Compiler Warnings

[SteveFosdick]

Thanks for A09. When compiling on Linux with gcc 9.1.0 I get some warnings, of which the first looks to be of some concern:

a09.c:2361:38: warning: division ‘sizeof (struct regrecord *) / sizeof (struct regrecord)’ does not compute the number of array elements [-Wsizeof-pointer-div]
 2361 | for (i = 0; i < (sizeof(bitregtable) / sizeof(bitregtable[0])); i++)
      |                                      ^
a09.c:1658:19: note: first ‘sizeof’ operand was declared here
 1658 | struct regrecord *bitregtable = bitregtable09;
      |                   ^~~~~~~~~~~
a09.c: In function ‘expandmacro’:
a09.c:6055:5: warning: unused variable ‘nMacLine’ [-Wunused-variable]
 6055 | int nMacLine = 1;                       /* current macro line                */
      |     ^~~~~~~~
a09.c: In function ‘pseudoop’:
a09.c:5956:5: warning: ‘strncpy’ output may be truncated copying 8 bytes from a string of length 32 [-Wstringop-truncation]
 5956 |     strncpy(modulename, namebuf, 8);
      |     ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

[Arakula]

You are right regarding the first warning - that's bad, and I wonder why it hasn't been flagged by my compiler (Microsoft C++ 12 / Visual Studio 2008). It should have been. Obviously, this one crept in when I added the 6800 support. I'll upload a fix as soon as I can find the time. Maybe tonight, but I wouldn't count on it.

If you need it now, the remedy is simple; add a line int bitregtablesize = sizeof(bitregtable09) / sizeof(bitregtable09[0]); after the declaration of bitregtable, then replace for (i = 0; i < (sizeof(bitregtable) / sizeof(bitregtable[0])); i++) with for (i = 0; i < bitregtablesize; i++) ... that's it. A complete implementation (which I'll upload) would also set bitregtablesize each time the bitregtable pointer is modified, but the size is always 3 in the current implementation, so that's not really necessary to get it to work.

[Arakula]

As far as I can see, this should be fixed now. In theory, the other two warnings should be gone, too, but I didn't test that thoroughly (only a quick test with gcc 7.4.0 in the Windows 10 Bash using the command line gcc -Wall -oa09 a09.c which kept its mouth shut).