Volker Nitsch napsal(a): >On 1/19/06, Coussement Christophe <[EMAIL PROTECTED]> wrote: > > >>Hi list, >> >>Our new system, written with REBOL, is running now on some 150 PC's, >>distributed over 13 different locations through Belgium. Those PC's are >>using M$ NT, 2000 or XP and we do not have any control over the >>configuration. >> >>In 90% of the cases runs the software without problems. But in some >>case, never the same, never at the same time or in the same conditions, >>we got an howful "Crash (Should not happend). Corrupt datatype 54 at >>101" -error. >> >>The application is intensively making use of VID. We did optimalize the >>memory use (after the generation of 430 screens, we only use some 2MB >>RAM). We do not make use of 'recycle instructions in the code. >> >> Actually, it is quite unusual to use RECYCLE in your code (I don't do it in my code e.g.), because REBOL uses automatic garbage collection.
>>Can anybody give us some clue about the conditions that cause an "Crash >>(Should not happend). Corrupt datatype 54 at 101" -error and how we do >>have to interpret it ? >> >> >> > >Usually a c-level memory-error. A pointer may be wrong, or some native >writes out of bounds, or the gc freed memory which was still in use. >Very hard to track, happen randomly, some such errors still exists. >One Carl found lately: When text is longer then the area and is >clipped, it writes out of bounds. >Also with async or awake there are sometimes problems. >Don't know if these issues are fixed. > > Yes, these errors are hard to track, did you succeed to reproduce the behaviour, or do you have just reports from your users? I succeeded to isolate a crash of this kind a couple of times too, when I had the code causing the crash. It is very helpful to try to simplify the code as much as possible while retaining the crash to enable Carl to easily debug it. > > >>TIA, >> >>--christophe >> >>This e-mail and any attachments may contain confidential and >>privileged information. If you are not the intended >>recipient, >>please notify the sender immediately by return e-mail, >>delete >>this >>e-mail and destroy any copies. Any dissemination or use of >>this >>information by a person other than the intended recipient is >>unauthorized and may be illegal. >>-- >>To unsubscribe from the list, just send an email to >>lists at rebol.com with unsubscribe as the subject. >> >> >> >> > > >-- >-Volker > >"Any problem in computer science can be solved with another layer of >indirection. But that usually will create another problem." David >Wheeler > > -- To unsubscribe from the list, just send an email to lists at rebol.com with unsubscribe as the subject.
