928786 - migrate legacy "has been marked as a duplicate of this bug" comments to CMT_HAS_DUPE

Status: NEW

(Bugzilla :: Database, enhancement)

Description

[:glob ✱]

there's currently two ways in the database which we track "has been marked as a duplicate" comments:

1. type == CMT_HAS_DUPE
2. comment is '*** Bug \d+ has been marked as a duplicate of this bug. ***'

the second form is from pre-bugzilla 3.0 duplicates (bug 353656).


the problem this causes is code which has to detect duplicates, such as the inline-history extension, have to perform a slow LIKE search on comments instead of just looking at the 'type' column.


until bug 200202 is fixed we should migrate existing "has been marked as a duplicate" comments to CMT_HAS_DUPE, taking care to consider localised installations.

Comment 1

[Frédéric Buclin]

There is no need to take care of localized installations as localized dupe strings are already ignored anyway.

Do you have an idea of the time it would take to scan all comments to detect the dupe strings? (One way to make it faster would be to detect the first comment with type != CMT_NORMAL and only scan older comments).