Why I abandoned PBP & switched to Swordfish

[Richard Bowser]

The following “discussion” was posted on the PBP forum:

Original “Linker/COFF debugging in MPLAB SIM” post, July 10, 2006:
I am trying to use MPLAB SIM to check out code I’m writing for an 18F8722. I have a brand new installation of MPLAB V7.40, and I can certainly build and execute runnable projects. EXCEPT I cannot enable Linker/COFF debugging, which I need. I run Windows XP Pro here. I have a friend with the same OS & the same MPLAB who does the same thing and can build and simulate with no problems. Yet whenever I try and build my project it gets far enough to report assembly success, I immediately get a Microsoft trap “PBP has encountered a problem and needs to close.” Then the PBP message tells me the .O file is not in valid COFF format. I have easily demonstrated the suite works fine if I do not enable Linker/COFF debugging.

I would like to look a bit closer to narrow down the reason for the problem. It seems the Linker/COFF debugging is unique to PBP. Can anyone offer any insight about what may be going on? It is at least conceivable that I have a problem with some PBP package file.

The first (and only) reply, January 3, 2007:
Hi,

I'm having precisely the same problem, compiling for the PIC18F4431. Did you find a solution?

Thanks!
Jim

My “Linker/COFF debugging” response, January 4, 2007:

No, I was never able to correct the problem. I totally removed PBP from my system and re-did the installation, this time using the installation sources of my friend, who does not experience the problem: But I got the identical failure yet again.

Nobody has been able to shed any light on the issue. Nobody has been able to suggest even a way to capture relevant diagnostic data. The trail seems to have hit a stone wall. I have been forced to abandon PIC Basic, due to this issue. (You are the first person who has ever shown concern about it.) I was really hoping someone would help me figure it out. I do not have enough knowledge of PBP internals to debug this situation.

Frankly, I am quite unhappy with MELabs and the entire user community. It would be different if anybody cared whether the product could be used or not. I am switching to the Swordfish compiler, which is being actively supported by the developer, David John Barker (who also wrote PIC Basic Pro). Swordfish is an advanced complier for Microchip 18XX parts. It has wonderful interrupt support and a truly sophisticated structure. I heartily recommend it.

[TimB]

I'm sorry you left PBP for those reasons I gave it for another "Flat" compiler years ago for other reasons.

I will warn you though that currently no COFF files are produced by SF, its a massive task to produce them, made worse by the fact that its structured compiler and to do the job properly will not be easy.

Lastly just to clarify PBpro was written by Lenard Zimmerman and Geff from Melabs. Dave produced an IDE (Microcode Studio)

The aim is to get it running in ISIS (you can now) with high level debugging.

[dwayner]

Posted: Thu Jan 04, 2007 10:00 am Post subject: Why I abandoned PBP & switched to Swordfish

Original “Linker/COFF debugging in MPLAB SIM” post, July 10, 2006:
I am trying to use MPLAB SIM to check out code I’m writing for an 18F8722. I have a brand new installation of MPLAB V7.40, and I can certainly build and execute runnable projects. EXCEPT I cannot enable Linker/COFF debugging, which I need. I run Windows XP Pro here. I have a friend with the same OS & the same MPLAB who does the same thing and can build and simulate with no problems. Yet whenever I try and build my project it gets far enough to report assembly success, I immediately get a Microsoft trap “PBP has encountered a problem and needs to close.” Then the PBP message tells me the .O file is not in valid COFF format. I have easily demonstrated the suite works fine if I do not enable Linker/COFF debugging.

I would like to look a bit closer to narrow down the reason for the problem. It seems the Linker/COFF debugging is unique to PBP. Can anyone offer any insight about what may be going on? It is at least conceivable that I have a problem with some PBP package file.

<end of quoted message>

I have a suggestion that might be worth checking out: remove MPLAB from the system. Then re-install MPLAB but specify the install path as c:\mp740

The problem most people seem to have is that portions of MPLAB are based on 16 bit (DOS) rules rather than 32 bit. One of those limitations is path length, another is spaces in filenames.

Not saying this WILL fix the problem but should mention that I have not had these problems with any of *my* MPLAB installations.

Also note that MPLAB 7.x seems to be completely intolerant of any remnants of MPLAB 6.x remaining on the system. My most recent MPLAB installation (7.50) required removing all previous 6.xx and 7.xx versions before the ICD2 would function.

However, 7.50 is nice enough to tolerate the continued existance of 5.70.40 - which I still use when going back to projects developed with that version.

dwayne