On Tue, 26 Feb 2002, Anthony Giorgio wrote:

> I've compiled Splint on an IBM zServer running z/OS R2.  When I
> attempt to use splint on any non-trivial source file, I get the following
> error:
> ANGIO@AQMVSOE ~/unicode> splint unicode.c -I /u/angio/root/usr/include/
> Splint --- 26 Feb 2002
> /u/dbxdev/local//share/splint/lib/lclinit.lci:1,1: invalid C number
> *** Cannot continue.
> I've examined lclint.lci, and as far as I can tell, the file is identical
> to the one on my Linux machine.  What does the error message mean?  I know
> my platform isn't the most common place for UNIX applications to be run,
> but most work fairly well.  One caveat is that the character set the
> machine uses is EBCDIC, as opposed to ASCII.

This is unfortunately the problem.  The LCL initialization parsing code is
ASCII-specific (see lclscanline.c if you need to fix this quickly
yourself, otherwise we will get to it eventually).  If you are not
processing LCL files, you shouldn't need this, though.  The problem is the
-I flag syntax --- it shouldn't have a space.  Use
-I/u/angio/root/usr/include instead.

--- Dave

> The reason that I specify a different location for the source header files
> is that they have "interesting" syntax that splint gags on.  I wrote a perl
> script that fixes the issues.
> I never had this problem with lclint 2.5 - is this a new feature?
> Anthony Giorgio
> dbx Development
> IBM Corporation
> MS/P334
> 2455 South Road
> Poughkeepsie, NY 12601
> Phone: (845) 435-9115
> Tie Line: 295-9115
> "Thou shalt not follow the Null Pointer, for at its end Chaos and Madness
> lie."

Reply via email to