fix the first error, since this is obviously the default state of the machines,
Georg
comment:8 Changed 12 years ago by
- Status changed from new to needs_review
I've uploaded a patch to fix the second (i.e. the linker) error. Yet only tested on Ubuntu; should work for Fedora 13...
=== pari-2.3.5.p3 (Leif Leonhardy, August 11th 2010) === * #9722: Explicitly link against libstdc++ if FLTK is used, to support Fedora 13. * TODO/FIXME: On Fedora 13, PARI tries to use FLTK even if just the library (as opposed to the required developer) package of FLTK is installed.
comment:9 in reply to: ↑ 7 Changed 12 years ago by
Replying to ggrafendorfer:
Replying to leif:
Replying to ggrafendorfer:
btw, this bug exists since version after 4.3.4, which was the last which built correctly,
Hmmm, I don't think this error showed up before Fedora 13 was released... ;-)
I of course meant the linker error.
We had Fedora 11 before, where the same error occured, my second report is here:
Nice. A trivial work-around, if you simply want to compile Sage/PARI without FLTK is:
export PARI_EXTRA_OPTS="--with-fltk=no"
(This will complain that the directory .../no
does not exist, but proceed building without FLTK.)
Anyway, I think I have a trivial patch/new spkg ready.
Are you familiar with patching spkgs?
no, is it difficult, where is the new spkgs?
I don't have an account on sage.math, so I could only put it into the volatile ftp directory. (But someone else could move it.)
Applying patches to spkgs is relatively simple. If you have Mercurial (hg
) installed:
$ tar xjf /path/to/pari-2.3.5.p2.spkg $ mv pari-2.3.5.p2 pari-2.3.5.p3 $ cd pari-2.3.5.p3 $ hg import /path/to/spkg-patch.patch # replace with appropriate filename $ cd .. $ tar cjf pari-2.3.5.p3.spkg pari-2.3.5.p3
Then you can install the new spkg the usual way (./sage -i ...
, but perhaps first copy it over to $SAGE_ROOT/spkg/standard
to get around limitations of sage-spkg
, and then type ./sage -i pari-2.3.5.p3
).
It does not fix the first error, i.e. PARI attempting to use FLTK despite no headers being present, since I currently cannot reproduce that. I could perhaps fix that later though.)
Since all the machines are the same, it should also fix the first error, since this is obviously the default state of the machines,
Ok, I've read this too late...
Wouldn't be bad then to install libfltk-dev
on all machines... (I would assume the machines get configured/installed from a central server anyway.)
comment:10 Changed 12 years ago by
- Cc jsp added
comment:11 follow-ups: ↓ 14 ↓ 15 Changed 12 years ago by
ggrafendorfer: could you please try to see if the problem is there when you compile PARI/GP by itself? That is, download http://pari.math.u-bordeaux.fr/pub/pari/unix/pari-2.3.5.tar.gz, do ./Configure && make
and see what happens. If you can, please try it both with fltk-devel uninstalled and installed. If it doesn't compile like that, it should be reported upstream.
comment:12 follow-up: ↓ 13 Changed 12 years ago by
If it helps, I don't mind merging an intermediate / incremental PARI package in the 4.5.3 series.
Leif, I can make your package available on sage.math. Since it's ~2 MB, just send it to me by email.
Although I haven't tried it myself, it might be easier to set up and use spkg-upload.
comment:13 in reply to: ↑ 12 Changed 12 years ago by
Replying to mpatel:
If it helps, I don't mind merging an intermediate / incremental PARI package in the 4.5.3 series.
Ok.
Leif, I can make your package available on sage.math. Since it's ~2 MB, just send it to me by email.
I'm going to set up a Fedora box and will try to address the configuration error as well.
Although I haven't tried it myself, it might be easier to set up and use spkg-upload.
I'd need credentials from Minh, but the sun is just rising down under... :)
comment:14 in reply to: ↑ 11 Changed 12 years ago by
Replying to jdemeyer:
ggrafendorfer: could you please try to see if the problem is there when you compile PARI/GP by itself? That is, download http://pari.math.u-bordeaux.fr/pub/pari/unix/pari-2.3.5.tar.gz, do
./Configure && make
and see what happens. If you can, please try it both with fltk-devel uninstalled and installed. If it doesn't compile like that, it should be reported upstream.
