Article View: comp.sys.mac.programmer.codewarrior
Article #703Re: [ANN] New Debugger Features for CWDS Mac OS 9
From: Kurt Bigler
Date: Tue, 02 Sep 2003 22:14
Date: Tue, 02 Sep 2003 22:14
138 lines
7510 bytes
7510 bytes
in article mwron-051657.16252302092003@enews.newsguy.com, MW Ron at mwron@metrowerks.com wrote on 9/2/03 1:25 PM: > In article <BB7A2EA3.B5DF%kkb@breathsense.com>, > Kurt Bigler <kkb@breathsense.com> wrote: > >> >> But I see that I was unclear. I thought it was just well understood by >> everyone that the CW8 OS X debugger is relatively useless. > > as compared to what, using GDB. I meant as compared to debugging the same application using the same 8.3 ide under OS 9. >> Viewing global >> variables is not the only problem. Only half of the local variables show, >> and of course this will usually turn out to exclude the ones you need to >> see, and IIRC (I use it so rarely so I forget whether this was fixed) you >> can not set breakpoints in certain modules. > > If you can't see breakpoints or variables it probably is an indication > that your project is too big. And you want to sound like something besides defensive here? > There is a limit to the number of > symbolics we could store and that was not big deal due to the > limitations of TOC. TOC limitations specific to OS X that don't create any problems under OS 9? > However with CW 7.2 people started making larger > and larger applications as they were not forced to write DLL's and such. > therefore we seen a lot more complaints. We have increased the limit to > something like 4 gig from 65k so this shouldn't be a problem. When did you increase this limit? With CW 9? And this was an OSX-specific limit? >> Everything I described has been acknowledged and discussed on this list and >> elsewhere. I was not aware that some people *could* use the debugger, e.g. >> Scott Ribe apparently is able to view his local variables and can set >> breakpoints where he needs to set them. > > In all fairness a good program design would not have this problem but I > understand that customers don't always have this luxury. Well all the bad design is my fault. But I don't have many globals. I have the powerplant globals. Everywhere all tolled in all the modules many of which manage global resources (e.g. MIDI, audio interfaces) and other things that must be instantiated exactly once, I might have 100 globals. I use an object wherever possible and create a single global instantiation of that object rather than making each of its instance variables a global object. Regarding functions, most are quite small. Some get up to 200 lines because they are full of diagnostic things. But the ones that fail to display local variables are often quite small, fitting on the screen easily. So what is the problem? Are you expecting that the "project is too big" problem should NOT have affected me under OS 9 running the same project? Because the problems as I have said is under OS X only. It sounds to me like you are describing generic 8.3 problems, which I apparently don't have. >> So I am waiting for a debugger that works. > > it has always been there, You need to refer back to some previous messages on this newsgroup. But I suppose you are gone. I was not the only one that found OS X debugging rather severely compromised. > and CW 8.3 is very effective, however it does > not work in the same manner as previous versions. Most problems are a > matter of procedural changes. Never had any of those problems. Everything I normally use is basically identical. Besides I am comparing 8.3 under OS X to 8.3 under OS 9. There are no problems under OS 9 debugging the exact same application. So what are these TOC problems? TOC is a PPC thing. OS 9 is no less PPC than OS X. > (Hey they bite me too, and it is > aggravating) or just the differences to Mach-O and how it works. Are there Mach-O specific problems related to what I am describing? >> Thanks, but you could help much more by trying to stop development from >> staying in such a feature-motivated