[tex-live] Build fails for FreeBSD 4.10

Vladimir Volovich vvv at vsu.ru
Thu Sep 30 21:09:10 CEST 2004


"ML" == Manfred Lotz writes:

 >> Eddie, for a long-term fix, you could either configure a check for
 >> that type and avoid declaring it if the system does (Manfred is
 >> compiling on FreeBSD), or, more simply, use another name for your
 >> unsigned short type that systems will not define, like my_uint16.
 >> <anything>_t is technically reserved by the !@#$ standards.

 ML> I'm a bit confused. Could it be that I got rubbish when I tried
 ML> rsync -v -a --exclude=Images tug.org::tldevsrc .

t1utils in TeX Live claims to be of version 1.31 (according to file
NEWS), but the files t1asm.c and others differ from the official
distribution's version by hacks added by Fabrice.

E.g. here is the diff of t1asm.c from t1utils-1.31 and from texlive:

--- official/t1asm.c 2004-01-24 05:28:23
+++ tex-live/t1asm.c 2004-03-01 16:05:09
@@ -64,6 +64,7 @@
 #include <lcdf/clp.h>
 #include "t1lib.h"
  
+typedef unsigned short uint16_t;
 #define LINESIZE 512
  
 #ifdef __cplusplus
@@ -346,7 +347,7 @@
 /* This function is used by the binary search, bsearch(), for command names in
    the command table. */
  
-static int command_compare(const void *key, const void *item)
+static int CDECL command_compare(const void *key, const void *item)
 {
   return strcmp((const char *) key, ((const struct command *) item)->name);
 }
@@ -551,7 +552,7 @@
 };
 static const char *program_name;
  
-void
+void CDECL
 fatal_error(const char *message, ...)
 {
   va_list val;
@@ -562,7 +563,7 @@
   exit(1);
 }
  
-void
+void CDECL
 error(const char *message, ...)
 {
   va_list val;

I asked Fabrice long time ago about this, and he said that he needed
these changes (re CDECL) and i may have forgot, but seems that he said
that he tried to contact Eddie on this.

Apparently, Fabrice did not do a very good job in syncing his changes
with official version and forgot to remove the typedef.

This shows that it's a very bad idea to have hacked inofficial
versions in TeX Live. We should ALWAYS send the changes upstream,
otherwise it will become a maintenance nightmare.

Best,
v.



More information about the tex-live mailing list