On 5/3/19, Manuel Rigger wrote:
> I just tried the examples on the int-real branch,
> and it seems that they do not cause any errors.
>
The int-real branch has now been fully tested and merged to trunk.
--
D. Richard Hipp
d...@sqlite.org
___
Unfortunately, I deleted some other similar cases that I found before
reading your reply. I'll gather any new cases (I assume you are talking
about cases that the int-real branch should address).
Best,
Manuel
On Fri, May 3, 2019 at 2:34 PM Richard Hipp wrote:
> On 5/3/19, Manuel Rigger wrote:
On 5/3/19, Manuel Rigger wrote:
> I just tried the examples on the int-real branch,
> and it seems that they do not cause any errors.
That is good to hear.
It would be helpful if you could gather together all of your test
cases and send them to us, so that we could add them to the SQLite
Thanks for the feedback! I just tried the examples on the int-real branch,
and it seems that they do not cause any errors.
So I guess these two examples can be considered to be duplicate bug reports
of my "Index on REAL column malfunctions when multiplying with a string"
email.
Best,
Manuel
On
On 5/3/19, Manuel Rigger wrote:
> I found another similar example with GLOB:
>
> CREATE TABLE test (c0 REAL);
> CREATE UNIQUE INDEX index_0 ON test(('0' GLOB -c0));
> INSERT INTO test(c0) VALUES (1.0), (0.0);
> REINDEX;
>
> Is the issue related to the implementation of the unary minus operator
>
I found another similar example with GLOB:
CREATE TABLE test (c0 REAL);
CREATE UNIQUE INDEX index_0 ON test(('0' GLOB -c0));
INSERT INTO test(c0) VALUES (1.0), (0.0);
REINDEX;
Is the issue related to the implementation of the unary minus operator
(i.e., do both examples trigger the same bug)?
Hi,
I found another very specific corner case for which an index fails ("UNIQUE
constraint failed: index 'index_0'"):
CREATE TABLE test (c0 REAL);
CREATE UNIQUE INDEX index_0 ON test(LENGTH(-c0));
INSERT INTO test(c0) VALUES (0.0), ('10:');
REINDEX;
Best,
Manuel
7 matches
Mail list logo
