Posts
I wanted to set up a2ps 4.14 on an old SCO Unix machine (which suprsingly had a gcc compiler installed) for Transforming Ingres' legacy reports with PostScript
Before you set up a2ps, you have to set gperf up first. Its installation was a breeze, so I assumed that a2ps would install easily too.....However that was not the case; when reaching the 'make' phase I was getting the following error :
liba2ps.a is a collection of libraries, whose dissasembly reveals :
that 'va_copy' is referenced in the printlen.o library and printlen.c file.Examining printlen.c we find a call to it at line 102 :
va_copy (ap, args);
Also since it is a GLOB binding we gather that it is defined elsewhere and not inside printlen.o; as a matter of fact it is not defined in none of the a2ps libraries but belongs to the underlying c library (or gcc library in this case) with its definition appearing in the 'stdarg.h' header :
vi ./usr/gnu/lib/gcc-lib/i586-unknown-sysv5/2.95.3pl1/include/stdarg.h
/* Copy __gnuc_va_list into another variable of this type. */
#define __va_copy(dest, src) (dest) = (src)
#endif /* _STDARG_H */
the issue here is that the gcc version is outdated,hence the gcc library itself, and subsequnetly 'va_copy''s signature does not match the one of the prototype inside printlen.c
Getting a new version of gcc was not an option; I was happy enough to see that there was a gcc instalation after all,let alone installing a new one!
So I was set for a workaround; modify the macro's definition from :
/* Copy __gnuc_va_list into another variable of this type. */
#define __va_copy(dest, src) (dest) = (src)
#endif /* _STDARG_H */
to :
/* Copy __gnuc_va_list into another variable of this type. */
#define va_copy(d,s)
#endif /* _STDARG_H */
This little change satisfied 'make',compiled successfully and a2ps was set up, at last. Now a test was in order to see if it actually works; fortunatelly it did and a2ps transformed a text file into a perfect postscript one
Before you set up a2ps, you have to set gperf up first. Its installation was a breeze, so I assumed that a2ps would install easily too.....However that was not the case; when reaching the 'make' phase I was getting the following error :
Making all in lib
make all-am
Making all in src
make all-am
/bin/ksh ../libtool --tag=CC --mode=link gcc -g -O2 -o a2ps main.o
read.o sshread.o ssheet.o select.o generate.o delegate.o regex.o buffer.o ve
rsions.o ffaces.o version-etc.o long-options.o parsessh.o lexssh.o lexps.o sh
eets-map.o ../lib/liba2ps.la -lm
gcc -g -O2 -o a2ps main.o read.o sshread.o ssheet.o select.o generate.o delegate
.o regex.o buffer.o versions.o ffaces.o version-etc.o long-options.o parsessh.o
lexssh.o lexps.o sheets-map.o ../lib/.libs/liba2ps.a -lm
Undefined first referenced symbol in file va_copy liba2ps.a(printlen.o)
UX:ld: ERROR: Symbol referencing errors. No output written to a2ps collect2: ld returned 1 exit status *** Error code 1 (bu21) UX:make: ERROR: fatal error. *** Error code 1 (bu21) UX:make: ERROR: fatal error. *** Error code 1 (bu21) UX:make: ERROR: fatal error. *** Error code 1 (bu21) UX:make: ERROR: fatal error.
After a minute of constant....swearing (!!), I found my cool again and set to examine liba2ps.a
liba2ps.a is a collection of libraries, whose dissasembly reveals :
Symbols from ./lib/.libs/liba2ps.a[printlen.o]: [Index] Value Size Type Bind Other Shndx Name [1] |0 |0 |FILE |LOCL |0 |ABS |printlen.c [2] |0 |0 |SECT |LOCL |0 |1 | [3] |0 |0 |SECT |LOCL |0 |3 | [4] |0 |0 |SECT |LOCL |0 |4 | [5] |0 |0 |NOTY |LOCL |0 |1 |gcc2_compiled. [6] |0 |0 |SECT |LOCL |0 |6 | [7] |0 |0 |SECT |LOCL |0 |7 | [8] |0 |0 |SECT |LOCL |0 |9 | [9] |0 |0 |SECT |LOCL |0 |12 | [10] |0 |253 |FUNC |LOCL |0 |1 |int_printflen [11] |0 |0 |SECT |LOCL |0 |14 | [12] |0 |0 |NOTY |LOCL |0 |14 |__FRAME_BEGIN__ [13] |0 |0 |SECT |LOCL |0 |16 | [14] |0 |0 |SECT |LOCL |0 |18 | [15] |0 |0 |SECT |LOCL |0 |5 | [16] |0 |0 |SECT |LOCL |0 |11 | [17] |0 |0 |NOTY |GLOB |0 |UNDEF |strchr [18] |256 |43 |FUNC |GLOB |0 |1 |vprintflen [19] |0 |0 |NOTY |GLOB |0 |UNDEF |va_copy [20] |300 |26 |FUNC |GLOB |0 |1 |printflen
that 'va_copy' is referenced in the printlen.o library and printlen.c file.Examining printlen.c we find a call to it at line 102 :
va_copy (ap, args);
Also since it is a GLOB binding we gather that it is defined elsewhere and not inside printlen.o; as a matter of fact it is not defined in none of the a2ps libraries but belongs to the underlying c library (or gcc library in this case) with its definition appearing in the 'stdarg.h' header :
vi ./usr/gnu/lib/gcc-lib/i586-unknown-sysv5/2.95.3pl1/include/stdarg.h
/* Copy __gnuc_va_list into another variable of this type. */
#define __va_copy(dest, src) (dest) = (src)
#endif /* _STDARG_H */
the issue here is that the gcc version is outdated,hence the gcc library itself, and subsequnetly 'va_copy''s signature does not match the one of the prototype inside printlen.c
Getting a new version of gcc was not an option; I was happy enough to see that there was a gcc instalation after all,let alone installing a new one!
So I was set for a workaround; modify the macro's definition from :
/* Copy __gnuc_va_list into another variable of this type. */
#define __va_copy(dest, src) (dest) = (src)
#endif /* _STDARG_H */
to :
/* Copy __gnuc_va_list into another variable of this type. */
#define va_copy(d,s)
#endif /* _STDARG_H */
This little change satisfied 'make',compiled successfully and a2ps was set up, at last. Now a test was in order to see if it actually works; fortunatelly it did and a2ps transformed a text file into a perfect postscript one
Comments