MultiValue Tools

I am working on a SB+ screen app where when I bring up a record, information gets displayed, and then when I immediately press F2, this error (in the subject) is displayed followed by:

index: 0 to one dimension matrix is out of range

..at which time SB+ aborts to the colon prompt.

I've checked my code to see if I'm updating some common variable in an unusual fashion but not finding anything like that.  Can anyone with access to the source please tell me which variable is whacked and what I can do about it?

Another oddity, I am pressing F2 while the common variable CNT = 1. However, when inspect ​CNT, it's always 1315 - which is about 22% through the mv's being shown on the screen.  This doesn't happen all the time, but with one specific record I can make it happen EVERY time.  Also, I checked the record for any weird control characters and there are none.  I even removed the mv set of data with CNT = 1315 and CNT is still 1315 when it blows up.  

Any input about what variable is whacked would be most appreciated.

------------------------------
Kevin King
President
Precision Solutions, Inc.
Longmont CO US
------------------------------

Hello Kevin,

This looks like a known issue SBXA-17716 screen driver corruption that was fixed for release 6.5.6.
It occurs when a screen, or a set of linked screens, contains more than 21 mandatory fields - the list of mandatory field steps at the foot of the screen driver record can be corrupted if the step number[2, 999] = 1.
You can work around this by inserting a dummy process to shift the step numbers.
Try adding a dummy process into the 'Process Before' slot of the PD.I process  The process can simply EXIT.

Please go ahead and open a support case for the other oddity.  We need more information to reproduce this issue.

Thanks, Mark.



------------------------------
Mark Shannon
Rocket Internal - All Brands
North Sydney NSW AU
------------------------------

Original Message:
Sent: 10-08-2022 13:49
From: Kevin King
Subject: RHEL, Unidata 8.2.1 hf6, SB+ 5.5.4, _SB554_SB.INTERP.I non-numeric found where numeric required on line 1411

I am working on a SB+ screen app where when I bring up a record, information gets displayed, and then when I immediately press F2, this error (in the subject) is displayed followed by:

index: 0 to one dimension matrix is out of range

..at which time SB+ aborts to the colon prompt.

I've checked my code to see if I'm updating some common variable in an unusual fashion but not finding anything like that.  Can anyone with access to the source please tell me which variable is whacked and what I can do about it?

Another oddity, I am pressing F2 while the common variable CNT = 1. However, when inspect ​CNT, it's always 1315 - which is about 22% through the mv's being shown on the screen.  This doesn't happen all the time, but with one specific record I can make it happen EVERY time.  Also, I checked the record for any weird control characters and there are none.  I even removed the mv set of data with CNT = 1315 and CNT is still 1315 when it blows up.  

Any input about what variable is whacked would be most appreciated.

------------------------------
Kevin King
President
Precision Solutions, Inc.
Longmont CO US
------------------------------

Thank you Mark.  Here are the closing lines to my driver:

027: D 28 9²0²12²21²²19²²63²0²²Y²²²0²²²²ERROR²²²²0²²
028: E 5 1²0²²3
029: M 30 8²9²21
030: P 31 P:EARA0014,ALL
031: W 2

Are you saying that the mandatory step 29 is the issue?  This screen has only these 3 mandatory fields but one of them is step 21.

------------------------------
Kevin King
President
Precision Solutions, Inc.
Longmont CO US
------------------------------

Original Message:
Sent: 10-09-2022 18:47
From: Mark Shannon
Subject: RHEL, Unidata 8.2.1 hf6, SB+ 5.5.4, _SB554_SB.INTERP.I non-numeric found where numeric required on line 1411

Hello Kevin,

This looks like a known issue SBXA-17716 screen driver corruption that was fixed for release 6.5.6.
It occurs when a screen, or a set of linked screens, contains more than 21 mandatory fields - the list of mandatory field steps at the foot of the screen driver record can be corrupted if the step number[2, 999] = 1.
You can work around this by inserting a dummy process to shift the step numbers.
Try adding a dummy process into the 'Process Before' slot of the PD.I process  The process can simply EXIT.

Please go ahead and open a support case for the other oddity.  We need more information to reproduce this issue.

Thanks, Mark.



