Wow, I think I can count at least four things wrong with that one comment. Now, I guess that that comment is in the middle of a global data header file, as opposed to being in the planning documentation?

drachenstern:

Wow, I think I can count at least four things wrong with that one comment. Now, I guess that that comment is in the middle of a global data header file, as opposed to being in the planning documentation?

pg_char	 	name[FONT_SIZE + BOM_HEADER];		/* "Name" of font */

Carnildo:

/* Font names are stored as null-terminated UTF-16 Pascal strings. */

So what's your problem?  I'd die for comments like that!  It's a Pascal string -- that means it has a leading length count, probably 2 bytes long (pg_char type in your example).  Then the data also has a null terminator, probably actually two nulls.  This makes it more easily usable from a C context.  Then it tells you that it's in UTF-16 encoding.

I repeat, so what's your problem?

 Why bother null terminating it if you've got a length count?

vt_mruhlin:

 Why bother null terminating it if you've got a length count?

Didn't Joel Sposky call those Fucked Strings once?

They'll get out of sync if the string is at all mutable.

vt_mruhlin:

 Why bother null terminating it if you've got a length count?

Because you can simply pass them to WinAPI.

UpNDown:

So what's your problem?  I'd die for comments like that!  It's a Pascal string -- that means it has a leading length count, probably 2 bytes long (pg_char type in your example).  Then the data also has a null terminator, probably actually two nulls.  This makes it more easily usable from a C context.  Then it tells you that it's in UTF-16 encoding.

I repeat, so what's your problem?

That we still don't know whether it's UTF-16BE or UTF-16LE!

Duhhhhhhhh

Carnildo:

/* Font names are stored as BSTR:s. */