hebrew in stored procedure problem

[gad]

Hello,
I use the package body editor to write stored procedure. The procedure has some constatnts in hebrew.
I can write compile and test and all is fine.
But if I close the package body window, and open the same package again, the hebrew constants apear as gibrish '????'. Now I can't compile again as the gibrish is saved to the database as seen on screen. I have to correct all hebrew constants before new compiltaion.
My nls setting in the database are:
NLS_CHARACTERSET AL32UTF8
NLS_NCHAR_CHARACTERSET AL16UTF16

and the local computer I work on registry key (under oracle_home):
nls_lang AMERICAN_AMERICA.IW8ISO8859P8

Data I enter through the grid or through other programs connected to the database apears correct after reading it back from the database.

Please help.

TIA
gad

 

[gad]

I figurred it out. May be it's a bug or somthing wrong with the nls settings on my computer,
but if uncheck the unicode flag under preffrences/oracle/options the package editor reads and displays the hebrew constants OK.
What is that flag for and how is unchecking going to affect plsdev?

gad

 

[Marco Kalter]

Using UTF8 as the client character set will indeed cause this problem, since the editor is not Unicode compliant. This will be added in 7.1.

Disabling this option will implicitly convert your Oracle UTF8 data to the local character set of your Windows PC (presumably Hebrew). This will work fine as long as you do not mix it with other UTF8 languages.