|
Get
The Update
VB/ISAM
Service Pack 01/10/2005 Version 50.32.5.9 The
following fixes are cumulative and include all fixes
from previous Service Packs
|
|
Version
|
Issue
|
Fix
|
|
9.
|
50.32.6.2
05/02/08
|
Certain Sharing access will result in missing keys during a read with nodata at the end of file (dataset). The Data is intact but the software will not find all keys during a readnext operation.
|
Modified the access to the Map validation routine to determine if a new copy is required before the readnext operation.
|
|
8.
|
50.32.5.9
01/10/05
|
DataSet
Formats with arrays of 200 or more strings (a250$) would
cause the system to crash when the dataset was
opened.
|
While
fixing an error causing Compound Fields to be
processed, we created an array of 256 objects
that turned out could become larger under the
circumstances when a format string specified
"A250$". We now use an array that
is much larger and cannot overflow!
|
|
7.
|
50.32.5.8
10/22/04
|
When
a dataset grows to its maximum, the
last record update will corrupt the
database requiring a rebuild to recover.
|
Modified
the low level binary search algorthm
to account for the specific end point
contition.
A
new download is available on the Products
page. Please send Email to receive a
newer version of Vtl5032.dll
|
|
6.
|
50.32.5.6
09/16/04
|
Certain
Endpoint cases of multiple users updating
different records in the last group
could result in erroneous reporting
of VIS_DISK_ERROR.
|
Corrected
the test for dataset full so that a
record update will not be attempted
if there is not enough space for the
worst case condition. VB/ISAM will now
correctly return VIS_NO_ROOM!
|
|
5.
|
50.32.5.5
07/18/04
|
A
bug in the C++ library "realloc" function
caused record growth to overlay other memory when
modifying field data.
|
Created
an internal bug-free version for VB/ISAM.
|
|
4.
|
50.32.5.5
07/18/04
|
A
bug caused the low level Dataset Open function to
clobber certain parts of memory when one of the
component dataset files (.ism) was missing.
|
Re-coded
the low level function to correct this issue.
|
|
3.
|
50.32.5.5
07/18/04
|
VIS_DISK_ERROR
now returns an extended error "14xx" where
xx identifies the specific failure. This
can be a value from 01 to 50. These errors occure
primarly because a block of data that we expected
to be a certain size was either larger or (mostly)
smaller than we expected.
|
This
helps to pinpoint system failures for technical support.
|
|
2.
|
50.32.5.5
07/18/04
|
A bug
related to small primary key sizes and VmxGet/XLOOKUP
would incorrectly return VIS_INVALID_KEY
|
Modified
VmxGet to interpret the stored format string properly
for this case.
|
|
1.
|
50.32.5.5
07/18/04
|
A bug
in VmxOpen & VmxCreate would not process path/filename
beyond 128 characters.
|
Changed
the create and open modules to respond properly
by calling a more appropriate windowAPI to determine
Full_Path!
|
|