Done, on two different machines, one with ftlk + fltk-devel, one with fltk but without fltk-devel installed, both on Fedora 13 64bit, both failed after successful configuration, I reported this to the pari project,
Georg
comment:15 in reply to: ↑ 11 ; follow-up: ↓ 16 Changed 12 years ago by
Here is my report to the pari project and the answer:
Hi,
there are some problems with the pari package when building sage (http://www.sagemath.org) reported here:
http://trac.sagemath.org/sage_trac/ticket/9722
I tried to build pari separately (stand alone) but also does not work:
After successful configuration pari does not build in 2 different cases:
AMD Phenom II X4 925, Fedora 13, with fltk and fltk-devel installed: for log-message see file "log_with_fltk-devel" attached
Esprimo 64bit, Fedora 13, with fltk, but not fltk-devel installed: for log-message see file "log_without_fltk-devel" attached
Hello Georg,
On both system, there is libfltk library installed in /usr/local/lib64/, but no matching header files in /usr/local/include is provided, so compiling fltk does not work. Configure output a warning:
### ### libX11.so not found. Please install X11 development files. ### They usually come in XFree86-devel (RPM) or xlibs-dev (Debian) packages ### ### X11 not found ###
So either install XFree86-devel (or whathever is the correct name of fedora) or use --graphic=none to disable graphics.
Installing fltk and fltk-devel does not work because the version of fltk it provides is not compatible with the one in /usr/local/lib64/ (this is only a guess).
Cheers, Bill.
comment:16 in reply to: ↑ 15 Changed 12 years ago by
- Keywords FLTK DSO libstdc++ plotfltk FL/Fl.H added
Replying to ggrafendorfer:
Here is my report to the pari project and the answer:
Hi,
there are some problems with the pari package when building sage (http://www.sagemath.org) reported here:
http://trac.sagemath.org/sage_trac/ticket/9722
I tried to build pari separately (stand alone) but also does not work:
After successful configuration pari does not build in 2 different cases:
AMD Phenom II X4 925, Fedora 13, with fltk and fltk-devel installed: for log-message see file "log_with_fltk-devel" attached
Esprimo 64bit, Fedora 13, with fltk, but not fltk-devel installed: for log-message see file "log_without_fltk-devel" attached
Hello Georg,
On both system, there is libfltk library installed in /usr/local/lib64/, but no matching header files in /usr/local/include is provided, so compiling fltk does not work. Configure output a warning:
### ### libX11.so not found. Please install X11 development files. ### They usually come in XFree86-devel (RPM) or xlibs-dev (Debian) packages ### ### X11 not found ###
So either install XFree86-devel (or whathever is the correct name of fedora) or use --graphic=none to disable graphics.
Installing fltk and fltk-devel does not work because the version of fltk it provides is not compatible with the one in /usr/local/lib64/ (this is only a guess).
Cheers, Bill.
In any case, PARI shouldn't attempt to build with FLTK support if no (or not matching) header files are present; the second (linker) error is also definitely a bug, which currently only shows up on Fedora 13 (and derivatives) due to stricter linkage rules.
I couldn't (yet) reproduce the errors on 32-bit Fedora 13 though, most probably because other headers and libraries weren't installed. (I tested installing Sage's PARI spkg with only fltk-1.1.10-1.fc13
, and then in addition fltk-devel-1.1.10-1.fc13
installed; both with PARI 2.3.5 and the new PARI 2.4.3 svn snapshot from #9343.)
I'll examine that later in more detail.
comment:17 Changed 12 years ago by
P.S.: Btw, we do (unconditionally) configure PARI with --graphic=none
...
(Although the user can set PARI_EXTRA_OPTS
to pass additional configuration options; cf. my trivial work-around above.)
comment:18 follow-up: ↓ 19 Changed 12 years ago by
Should we reply to the PARI developers with Leif's points?
For now, can we bypass PARI's FLTK checks altogether because/when we use --graphic=none
?
comment:19 in reply to: ↑ 18 Changed 12 years ago by
Replying to mpatel:
Should we reply to the PARI developers with Leif's points?
