RFA gdb/utils.c (gdb_realpath) fallback patch for sco3.2v5

RFA gdb/utils.c (gdb_realpath) fallback patch for sco3.2v5

[rbrown64]

This permits the 0211 snapshot to link & break in main on
i586-pc-sco3.2v5.0.2  (debugging itself).
Possibly this should be done with the configury and set a PATH_MAX
value in config.h using

$ getconf PATH_MAX /
1024

2002-02-20  Rodney Brown  <rbrown64@csc.com.au>

      * utils.c (gdb_realpath): Add pathconf fallback for sco3.2v5.

*** utils.c.orig Sun Feb 10 14:08:42 2002
--- utils.c Wed Feb 20 19:30:16 2002
***************
*** 2522,2527 ****
    char buf[PATH_MAX];
  #elif defined (MAXPATHLEN)
    char buf[MAXPATHLEN];
  #else
  #error "Neither PATH_MAX nor MAXPATHLEN defined"
  #endif
--- 2522,2529 ----
    char buf[PATH_MAX];
  #elif defined (MAXPATHLEN)
    char buf[MAXPATHLEN];
+ #elif defined (HAVE_UNISTD_H) && defined(HAVE_ALLOCA)
+   char *buf = alloca ((size_t)pathconf ("/", _PC_PATH_MAX));
  #else
  #error "Neither PATH_MAX nor MAXPATHLEN defined"
  #endif

Re: RFA gdb/utils.c (gdb_realpath) fallback patch for sco3.2v5

[Andrew Cagney]

> This permits the 0211 snapshot to link & break in main on
> i586-pc-sco3.2v5.0.2  (debugging itself).
> Possibly this should be done with the configury and set a PATH_MAX
> value in config.h using
> 
> $ getconf PATH_MAX /
> 1024

Jason Thorp pointed out similar issues with the code.

We can't define PATH_MAX since that would polute the system name space 
(cf the true/false problem).  However, a namespace proof GDB_PATH_MAX 
might just do the trick.  Getting it via configury could be tricky 
though - it needs to be a compile time, rather than runtime, test.

Hmm, perhaphs autoconf has already solved this problem?

Anyway, yes ok to the change.

Andrew

Re: RFA gdb/utils.c (gdb_realpath) fallback patch for sco3.2v5

[Andrew Cagney]

PS: Feel free to add yourself to the MAINTAINERS file under 
write-after-approval as a separate, obvious fix, patch.

Andrew