> From: [EMAIL PROTECTED] > [mailto:[EMAIL PROTECTED] On Behalf Of > Ludolf Holzheid > Sent: Tuesday, 20 November, 2007 02:59 > > On Mon, 2007-11-19 06:48:08 -0800, Michael Wojcik wrote: > > > > You're not going to be able to configure Splint to build under DOS. > > The > > FAT32 filesystem doesn't support long file names, for one thing. > > I'm not sure splint depends on long file names.
I'm not sure it does either, but "configure" is nine characters long. Quite a few of the Splint source files have names that aren't 8.3. > > To get a splint binary that works under DOS you'd have to do a > > wholesale port of the technology. I don't think it's feasible. > > If no POSIX environment is available to run the configure > script, the main part would be to generate config.h and > Makefile (from config.h.in and Makefile.in) by hand. Sure. That'll be fun. And emulate M4 to process aclocal.m4, and emulate Flex and Bison to build the lexer and parser, and ... Without a POSIX environment I think building Splint might be something of a chore. > It may be somewhat tricky, but feasible in any case. Manually configuring it, yes. Getting it to run under DOS is a much larger task. Though since it turns out the OP is using DOS 7.1, it might be a bit more plausible than I initially thought. In general, porting non-trivial software designed for a 32-bit virtual-memory OS with a reasonably capable API to the limited monitor that DOS provides is not simple. Back in the day I ported quite a bit of Unix software to DOS, including things like diff and RCS, and it was a chore. Now, that was circa DOS 4 and 5, and DOS 7 is more capable. But this isn't a project I would personally want to devote my resources to. > > Since you have a Windows system available, why don't you copy your > > sources (and any headers they use) to that, and run the Windows splint > > binary against them there? You don't need to run splint on your target > > system. > > But the DOS box seems to be the development system ... So what? You don't need to run Splint on the development system. The software I check with Splint is built on a dozen or so platforms, but I only run Splint on one of them - and I could check it with Splint on a platform that it's not built on, if that were my only choice. Splint doesn't have to be run as part of the build process. It's not necessary to check the code after every single change; it's probably not even desirable. If it were me, I wouldn't even be working on the DOS machine. I'd do my development on Linux or Windows, with DOS running in a VM, and have my development environment use the VM scripting APIs to drive remote compilation. Has no one else here ever done any cross-system development? This isn't a novel suggestion. Hey, if the OP really wants to port Splint to DOS 7, I won't stand in his way. It's certainly likely to be an educational experience, anyway. I'm just suggesting that it may not be the optimal approach. -- Michael Wojcik Principal Software Systems Developer, Micro Focus _______________________________________________ splint-discuss mailing list splint-discuss@mail.cs.virginia.edu http://www.cs.virginia.edu/mailman/listinfo/splint-discuss
