dBASE Bugs - Will there ever be help for my random dBASE5.7 errors ?

Author

Will there ever be help for my random dBASE5.7 errors ?

U. Keller

2006-02-10, 7:23 am

I have started a thread "dBASE5.7: Bugs where spaces are not 'seen' as such!" in this newsgroup, the only answer was in fact returning a question to me (what in German is called 'Rueckfrage'); I answered it on Jan29, since then silence in the thread ! Tod
ay I tried searching the archive of this NG, but found nothing that could help. Except may-be in some sense this: someone argues that software bugs are never random, and if they seem random, then there is a hidden pattern that is still unrecognized. OK I
agree, but in many cases (and in mine) this is exactly the *big* problem: how does one find this hidden source ? I have made several hypotheses - all were either false (i.e. did not show the true source of the problem) or need huge work to verify ! So you
must understand that I hope(d) that 'my' bugs were known by professionals of dBASE. May-be I'm wrong with this because if it were true, there would probably be a trace of this knowledge in the said archive. But not necessarily because the bug might have
been handled in other forums / discussions or I may just not have found the good search words to find the relevant thread. I invite to continue the stopped thread at least to tell me that one cannot help with what I communicated. I have made a small test
program containing the core of the code where these illogic and random bugs occur, but this one does not seem to exhibit the bug - it misses the presence of the source of the problem, which may need some 'critical mass' (size of code or number of source f
iles built together or number of dbf's / var's / ...) or the probability of occurrence of such a bug is roughly proportional to such a size parameter (in my appl. it may-be between 1% and 5%) so that only after thousands of starts the bug would happen ...
Here are some of the hypotheses about the bug source which remain undecided: 1) I use some long file names (the small test prog. has one, but this seems insufficient) which can make problems with dbase5.7 (with which I developed the appl. that I deploy o
n a PC without dbase under Windows XP) - in one case I got a reproducible bug with image files addressed from a form, but in the bugs for which I seek a solution, they are 'random' and did never occur in the devel. environ' or even with the exe-file on th
e Win98-PC where I have it. 2) Although the appl. usually work with all its stuff on the WinXP-PC, there may be limited compatibility between this OS and either the use BDE version or the runtime engine associated with dbase5.7 - a bad thing being that in
fact the OS always changes by all these security fixes that many software producers insist one should download & install, in fact I suspect at least one apparently new bug - affecting (also randomly) the running of the tool crystal.vbx - to have been tri
ggered by the install. of such a fix (these things being the only recent change I can accuse). 3) Like the well-known dbase IV bug with (re-)indexing and similar operations, dbase5.7 might contain things that don't work any more on modern fast computers -
the difference being that that old bug was constantly reproducible, but this does not seem to me to be a valid argument against this hyp. - - - Other suggestions to get rid of these bugs? (other than switching to a 32-bit dbase like the dB2K that I have
already tested but badly enough would need much code change work to make the migration work !) Here one of mine: if some people would agree to test my appl's on div. platforms (soft- and hardware) for the ocurrence of 'my' bugs, this could help to at leas
t partly determine which environments make them possible - handling the compatibility issue. The appl's are in the WWW under www.ulyssek.ch (a German language site !)

kathy kolosky

2006-02-10, 9:23 am

Hi U.
The bug-reports Newsgroups are set for bug reporting on the current version
of dBASE.
Unfortunately, 5.7 is an older 16 bit version of the dBASE and is no longer
in development.

You can however obtain programming help with any version in the
dbase.programming Newsgroups.
It would be more appropriate to post your question there.
Also, you may want to make your post more readable for those dedicated
developers who give their time to help.

Thanks, Kathy
QA
dBI

"U. Keller" <ulysse@ulyssek.ch> wrote in message
news:CCoOjsjLGHA.1152@news-server...
>I have started a thread "dBASE5.7: Bugs where spaces are not 'seen' as
>such!" in this newsgroup, the only answer was in fact returning a question
>to me (what in German is called 'Rueckfrage'); I answered it on Jan29,
>since then silence in the thread ! Today I tried searching the archive of
>this NG, but found nothing that could help. Except may-be in some sense
>this: someone argues that software bugs are never random, and if they seem
>random, then there is a hidden pattern that is still unrecognized. OK I
>agree, but in many cases (and in mine) this is exactly the *big* problem:
>how does one find this hidden source ? I have made several hypotheses - all
>were either false (i.e. did not show the true source of the problem) or
>need huge work to verify ! So you must understand that I hope(d) that 'my'
>bugs were known by professionals of dBASE. May-be I'm wrong with this
>because if it were true, there would probably be a trace of this knowledge
>in the said archive. But not necessarily because the bug might have been
>handled in other forums / discussions or I may just not have found the good
>search words to find the relevant thread. I invite to continue the stopped
>thread at least to tell me that one cannot help with what I communicated. I
>have made a small test program containing the core of the code where these
>illogic and random bugs occur, but this one does not seem to exhibit the
>bug - it misses the presence of the source of the problem, which may need
>some 'critical mass' (size of code or number of source files built together
>or number of dbf's / var's / ...) or the probability of occurrence of such
>a bug is roughly proportional to such a size parameter (in my appl. it
>may-be between 1% and 5%) so that only after thousands of starts the bug
>would happen ... Here are some of the hypotheses about the bug source which
>remain undecided: 1) I use some long file names (the small test prog. has
>one, but this seems insufficient) which can make problems with dbase5.7
>(with which I developed the appl. that I deploy on a PC without dbase under
>Windows XP) - in one case I got a reproducible bug with image files
>addressed from a form, but in the bugs for which I seek a solution, they
>are 'random' and did never occur in the devel. environ' or even with the
>exe-file on the Win98-PC where I have it. 2) Although the appl. usually
>work with all its stuff on the WinXP-PC, there may be limited compatibility
>between this OS and either the use BDE version or the runtime engine
>associated with dbase5.7 - a bad thing being that in fact the OS always
>changes by all these security fixes that many software producers insist one
>should download & install, in fact I suspect at least one apparently new
>bug - affecting (also randomly) the running of the tool crystal.vbx - to
>have been triggered by the install. of such a fix (these things being the
>only recent change I can accuse). 3) Like the well-known dbase IV bug with
>(re-)indexing and similar operations, dbase5.7 might contain things that
>don't work any more on modern fast computers - the difference being that
>that old bug was constantly reproducible, but this does not seem to me to
>be a valid argument against this hyp. - - - Other suggestions to get rid of
>these bugs? (other than switching to a 32-bit dbase like the dB2K that I
>have already tested but badly enough would need much code change work to
>make the migration work !) Here one of mine: if some people would agree to
>test my appl's on div. platforms (soft- and hardware) for the ocurrence of
>'my' bugs, this could help to at least partly determine which environments
>make them possible - handling the compatibility issue. The appl's are in
>the WWW under www.ulyssek.ch (a German language site !)

Mark Ziburis

2006-02-10, 11:23 am

In addition, I don't think that 5.7 was ever supported by dBase Inc,
rather was owned by Borland.

Mark Ziburis

kathy kolosky

2006-02-10, 1:23 pm

Hi Mark. In the very beginning dBASE 5.7 was supported by dBASE Inc.
In fact 5.7 was an update to Borland's 5.6 (many bug fixes). But, that was
the last
build of a 16 bit version of dBASE.

Thanks, Kathy
QA
dBI

"Mark Ziburis" <mziburis@ix.netcom.com> wrote in message
news:4aslQqlLGHA.1148@news-server...
> In addition, I don't think that 5.7 was ever supported by dBase Inc,
> rather was owned by Borland.
>
> Mark Ziburis

Sponsored Links