提交 0d1aa5ac 编写于 作者: T Tom Lane

pg_dump -z was careless about ownership of indexes.

Now it's a little less so.
上级 f4473440
......@@ -21,7 +21,7 @@
*
*
* IDENTIFICATION
* $Header: /cvsroot/pgsql/src/bin/pg_dump/pg_dump.c,v 1.86 1998/09/23 04:22:14 momjian Exp $
* $Header: /cvsroot/pgsql/src/bin/pg_dump/pg_dump.c,v 1.87 1998/10/01 01:49:12 tgl Exp $
*
* Modifications - 6/10/96 - dave@bensoft.com - version 1.13.dhb
*
......@@ -2678,6 +2678,11 @@ dumpIndices(FILE *fout, IndInfo *indinfo, int numIndices,
if (!tablename || (!strcmp(indinfo[i].indrelname, tablename)))
{
/* We make the index belong to the owner of its table,
* which is not necessarily right but should answer 99% of the
* time. Would have to add owner name to IndInfo to do it right.
*/
becomeUser(fout, tblinfo[tableInd].usename);
strcpy(id1, fmtId(indinfo[i].indexrelname));
strcpy(id2, fmtId(indinfo[i].indrelname));
......
Markdown is supported
0% .
You are about to add 0 people to the discussion. Proceed with caution.
先完成此消息的编辑!
想要评论请 注册