mode when there are so many >> show-stopping bugs. > > You made my day. The worry has been that by putting 8 months into > nothing but quality enhancements and fixing all prioriity one and two > bugs that our customers would not notice the newer features. > > Bless you !!!! Glad to help the cause. But I didn't much in the way of quality enhancements mentioned on the long long list of features. That was my point. That's where you should start if you want to sell. People love features, but they love them 2nd. First we like things to work. We do realize you have a hard job on your hands and, speaking for myself, I am willing to pay for bug fixes, and very unwilling to pay for a single feature, lacking essential bug fixes. >> I just want to know when the debuger will work. I would like to upgrade >> then. I was hoping to hear that I had it all wrong and this had been fixed, >> and then I would happily write my check. Lacking that I would like to hear >> that all other develpment will be stopped until the debugger is working. > > The debugger has added features that allows it to work in a manner you > seem to be desiring. It lets you to have mammoth files and number of > files without losing variables or breakpoints. Please again clarify whether you are talking about fixes compared to the generic 8.3, which means basically that it wouldn't help me, or if you are talking about fixes on the OS X host side specifically, in which case it might help me. > It has a way now that > lets you see global and static variables in the local variables pains > as it used to do with PowerPC programs for many years. That's a great convenience, but merely a convenience. If given the choice between global variables displayed wrong in the local variables pane and global variables whose values actually show correctly in a separate window, I'd chose the latter. I'd like to determine whether it is really worth ordering this thing, and I'd like you to be just a tad clearer so I can understand what fixes are likely to help me. I hope my responses above can help you to further clarify. -Kurt > > Ron
Technical Details
Message-ID:
<BB7A645F.B5FF%kkb@breathsense.com>Path:
arc1-ca.pugleaf.net!archiver1.google.com!news1.google.com!newsfeed.stanford.edu!cyclone.bc.net!in.100proofnews.com!in.100proofnews.com!attla2!ip.att.net!attbi_feed3!attbi_feed4!attbi.com!rwcrnsc51.ops.asp.att.net.POSTED!not-for-mailReferences:
<mwron-8CACDA.14184825082003@enews.newsguy.com> <BB79584E.B5AD%kkb@breathsense.com> <mwron-4158FF.11431902092003@enews.newsguy.com> <BB7A2EA3.B5DF%kkb@breathsense.com> <mwron-051657.16252302092003@enews.newsguy.com>Show Full Headers
Complete Article Headers
X-Google-Language: ENGLISH,ASCII-7-bit X-Google-Thread: 10bcce,39da388965422076 X-Google-Attributes: gid10bcce,public X-Google-Thread: f635e,39da388965422076 X-Google-Attributes: gidf635e,public X-Google-ArrivalTime: 2003-09-02 15:14:41 PST Path: arc1-ca.pugleaf.net!archiver1.google.com!news1.google.com!newsfeed.stanford.edu!cyclone.bc.net!in.100proofnews.com!in.100proofnews.com!attla2!ip.att.net!attbi_feed3!attbi_feed4!attbi.com!rwcrnsc51.ops.asp.att.net.POSTED!not-for-mail User-Agent: Microsoft-Outlook-Express-Macintosh-Edition/5.0.3 Subject: Re: [ANN] New Debugger Features for CWDS Mac OS 9 From: Kurt Bigler <kkb@breathsense.com> Newsgroups: comp.sys.mac.programmer.codewarrior,codewarrior.mac Message-ID: <BB7A645F.B5FF%kkb@breathsense.com> References: <mwron-8CACDA.14184825082003@enews.newsguy.com> <BB79584E.B5AD%kkb@breathsense.com> <mwron-4158FF.11431902092003@enews.newsguy.com> <BB7A2EA3.B5DF%kkb@breathsense.com> <mwron-051657.16252302092003@enews.newsguy.com> Mime-version: 1.0 Content-Type: text/plain; charset="US-ASCII" Content-transfer-encoding: 7bit Lines: 139 NNTP-Posting-Host: 12.240.241.59 X-Complaints-To: abuse@comcast.net X-Trace: rwcrnsc51.ops.asp.att.net 1062540879 12.240.241.59 (Tue, 02 Sep 2003 22:14:39 GMT) NNTP-Posting-Date: Tue, 02 Sep 2003 22:14:39 GMT Organization: Comcast Online Date: Tue, 02 Sep 2003 22:14:40 GMT