------------------------------
Mark Shannon
Rocket Internal - All Brands
North Sydney NSW AU

Original Message:
Sent: 10-08-2022 13:49
From: Kevin King
Subject: RHEL, Unidata 8.2.1 hf6, SB+ 5.5.4, _SB554_SB.INTERP.I non-numeric found where numeric required on line 1411

I am working on a SB+ screen app where when I bring up a record, information gets displayed, and then when I immediately press F2, this error (in the subject) is displayed followed by:

index: 0 to one dimension matrix is out of range

..at which time SB+ aborts to the colon prompt.

I've checked my code to see if I'm updating some common variable in an unusual fashion but not finding anything like that.  Can anyone with access to the source please tell me which variable is whacked and what I can do about it?

Another oddity, I am pressing F2 while the common variable CNT = 1. However, when inspect ​CNT, it's always 1315 - which is about 22% through the mv's being shown on the screen.  This doesn't happen all the time, but with one specific record I can make it happen EVERY time.  Also, I checked the record for any weird control characters and there are none.  I even removed the mv set of data with CNT = 1315 and CNT is still 1315 when it blows up.  

Any input about what variable is whacked would be most appreciated.

------------------------------
Kevin King
President
Precision Solutions, Inc.
Longmont CO US
------------------------------

Hello Kevin,

It looks like this does not quite fit with our known issue.

Please go ahead and open a support case to gather more information and dig deeper.

Thanks, Mark.

------------------------------
Mark Shannon
Rocket Internal - All Brands
North Sydney NSW AU
------------------------------

Original Message:
Sent: 10-10-2022 15:07
From: Kevin King
Subject: RHEL, Unidata 8.2.1 hf6, SB+ 5.5.4, _SB554_SB.INTERP.I non-numeric found where numeric required on line 1411

Thank you Mark.  Here are the closing lines to my driver:

027: D 28 9²0²12²21²²19²²63²0²²Y²²²0²²²²ERROR²²²²0²²
028: E 5 1²0²²3
029: M 30 8²9²21
030: P 31 P:EARA0014,ALL
031: W 2

Are you saying that the mandatory step 29 is the issue?  This screen has only these 3 mandatory fields but one of them is step 21.

------------------------------
Kevin King
President
Precision Solutions, Inc.
Longmont CO US

Original Message:
Sent: 10-09-2022 18:47
From: Mark Shannon
Subject: RHEL, Unidata 8.2.1 hf6, SB+ 5.5.4, _SB554_SB.INTERP.I non-numeric found where numeric required on line 1411

Hello Kevin,

This looks like a known issue SBXA-17716 screen driver corruption that was fixed for release 6.5.6.
It occurs when a screen, or a set of linked screens, contains more than 21 mandatory fields - the list of mandatory field steps at the foot of the screen driver record can be corrupted if the step number[2, 999] = 1.
You can work around this by inserting a dummy process to shift the step numbers.
Try adding a dummy process into the 'Process Before' slot of the PD.I process  The process can simply EXIT.

Please go ahead and open a support case for the other oddity.  We need more information to reproduce this issue.

Thanks, Mark.



------------------------------
Mark Shannon
Rocket Internal - All Brands
North Sydney NSW AU

Original Message:
Sent: 10-08-2022 13:49
From: Kevin King
Subject: RHEL, Unidata 8.2.1 hf6, SB+ 5.5.4, _SB554_SB.INTERP.I non-numeric found where numeric required on line 1411

I am working on a SB+ screen app where when I bring up a record, information gets displayed, and then when I immediately press F2, this error (in the subject) is displayed followed by:

index: 0 to one dimension matrix is out of range

..at which time SB+ aborts to the colon prompt.

I've checked my code to see if I'm updating some common variable in an unusual fashion but not finding anything like that.  Can anyone with access to the source please tell me which variable is whacked and what I can do about it?

Another oddity, I am pressing F2 while the common variable CNT = 1. However, when inspect ​CNT, it's always 1315 - which is about 22% through the mv's being shown on the screen.  This doesn't happen all the time, but with one specific record I can make it happen EVERY time.  Also, I checked the record for any weird control characters and there are none.  I even removed the mv set of data with CNT = 1315 and CNT is still 1315 when it blows up.  

Any input about what variable is whacked would be most appreciated.

------------------------------
Kevin King
President
Precision Solutions, Inc.
Longmont CO US
------------------------------