For now, can we bypass PARI's FLTK checks altogether because/when we use
--graphic=none
?
Cf. also #9343. I couldn't reproduce PARI attempting to build FLTK on any system, including Fedora 13 (32-bit though).
And we already pass --graphic=none
to PARI's Configure
...
(I wonder if Georg wanted PARI to use X11/FLTK.)
The missing libstdc++ in plotfltk
's libraries is of course a bug (which my patch solves).
comment:20 Changed 12 years ago by
s/build FLTK/build using FLTK/
comment:21 Changed 12 years ago by
We could in addition bypass (if false && test ...
) this part of Configure
for use with Sage:
... if test "$optimization" != profiling; then . ./get_X11 # X11, X11_INC, X11_LIBS . ./get_graphic_lib # which_graphic_lib . ./get_fltk # FLTKDIR, FLTK_LIBS . ./get_Qt # QTDIR, QTLIB echo "Hi-Res Graphics: $which_graphic_lib" . ./get_readline # $_readline_list (includes 'readline') else which_graphic_lib=none fi ...
But that's not very nice, and unless I can reproduce the problem...
And this would of course break graphics support for other users completely.
comment:22 follow-up: ↓ 25 Changed 12 years ago by
Georg, could you perhaps attach (upload) the two logfiles you sent to the PARI developers here, too?
comment:23 Changed 12 years ago by
Georg, could you please check your environment settings?
user@host$ printenv | grep fltk
I found PARI not resetting/initializing variables presumably intended to be locally used only...
comment:24 Changed 12 years ago by
- Status changed from needs_review to needs_info
Changed 12 years ago by
Changed 12 years ago by
comment:25 in reply to: ↑ 22 ; follow-ups: ↓ 26 ↓ 27 Changed 12 years ago by
Replying to leif:
Georg, could you perhaps attach (upload) the two logfiles you sent to the PARI developers here, too?
done, and here is my environment setting:
ggeorg@maschke% printenv | grep fltk ggeorg@maschke% rpm -qa | grep fltk fltk-devel-1.1.10-1.fc13.x86_64 fltk-1.1.10-1.fc13.x86_64 ggeorg@maschke%
comment:26 in reply to: ↑ 25 Changed 12 years ago by
Replying to ggrafendorfer:
Replying to leif:
Georg, could you perhaps attach (upload) the two logfiles you sent to the PARI developers here, too?
done, and here is my environment setting:
ggeorg@maschke% printenv | grep fltk ggeorg@maschke% rpm -qa | grep fltk fltk-devel-1.1.10-1.fc13.x86_64 fltk-1.1.10-1.fc13.x86_64 ggeorg@maschke%
Hmmm, I could reproduce your compilation error on Ubuntu, but only by e.g. export with_fltk=yes
(with FLTK, but not its header files/devel package installed)...
comment:27 in reply to: ↑ 25 ; follow-up: ↓ 28 Changed 12 years ago by
Replying to ggrafendorfer:
Replying to leif:
Georg, could you perhaps attach (upload) the two logfiles you sent to the PARI developers here, too?
done
Thanks, does the same happen if you run ./Configure --graphic=none
?
comment:28 in reply to: ↑ 27 ; follow-up: ↓ 29 Changed 12 years ago by
Replying to leif:
Replying to ggrafendorfer:
Replying to leif:
Georg, could you perhaps attach (upload) the two logfiles you sent to the PARI developers here, too?
done
Thanks, does the same happen if you run
./Configure --graphic=none
?
yes, I'll attach the log file,
Georg
Changed 12 years ago by
comment:29 in reply to: ↑ 28 ; follow-up: ↓ 32 Changed 12 years ago by
- Status changed from needs_info to needs_work
Replying to ggrafendorfer:
Replying to leif:
Thanks, does the same happen if you run
./Configure --graphic=none
?yes, I'll attach the log file,
LOL, that's because you don't have X11 development packages installed! :)
config/get_fltk
:
if test -z "$with_fltk"; then case "$which_graphic_lib" in fltk) with_fltk=yes;; esac if test -z "$X11"; then with_fltk=yes; fi fi ...
