![]() |
Telelogic DOORS (steve huntington) | ![]() |
new topic :
profile :
search :
help :
dashboard :
calendar :
home
|
||
Latest News:
|
|
Topic Title: Script compatibility when upgrading Topic Summary: Created On: 9-Feb-2006 18:13 Status: Post and Reply |
Linear : Threading : Single : Branch |
![]() |
![]()
|
![]() |
|
I am planning to upgrade from 6.0 SR1 to 7.1.
I've been told to make sure my scripts work properly after upgrading. What sort of things change in DXL that would prevent a script from running? AL |
|
![]() |
|
![]() |
|
[1] There were some issues with constrained DBE placement. If you use a command like:
dbeErrors ->"left" ->"spaced" ->dbeList to position a DBE, it looks like they will need to be tweaked. Constrained placement is flakey. Even after all these years I still use the monkey-at-the-typewriter method of contrained placement, have NEVER got it rght the first time, and I cannot describe the differences. You can see the problem when the Dialog Box is displayed. Then hack at it. [2] If you use function "copyAttribute_" you had to #include <utils/copyops.inc> in v7.0 and below; but you cannot issue that include statement in v7.1 (its already included). [3] CPS behaves differently in v7.1 than it did in v6.0. The v7.1 rules are: // Doors v7.1: When applying, // Deletion CPs: Proposed values are always ignored (as are the Using flags) // Insertion CPs: Proposed values are always Used (ignoring the Using Flags) // Modification Cps: Proposed values are used only when the Using flag is True I don't recall the v6 rules but suspect that the Proposed values were only used when the corresponding Using flag was set, not matter the type of the CP. [4] In v7.0 if you manually delete the text from a Text attribute, it retains its residual, useless, and invisible RichText markup. Thus, string Text = richText(o."Object Text") and then testing for null Text isn't going to work. Be sure you test for an "empty" attribute using the get-raw-text form: Text = o."Object Text". [5] I see a reference in my comments and will investigate if you need to; it had something to do with v7.1 not deleting columns in a view correctly. [6] If you get the rich text from an object and set a different object with it; then retrieve from the second object you'll find it is NOT the same as the originally pasted text. The raw text is the same and the functional rich text is the same (bolding is still bolding), but the way formatting is presented can change: {/b THIS IS BOLD} may look like this: /b THIS IS BOLD/b0. Preamble font declaration formatting is also different. I found some way to get them to use the same formatting, by pasting each text to a Text Dialog Element and then reading each from the Element, THEN comparing them. Yuuuuck. I can think of no good way to determine if Rich Texts are the same. Command richTextFragment didn't help all that much, and is very inefficient in any case. [7] hardDelete(mod) or delete(mod) works differently, but I don't recall how. IIRC, you need to undelete the module before you can permanantly hardDelete it. IIRC, the same issue exists for hardDelete(obj). That's all I see in the comments of my pile of DXL. - Louie |
|
![]() |
Telelogic DOORS
» DXL Exchange
»
Script compatibility when upgrading
|
![]() |
FuseTalk Standard Edition v3.2 - © 1999-2009 FuseTalk Inc. All rights reserved.