Sibsankar constructs a database with the following schema:

CREATE TABLE t1(a integer primary key, b varchar collate engLish);
CREATE TABLE t2(a integer primary key, b varchar);
CREATE TABLE t3(a integer primary key, b varchar collate engLish);
CREATE TABLE t4(a integer primary key, b varchar collate engLish);
CREATE TABLE t5(a integer primary key, b varchar collate engLish);
CREATE TABLE t6(a integer, b varchar);
CREATE INDEX t1Idx1 on t1(b);
CREATE INDEX t2Idx2 on t2(b collate engLish);
CREATE INDEX t6Idx6 on t6(b collate engLish);

Then in a new SQL connection that does not have a collating
function defined for engLish he does:  "REINDEX t1".  This
gives a segfault.

The correct behavior should be to call the collation-needed
function and if that does not exist or if an appropriate
collating function is not found, it should return an error
of some kind.  
-- 
D. Richard Hipp <[EMAIL PROTECTED]>

Reply via email to