PARI is really funny. I'll fix all three issues and upload a second patch here later...
comment:30 follow-up: ↓ 31 Changed 12 years ago by
Whatever you do, it would be good to report these issues upstream. The PARI developers usually accept patches which fix bugs (although they are sometimes slow).
I haven't followed up on this ticket so much, but does the new version (see #9343) also need patching?
comment:31 in reply to: ↑ 30 Changed 12 years ago by
Replying to jdemeyer:
Whatever you do, it would be good to report these issues upstream.
I thought you were our PARI spokesman... ;-)
(I guess you'll speak French [too], btw.)
The PARI developers usually accept patches which fix bugs (although they are sometimes slow).
:) I'm not that sure what the intended logic behind the graphics detection scripts is, so at least one of the fixes is rather targeted at PARI/GP in Sage.
So I'll first provide a patch for PARI 2.3.5[.p2] / Sage 4.5.3, then we can see what we report upstream and what we fix in (our) 2.4.3 svn / for #9343 / Sage 4.6.
(The answer Georg got from upstream did not really address his problem btw.)
I haven't followed up on this ticket so much, but does the new version (see #9343) also need patching?
Yes, definitely, despite the scripts having changed between these versions. I've also commented on that at #9343 a few hours ago...
comment:32 in reply to: ↑ 29 Changed 12 years ago by
Replying to leif:
PARI is really funny. I'll fix all three issues and upload a second patch here later...
I would have guessed that adding --graphic=none
makes Configure
skip all graphics detection tests.
I can make an updated spkg with your new patch and host it on sage.math.
comment:33 Changed 12 years ago by
I'd like to release 4.5.3.rc0 by Monday, 23 August, and if all goes well, 4.5.3 by the end of next week. Since the necessary changes here aren't trivial --- we'd need to test them on "all" platforms --- I suggest that we fix this ticket in the new PARI spkg at #9343, which I'll merge into 4.6.alpha0. This presumes that #9343, #9591, and #9592 will be ready. I'll send a message to sage-devel about their status.
comment:34 follow-up: ↓ 49 Changed 12 years ago by
Georg, could you look at Jeroen's comment at #9343 and let us know any results?
comment:35 Changed 12 years ago by
Cross-replying to leif:
I'm working on #9722 btw, still hoping that it will make it into 4.5.3, since I don't believe 4.6 will be ready in time, and/or don't expect people to install a brand new major release (despite the ".0" missing) right before a new semester starts.
That's great! Do you think, for any reason, it'll take more than two days, say, to fix all three problems?
Changed 12 years ago by
SPKG patch, apply to PARI 2.3.5.p2. Fixes Fedora 13 link error when FLTK is used. (Updated version, only doc changes.)
comment:36 Changed 12 years ago by
- Status changed from needs_work to needs_review
New PARI spkg (p3), with the attached patch applied:
http://spkg-upload.googlecode.com/files/pari-2.3.5.p3.spkg
This only fixes the linker error on Fedora 13 (that occurred when fltk-devel
was installed).
If you run into the compiler error, convince your system administrator to
- either install
fltk-devel
- or remove an (invalid) unversioned link to the FLTK library installed. (I.e., if only
fltk
is installed, there should be nolibfltk.so
, but just versioned filenames like e.g.libfltk.so.1.1
, perhaps including versioned symbolic links to the actual libraries.)
I'll perhaps later upload an improved (p4) spkg that also makes the FLTK-devel detection more robust, s.t. the compiler error shouldn't occur either, even with such unexpected links present.
comment:37 Changed 12 years ago by
- Reviewers set to Jeroen Demeyer
- Status changed from needs_review to positive_review
For me, this has a positive_review, because:
- I viewed the patch, it makes sense (fltk is c++ after all) and looks good.
- It doesn't break anything on various systems which I tried (none of these is Fedora though).
- It fixes the problem for leif (I have not tested this).
