gas/README.coff

   1 The coff patches intend to do the following :
   2
   3    . Generate coff files very compatible with vanilla linker.
   4    . Understands coff debug directives.
   5
   6 Here are the guidelines of the work I have done :
   7
   8    . Encapsulate format dependent code in macros where it is possible.
   9    . Where not possible differenciate with #ifdef
  10    . try not to change the calling conventions of the existing functions.
  11         I made one exception : symbol_new. I would be pleased to hear about
  12         a better solution. (symbols.c)
  13    . Extend the use of N_TYPE_seg seg_N_TYPE tables so that segments can
  14         be manipulated without using their format dependent name. (subsegs.c)
  15    . Write a function to parse the .def debug directives
  16    . Write two small peaces of code to handle the .ln directive.
  17    . In write.c try to move all the cross compilation specifics (md_..) to
  18      format dependent files.
  19    . Encapsulate the data structures using generic types, macros calls.
  20    . Added too much code to resolve the complexity of the symbol table
  21      generated. Most of the code deals with debug stuff.
  22    . Create another makefile, shorter, cleaner.
  23    . Create a config.gas shell script to mimic the gcc,gdb... configuration
  24         mechanism. This reduce the complexity of the makefile.
  25    . Isolate the format dependent code in two files
  26         coff.c coff.h
  27         aout.c aout.h
  28         elf.c elf.h    [ Not yet ;-]
  29    . added a little stack management routine for coff in file stack.c
  30    . isolate os specific flags in m- files
  31
  32 If further development is planed on it is should solve the following problems :
  33
  34    . Encapsulate DESC & OTHER tests in a macro call. I'm not aware
  35         of their exact semantics.
  36    . Clean up the seg_N_TYPE N_TYPE_seg naming scheme
  37    . Try to remove as much reference to segment dependent names as possible
  38    . Find a cleaner solution for symbol_new.
  39    . Report the modifications on vax, ns32k, sparc machine dependent files.
  40         To acheive this goal, search for \<N_, sy_, symbol_new and symbolS.
  41    . Allow an arbitrary number of segments (spare sections .ctor .dtor .bletch)
  42    . Find a way to extend the debug information without breaking sdb
  43      compatibility. Mainly intended for G++.
  44    . should it do something to generate shared libraries objects ?
  45
  46 I have tested this code on the following processor/os. gcc-1.37.1 was
  47    used for all the tests.
  48
  49 386     SCO unix ODT
  50         gcc-1.37.1, gas, emacs-18.55
  51
  52 386     Esix rev C
  53         gas-1.37/write.s
  54
  55 386     Ix 2.02
  56         gas, all the X11R4 mit clients
  57
  58 386     CTIX 3.2
  59         xsol (X11R4 solitary game), gas
  60
  61 68030   unisoft 1.3
  62         the kernel (V.3.2) + tcp/ip extensions
  63         bash-1.05, bison-1.11, compress-4.0, cproto, shar-3.49, diff-1.14,
  64         dist-18.55, flex-2.3, gas-1.37, gcc-1.37.1, gdb-3.6, grep-1.5,
  65         kermit, make-3.58, makedep, patch, printf, makeinfo, g++-1.37.1,
  66         tar-1.08, texi2roff, uuencode, uutraf-1.2, libg++-1.37.2, groff-0.5
  67
  68 68020   sunos 3.5 (no, not coff, just to be sure that I didn't
  69                         introduce errors)
  70         gcc-1.37.1, gas, emacs-18.55, gdb-3.6, bison-1.11, diff-1.14,
  71         make-3.58, tar-1.08
  72
  73 68030   sunos 4.0.3 (idem)
  74         gas
  75
  76 I would be glad to hear about new experiences
  77
  78         Loic  (loic@adesign.uucp or loic@afp.uucp)
  79