making lint output useful

Earl H. Kinmonth ked at garnet.berkeley.edu
Thu Feb 23 03:34:03 AEST 1989


In a recent posting I complained about assbackward parameters in
setvbuf(). SCO called (I'm not under a contract) to explain that the
problem has been fixed in 2.3. So far so good.

One net reader suggested using lint to find such problems. I dutifully
dug out lint although I had tried it before and given up. It didn't
help, but I now remember why I stopped using lint!

My SCO Xenix documentation (Xenix Programmer's Guide, 4-2) says

==========

If lint discovers errors or inconsistencies in a source file, it
produces messages describing the problem. The message has the form:

filename(num):description

==========

This was apparently written by someone with a very perverse sense of
humor. The lint that came with my SCO Xenix produces output that has a
different format for every single message! Not only are the formats
different, but they are perverse. Writing an awk script to convert them
into the documented format for easy insertion in the source files is a
non-trivial task.

QUESTION: Is there a hidden option that causes SCO Xenix lint to
produce output in conformity with the documentation?

QUESTION: Has anyone written a script to massage SCO Xenix lint output
into some halfway usable form?

RHETORICAL QUESTION: Are professional programmers so afraid of
unemployment that they do things in a deliberately half-assed fashion
to guarantee themselves and their peers work fixing up earlier botches?



More information about the Comp.unix.xenix mailing list