home *** CD-ROM | disk | FTP | other *** search
- Xref: sparky comp.arch:8920 comp.lang.c:12392
- Newsgroups: comp.arch,comp.lang.c
- Path: sparky!uunet!europa.asd.contel.com!darwin.sura.net!mips!apple!news.oc.com!convex!ssimmons
- From: ssimmons@convex.com (Steve Simmons)
- Subject: >>>>>>>>> What would you like in a debugger?
- Sender: usenet@news.eng.convex.com (news access account)
- Message-ID: <ssimmons.713987445@convex.convex.com>
- Date: Sun, 16 Aug 1992 17:50:45 GMT
- Nntp-Posting-Host: pixel.convex.com
- Organization: Engineering, CONVEX Computer Corp., Richardson, Tx., USA
- X-Disclaimer: This message was written by a user at CONVEX Computer
- Corp. The opinions expressed are those of the user and
- not necessarily those of CONVEX.
- Lines: 44
-
- >
- > There are several more important problems that need to be solved
- > first. I'll be happy to define some way for debuggers to talk to
- > each other across machines once I find one or two that do what I
- > want on a single machine. To be absolutely frank, the current
- > state of debugging "technology" is little short of abysmal. I
- > don't use any debugger at the present time on UNIX systems because
- > I haven't found a single good one. I don't care whether the debugger
- > displays linked lists nicely in an X environment. This (without
- > fail) means that the programmers have concentrated on the GUI to
- > the detriment of the basic functionality of the debugger.
- >
- > Seeing as nobody else has written a debugger that I'm happy with,
- > I'm designing one myself. To be frank, I was somewhat disappointed
- > with many of the responses to my original question; people seem to
- > be more worried about the number of widgets they can twiddle than
- > whether their debuggers help them trace problems efficiently.
- >
- > -- Bryan
-
- Bryan:
-
- You are definitely a true developer ... You ask people their opinions...
- None of them agree with yours and so you go out and reinvent the
- wheel. ;-)
-
- I've followed the thread here. Most people have suggested key requirements
- within a debugger. In fact, most of them are mundane and have been
- implemented in one form (or another) in "a" debugger. The problem
- is that no one debugger provides it all.
-
- My advice is the following... Start with the source for one debugger...
- (maybe gdb or dbx). Then, rewrite the code to do you what you think is
- right. I can guarantee two things. Some people will love it and some
- people will hate it. We wrote ours from scratch and it took 6-8 people
- over one year to do it. Personally, we probably made a mistake in
- this tactic although we were limited by GNU's copyright. Also, we
- spent much time dealing with issues of optimized code debugging.
- Good luck from one CXdb developer (Convex's debugger).
-
- Thank you.
-
-
- Steve Simmons
-