Leif: keep me up to date with your potential .p4. I assume this patch has to be ported also to #9343?
comment:38 Changed 12 years ago by
I've compiled a pari-2.3.5.p4-preview.spkg
(changes not yet checked in, still testing on some systems with various combinations of parameters and installed packages):
=== pari-2.3.5.p4 (Leif Leonhardy, August 23th 2010) === * #9722: "--graphic=none" is now respected by PARI. We still look for graphics libraries, but don't use them in that case (see below, too). * #9722: Improved patch to config/get_fltk (makes sure the header files are present, too, if the FLTK library was found - fixes compiler error on IMHO misconfigured installations). We still look for FLTK components even if "--graphic=none" was specified, and print appropriate messages, but then simply do not use them. * #9722: New patch to config/get_X11: - Added/corrected/clarified messages regarding what was (not) found, and if this is really relevant. - An X11 library in */lib64/* will be found now, too (if we do a 64-bit build); on Fedora, libX11.so is located in /usr/lib64. - Added comments. * spkg-install: - Clear (unset) lots of variables that might (unintentionally) be used by PARI; perhaps upstream should initialize them. - Clear GP_INSTALL_PREFIX, which is interpreted by PARI. - Added help & informational message regarding PARI_EXTRA_OPTS; in addition, allow the user to override Sage's "--graphic=none" since this now works (see above; we don't want to annoy users that perhaps previously made use of GP's plotting support). * Clean-up in spkg-install: - Quote more environment variables (including SAGE_LOCAL). - Consistently use $UNAME instead of `uname` (set in sage-env). - Added some comments. * NOTE: At least some of these changes should be incorporated into our PARI 2.4.3 (svn 12577, see #9343), and/or reported upstream. I haven't changed config/get_Qt. It seems Qt support is broken on newer systems (because Qt <=3 is expected), but this doesn't break the build, only Qt will not be used.
Contact me if you are interested in testing this, too. (Appreciated.)
-Leif
comment:39 follow-up: ↓ 40 Changed 12 years ago by
I've attached a diff between the p3 and the p4-preview spkg, in case anybody is interested... ;-)
Note that this currently is just a diff, not a proper Mercurial patch ("changeset").
comment:40 in reply to: ↑ 39 Changed 12 years ago by
Replying to leif:
I've attached a diff between the p3 and the p4-preview spkg, in case anybody is interested... ;-)
That's a lot of stuff to look at :-)
I am personally a fan of GNU autoconf
for these situations because a lot of these checks can be simplified. However, I tried in vain to convince the PARI/GP developers to use autoconf
. Maybe I should do it for Sage just to prove a point...
comment:41 follow-up: ↓ 42 Changed 12 years ago by
Some people call it GNU autocrap...
Also a matter of taste (I personally find the configure
scripts readable), but autotools often unnecessarily slow down the installation process, and it's easy to build bad scripts with them, too.
Btw, I've meanwhile prepared a new, slightly corrected (wrt. --graphic=none
) p4 spkg, with also some more comments and improved messages.
I'll update the attached diff here, and after a bit more testing also a link to the (hopefully final) spkg.
(More interesting are the diffs of get_X11
and get_fltk
, i.e. patches/get_{X11,fltk
}.)
comment:42 in reply to: ↑ 41 ; follow-up: ↓ 43 Changed 12 years ago by
Replying to leif:
Some people call it GNU autocrap...
I know some people are against it (for reasons which I never fully understand), but given the choice between editing a configure.ac script and PARI's mess, I will choose the former.
Changed 12 years ago by
Just to ease seeing some of the changes I made... (Do not apply, it's in the patch, too.)
Changed 12 years ago by
Just to ease seeing some of the changes I made... (Do not apply, it's in the patch, too.)
comment:43 in reply to: ↑ 42 Changed 12 years ago by
Replying to jdemeyer:
Replying to leif:
Some people call it GNU autocrap...
I know some people are against it (for reasons which I never fully understand), but given the choice between editing a configure.ac script and PARI's mess, I will choose the former.
I'm with Jeroen on that. Properly used, autoconf
can simplify the build process considerably.
Of course, any tool badly used can result in a mess. There's one part of Sage where Makefile.am
is a file that one is expected to edit for "site customisation". Clearly the person using it did not have a clue what he/she was doing.
I would strongly suggest the Pari developers move to autoconf/automake
. I've personally not used libtool
for development, so can't really comment on that.
Pari's system, where there are specific targets, is always going to be a problem when versions of operating systems change and people have different search order for their PATH. In contrast, the autoconf
system tests what actually works, using the settings the person building the code has.
