| From: | Andrew Gierth <andrew(at)tao11(dot)riddles(dot)org(dot)uk> | 
|---|---|
| To: | coladict(at)gmail(dot)com | 
| Cc: | pgsql-bugs(at)postgresql(dot)org | 
| Subject: | Re: BUG #14599: Segmentation fault and database corruption with hstore | 
| Date: | 2017-03-27 15:39:23 | 
| Message-ID: | 87wpba7n5h.fsf@news-spur.riddles.org.uk | 
| Views: | Whole Thread | Raw Message | Download mbox | Resend email | 
| Thread: | |
| Lists: | pgsql-bugs | 
>>>>> "coladict" == coladict <coladict(at)gmail(dot)com> writes:
 coladict> 3. "CREATE CAST (varchar AS hstore) WITHOUT FUNCTION AS
 coladict> ASSIGNMENT;". If we skip step 4, everything works fine, but I
 coladict> was afraid it might fail with longer strings, so I did that
 coladict> anyway.
What on earth makes you think this is safe or useful? Creating a WITHOUT
FUNCTION cast opens a hole whereby you can feed values to the hstore
functions which are not actually valid hstores, so _of course_ it
crashes.  (There's a good reason why only superusers can create such
casts.)
Once you've stored such a value in a table, then subsequent accesses to
it will of course crash too.
-- 
Andrew (irc:RhodiumToad)
| From | Date | Subject | |
|---|---|---|---|
| Next Message | David G. Johnston | 2017-03-27 15:52:25 | Re: BUG #14599: Segmentation fault and database corruption with hstore | 
| Previous Message | coladict | 2017-03-27 14:32:51 | BUG #14599: Segmentation fault and database corruption with hstore |