Commit 29240b50 authored by Marko Mäkelä's avatar Marko Mäkelä

Correct a comment about incomplete records

The comment that I made in
commit 06299ddd
is inaccurate. Replace the comment, and make the assertion
debug-only, because I cannot remember any reports of
it ever failing in these 10 years.
parent 78716fff
/*****************************************************************************
Copyright (c) 1997, 2016, Oracle and/or its affiliates. All Rights Reserved.
Copyright (c) 2017, MariaDB Corporation.
Copyright (c) 2017, 2018, MariaDB Corporation.
This program is free software; you can redistribute it and/or modify it under
the terms of the GNU General Public License as published by the Free Software
......@@ -258,12 +258,14 @@ row_vers_impl_x_locked_low(
entry = row_build_index_entry(row, ext, index, heap);
/* entry may be NULL if a record was inserted in place
of a deleted record, and the BLOB pointers of the new
record were not initialized yet. But in that case,
prev_version should be NULL. */
ut_a(entry != NULL);
/* entry could only be NULL (the clustered index
record could contain BLOB pointers that are NULL) if
we were accessing a freshly inserted record before it
was fully inserted. prev_version cannot possibly be
such an incomplete record, because its transaction
would have to be committed in order for later versions
of the record to be able to exist. */
ut_ad(entry);
/* If we get here, we know that the trx_id transaction
modified prev_version. Let us check if prev_version
......
Markdown is supported
0%
or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment