carry@deakin.OZ.AU (Carry Varley) (08/22/90)
1. I recently received the 5.0c upgrade disks from Manx. cc, as and ln identify themselves as 5.0b, the ord utility is still 5.0a, dated Jan 1990, 8:08:05. Using ord on a collection of 83 Vn5.0 object files, it gets as far as printing out the first "**** cycle found, forcing out:" and hangs. This does not occur on smaller test sets of objects, even with multiple "cycles". I am just converting from 3.6a, and in an identical setup the 3.6 ord functions correctly - in fact, I am using it on a 3.6 shadow of the 5.0 file structure to generate a correctly ordered list. Using db, I found ord to be infinite-looping in the region $6d6- $6ec offset from the code hunk origin. It seems to cause no damage, and breaking it out at the bottom allows it to continue correctly. 2. With 5.0b as and ln, the line dc.l ((__H0_end-__H0_org)+(__H1_end-__H1_org)) incorrectly inserts the value of __H1_end, rather than the combined size of hunks 0 and 1. This used to work under 3.6a. A work-around is to equate the values, eg. H0size equ __H0_end-__H0-org H1size equ __H1_end-__H1_org dc.l H0size+H1size I don't know where the problem originates, and what else may fail because of it. ================ I am sorry to take up net bandwidth with this, but I have not been able to reach Manx by phone (no answer), and have received no reply to my email or airmail, sent some weeks previously. Which brings up the point : As the Manx BBS and BIX conferences are not really practical from Australia, is there some way the users here can be kept up to date with news. For example, information could be sent via email, ftp, etc to a collection point here, with access restrictions based on Manx registration if necessary. Should I pass on my registration number to someone in local-call distance from the Manx BBS, in return for email of update information, tips, etc? There MUST be a better way - any ideas ? David Varley C-Born Software Systems Geelong, AUSTRALIA +61 52 786530