[RFA] Correct problem in s390-opc.tab generation

[Pierre Muller]

  Using 
  s390-opc.tab generation leads to errors
due to longer description fields:

The longest description is 72 char long,
while s390-mkopc.c has a description array of size 60,
without any length limitation.
This leads to format array overwriting and uncompilable 
code...

 This is how I compute the longest description line:
$ cat s390-opc.txt | gawk -F '"' '{print $2}' |wc -L
72

Proposed patch:
1) Increase sdescription size to 80
2) Add a limitation that will generate a failure 
in the sscanf routine rather than creating trash.

  About point 2) I am not sure that adding a length
to that sscanf construct is supported by all scanners...

  Is this OK?


Pierre Muller
Pascal language support maintainer for GDB


opcodes ChangeLog entry:


2010-09-28  Pierre Muller  <muller@ics.u-strasbg.fr>

	* s390-mkopc.c (main): Change description array size to 80.
	Add maximum length of 79 to description parsing.

Index: s390-mkopc.c
===================================================================
RCS file: /cvs/src/src/opcodes/s390-mkopc.c,v
retrieving revision 1.16
diff -u -p -r1.16 s390-mkopc.c
--- s390-mkopc.c        27 Sep 2010 13:36:48 -0000      1.16
+++ s390-mkopc.c        29 Sep 2010 15:30:14 -0000
@@ -335,7 +335,7 @@ main (void)
       char  opcode[16];
       char  mnemonic[16];
       char  format[16];
-      char  description[64];
+      char  description[80];
       char  cpu_string[16];
       char  modes_string[16];
       int   min_cpu;
@@ -345,7 +345,7 @@ main (void)
       if (currentLine[0] == '#')
         continue;
       memset (opcode, 0, 8);
-      if (sscanf (currentLine, "%15s %15s %15s \"%[^\"]\" %15s %15s",
+      if (sscanf (currentLine, "%15s %15s %15s \"%79[^\"]\" %15s %15s",
                  opcode, mnemonic, format, description,
                  cpu_string, modes_string) == 6)
        {