Oracle dialect generates too long index names

[Former user]

This issue exists in 3.11

in _execute raise exceptions.SQLError(context.statement, context.parameters, e) sqlalchemy.exceptions.SQLError: (DatabaseError) ORA-00972: identifier is too long 'CREATE INDEX ix_r_5_0_was_r_notify_user_name ON r_5_0_was_r_notify (user_name)' {}

This patch works around the issue by truncating the index to the max_identifier_length() specified by the dialect.

[Former user]

This is a possible (general) patch for too long identifier.

[Mike Bayer]

im assuming the autosetinputsizes change is not part of the change...

[Former user]

Patch agaist SA 4.0 release

[Former user]

Added patch against SA 0.4 (pardon the dyslexia) autosetinputsizes is not really part of the change, however I seem to require it with Oracle 10g with UTF8 (testing on Win32 if that makes a difference).

[Former user]

(original author: ged) Doesn't this problem also exists for other DBMS? (I think I've had it with Postgres) and there is also a similar problem with table names (see ticket 571).

[Mike Bayer]

bd708ad176f936a94551280795c735d59ad32bf4 8c1f08c8aa9f7fe8253323a457b207a260435dde

[Mike Bayer]

Removing milestone: 0.4.xx (automated comment)