RHEL4 / Conquest 1.4.1.3beta and PostgreSQL

I've compiled, mv'd dicom.ini.postgres to dicom.ini and edited to reflect my database name, and I get the following behavior:

kokomk@halliday conquest]% ./dgate -v -o
Attempting to open database; test #1 of 10
Creating test table
Adding a record
Dropping test table
Closing database
Attempting to open database; test #2 of 10
Creating test table
Adding a record
Dropping test table
Closing database
Attempting to open database; test #3 of 10
Creating test table
Adding a record
Dropping test table
Closing database
Attempting to open database; test #4 of 10
Creating test table
Adding a record
Dropping test table
Closing database
Attempting to open database; test #5 of 10
Creating test table
Adding a record
Dropping test table
Closing database
Attempting to open database; test #6 of 10
Creating test table
Adding a record
Dropping test table
Closing database
Attempting to open database; test #7 of 10
Creating test table
Adding a record
Dropping test table
Closing database
Attempting to open database; test #8 of 10
Creating test table
Adding a record
Dropping test table
Closing database
Attempting to open database; test #9 of 10
Creating test table
Adding a record
Dropping test table
Closing database
Attempting to open database; test #10 of 10
Creating test table
Adding a record
Dropping test table
Closing database
---------- Succesful end of test -----------
[kokomk@halliday conquest]% ./dgate -v -i
Initializing Tables

At this point it hangs forever. No tables are created - database is still empty.

What am I missing?! Any insight appreciated.

SO - I've gotten much futher along:

Looking at the postgresql logs, it appears that the creation of the database (dgate -v -i) hangs on when dropping the old tables, so if I create dummy tables in the database and then run 'dgate -v -i' it completes successfully.

NOW - I've been trying to import images into the database and I get an error on insertion into the 'dicomstudies' table. The 'patientbir' column definition is char(8) and my data, which has evidently passed through someones anonymization routine, is trying to pass 'anonymous' into that field in the query. I've tried to alter the schema in 'dicom.sql' to change the column width, but that isn't working - upon recreation I still get a dicomstudies.patientbir field that is char(8). Should I be able to change this? How?

Thanks in advance,
Kris

I'll test it if I can find the older version - can you tell me where I can get it?

Question: have you been testing on PostgreSQL releases prior to 7.3? Apparently behavior changed with 7.3 such that insertion of an oversized string will fail rather than truncate...

I'm running 8.1.5 and insertion of a 9 char string in an 8 car field fails miserably on a test database...

Kris

Hi,
No, I edited the config file in response to an PostgreSQL error indicating that the insert failed as a result of the 9 character date field. (Recall date field contained 'anonymous'. Don't know why the server didn't truncate - is that something that I may have affected in a different configuration file?

For the purpose of demonstration I moved on to the dbaseIII backend, but this wouldn't be acceptable for larger scale use of the product going forward.

I've considered that if I couldn't find any other way around this problem I could always develop a BEFORE INSERT trigger to take care of it...but I really don't want to do this unless I have to.

Where might truncation be invoked (or not) in the server? I'll go back and have another look at the code...