Jeff Dean <jeff@aids-unix> (11/22/84)
I don't have SysV, but: have you tried the "-a" option? On AT&T versions of Unix (SysIII and after), it should suppress error messages about long assignments. In 4.1bsd, the "-a" option is reversed -- you use it to get (not suppress) messages about long assignments. The error message about incorrect sign extension is still there, but apparently someone at Berkeley decided that the message wasn't likely to be useful to anyone, and so it only comes out if you use the "-a" option two (or more) times (this is, of course, undocumented in the manual page). lint definitely needs a way to handle "I know that's a problem but I don't want to hear about it." Explicit casts (when type conversion is the problem) are very good for that purpose; the "lint comments" are also useful. Unfortunately, there are instances where neither suffice. I don't know of anyone who has come up with a better (more general purpose) scheme for doing this. If you feel that adding explicit casts is "cluttering" the code, you're unlikely to find a generally acceptable solution. I'd say that it is advantageous to explicitly mark any code that fails to pass through a program checker (such as lint) to assist future maintainers and others who may have to work with the code. jd
henry@utzoo.UUCP (Henry Spencer) (11/24/84)
> I don't have SysV, but: have you tried the "-a" option? > On AT&T versions of Unix (SysIII and after), it should > suppress error messages about long assignments. > > In 4.1bsd, the "-a" option is reversed -- you use it to get > (not suppress) messages about long assignments. ... Berkeley is not to blame for this difference; that's the way it was in the beginning, in V7. It was AT&T that changed it. -- Henry Spencer @ U of Toronto Zoology {allegra,ihnp4,linus,decvax}!utzoo!henry