|
Application Error during the import of DMIS program.
<< Domenico Varacalli -- 09/05/06 17:50:01>>
During the import of a DMIS program (Chorus NT, DMIS program) an Application Error appears.
The import of the program correctly works with the preceding version 4.1.
Look at the sequence of operations in the GIF.
<<END>>
<< Paola Pallo -- 09/13/06 18:24:57>>
Skype chat with Bret to clarify this modification.
<<END>>
<< Paola Pallo -- 09/12/06 15:36:05>>
Fixed.
During DMIS import, first of all PC-DMIS saves current PRG, then when Replace option is enabled, it creates a new PRG to be filled with the imported commands.
The issue was that in this phase it used "directly" the Serialize() method : recently, starting form SCHEMA=3051, commands are serialized with additional data ( bytes number and string description) , that are correctly managed only using SerializeCommandOut() and SerializeCommandIn(), defined in CPCDpart_program class.
For review :
V42\SOURCE\ARCHGVAR.CPP
V43B\SOURCE\ARCHGVAR.CPP
<<END>>
<< Changes made by Paola Pallo -- 09/12/06 15:36:41>>
Action: Paola Pallo to David Petrizze, Status: OPEN to REVIEW
<<END>>
<< Paola Pallo -- 09/12/06 13:17:01>>
Going ahead with the investigation, I found that this is caused by a new schema recently added and used in SerializeCommandOut() to protect part program from rogue command serialization bugs.
Fixing in progress.
<<END>>
<< Paola Pallo -- 09/11/06 16:21:08>>
Going ahead with the investigations, making a parallel with v41, that is correctly working.
In spite no differences in the code, for unknown reasons the archive file in v42 is bad .
The exception is in ser-flhd.cpp\SerializeFileHeader(), because strlength is loaded as 1610, instead of a value in-between 0 and 260 : it sounds like a 'dirty' value, since previous store was correct.
Being stuck, I will talk with Dave .
<<END>>
<< Paola Pallo -- 09/08/06 17:58:18>>
I noticed that if you select "Merge" instead of "Replace" option in DMIS Import Setup dialog, the import is working fine. Please use it as a work-around, in the meanwhile I'm looking for the real solution. Thanks.
<<END>>
<< Paola Pallo -- 09/06/06 17:16:00>>
Very strange situation, caused by the serialization of file header command on a new PRG. Investigating
<<END>>
<< Changes made by Tim Wernicke -- 09/06/06 08:47:25>>
Priority: to Critical
<<END>> |
|