Also, the autoconf/automake mailing lists are very active and helpful. In contrast, I've found with SCons
, which is another build tool used in Sage, that getting help is very difficult. That takes too much control away from the programmer. I know there's one part of Sage (I forget which) where a developer removed SCons code and replaced it by a Makefile.
Dave
Changed 12 years ago by
SPKG patch, now a proper Mercurial changeset. Apply to pari-2.3.5.p3.spkg. (Further improvements w.r.t. graphics support; some clean-up.)
comment:44 follow-up: ↓ 45 Changed 12 years ago by
After testing various combinations of installed (graphics) packages and PARI configure options ($PARI_EXTRA_OPTS
) on 64-bit Fedora 13, 32-bit and 64-bit Ubuntu 9.04, I'm now quite confident with it and will "release" PARI 2.3.5.p4.
I've replaced the p4-preview diff by a Mercurial patch, and will shortly provide a link to the "final" spkg, with all changes committed.
"Please build, test, and report! We'd love to hear about your experiences with this new package."TM
(I've so far only tested the spkg on top of Sage 4.5.3.alpha0 and 4.5.3.alpha1.)
comment:45 in reply to: ↑ 44 ; follow-up: ↓ 46 Changed 12 years ago by
Replying to leif:
[...] will shortly provide a link to the "final" spkg, with all changes committed.
New spkg now available: http://spkg-upload.googlecode.com/files/pari-2.3.5.p4.spkg
"Please build, test, and report! We'd love to hear about your experiences with this new package."TM
I think the easiest way to test this is to copy it into $SAGE_ROOT/spkg/standard
, set and export PARI_EXTRA_OPTS
as you like, and then do:
$ ./sage -f pari-2.3.5.p4 2>&1 | tee my_log-parameters.log
The "cumulative" spkg install log automatically created by Sage is $SAGE_ROOT/spkg/logs/pari-2.3.5.p4.log
.
Potential settings for PARI_EXTRA_OPTS
are:
- empty / unset (no graphics support; default)
--graphic=auto
(should use X11 if available)--with-fltk
--with-qt
(likely to not build on newer systems, requires some Qt development packages too, i.e. can lead to build errors; should disable graphics support if no Qt libraries are found)
Thanks for testing this; we really would like to release Sage 4.5.3[.rc0]...
P.S.: Should work on all reported machines at ETH Zürich, too (with or without fltk-devel
, even if --with-fltk
was specified). ;-)
Of course Sage can also be built from scratch with this spkg; just copy the spkg as mentioned above to spkg/standard
(Sage will automatically pick the newest), and then type make
or make build
.
(I've so far only tested the spkg on top of Sage 4.5.3.alpha0 and 4.5.3.alpha1.)
Also installed without problems on Sage 4.5.2, now running ptest
.
P.P.S.: The positive review of this ticket currently only applies to the p3 spkg.
comment:46 in reply to: ↑ 45 Changed 12 years ago by
- Status changed from positive_review to needs_info
comment:47 Changed 12 years ago by
- Status changed from needs_info to needs_review
comment:48 Changed 12 years ago by
I get a build failure with PARI_EXTRA_OPTS=--graphic=x11
on sage.math. I've put the log here.
comment:49 in reply to: ↑ 34 ; follow-up: ↓ 50 Changed 12 years ago by
Replying to mpatel:
Georg, could you look at Jeroen's comment at #9343 and let us know any results?
worked fine on fedora 13, http://trac.sagemath.org/sage_trac/ticket/9343#comment:307 sorry for the delay, it's my office machine,
Georg
comment:50 in reply to: ↑ 49 Changed 12 years ago by
- Description modified (diff)
I've updated the description with a link to the new package.
But does PARI 2.4.3.svn-12577 fix the problem anyway? Georg, could you try installing Jeroen's spkg with
./sage -f http://cage.ugent.be/~jdemeyer/sage/pari-2.4.3.svn-12577.p2.spkg
? Sage will not work properly without other the changes at #9343 (and related tickets; see NewPARI), but it would help to know. Thanks for your patience.
Georg, if I've correctly guessed your trac name, could you please update the Sage trac account name-real name map?