From: grant@rigel.cc.wmich.edu Sent: Wednesday, October 30, 2002 6:42 AM To: Info-VAX@Mvb.Saic.Com Subject: Re: OpenVMS 7.3-1 and CSWING In article , "John E. Malmberg" writes: > grant@rigel.cc.wmich.edu wrote: >> Just for the heck of it, I updated the copy of the cSWING sources I >> found so it compiles without errors. > > What settings did you use for the Compaq C Compiler? /OPT/NODEBUG Otherwise, defaults. > >> I think I got past the few cases >> of routines with inconsistent numbers of arguments, several cases of >> referencing non-existent structure members, and a library function which >> doesn't exist in the C headers (lib$erase_page). I also cleaned up all >> the IOSBs, corrected some longs which should have been shorts (channels, >> and lengths) and a few improperly initialized descriptors, but, there >> were a few coding errors I don't have time to pursue, including two >> global routines named "rest_cursor" and an attempt to assign a pointer >> to an integer. > > I have slowly been doing the same. > >> However, I don't have the time or ambition to do the kind of rewrite >> which is indicated by Craig Berry's observation. >> >> If anyone would like to either take on that job starting with a >> compilable copy of the source, or just try the compilable version to >> see whether it works any better than the old executable, you can mail >> be and I'll ship you the zip with the original and my changes. > > There seem to be enough latent bugs in the code that it probably will > take a long time and multiple eyes to shake out all but the most > apparent ones. > > Can someone who has a working version let me know if the help function > works correctly? It appears to be using the length of an array of char > as the address to return the amount of characters entered in to the > prompt. The program it self does not seem to use the data, and since I > do not have a link map of the original, I have no idea what memory cell > is being corrupted. > > There are also two arrays that look like they are supposed to be NULL > terminated, but instead are apparently being initialized to the ascii > characters '/' '0'. > > There is also a case where "==" is apparently being used for an > assignment of a return value. > > At the rate I am going over my lunch time, It will probably take a while > before I get all the modules compiled. > > -John > wb8tyw@qsl.network > Personal Opinion Only >