Re: [sqlite] Assertion Failed In sqlite3

2019-12-27 Thread Richard Hipp
On 12/27/19, Richard Hipp wrote: > On 12/27/19, Richard Hipp wrote: >> >> This is the third such false-positive bug like that this month > > And now there is a fourth: https://sqlite.org/src/info/5fbc159eeb092130c6f2 For those of you still keeping score... A fifth example of this is ticket

Re: [sqlite] Assertion Failed In sqlite3

2019-12-27 Thread Richard Hipp
On 12/27/19, Richard Hipp wrote: > > This is the third such false-positive bug like that this month And now there is a fourth: https://sqlite.org/src/info/5fbc159eeb092130c6f2 CREATE TABLE t0(c0 NOT NULL DEFAULT 1, c1 AS(c0) UNIQUE); REPLACE INTO t0 VALUES(NULL); If you understood my prior

Re: [sqlite] Assertion Failed In sqlite3

2019-12-27 Thread Richard Hipp
On 12/27/19, Manuel Rigger wrote: > Hi Yongheng and Rui, > > This might be a duplicate to a bug that I reported, since both test cases > trigger the same assertion error. See > https://sqlite.org/src/tktview?name=37823501c6. That particular assert() is sort of like an ASAN fault except that it

Re: [sqlite] Assertion Failed In sqlite3

2019-12-27 Thread Bigthing Do
Hi Manuel, Hh, what a coincident. It might be so. The test case looks very different though. Let’s wait for Richard to find it out then. Yongheng & Rui > On Dec 27, 2019, at 2:03 PM, Manuel Rigger wrote: > > Hi Yongheng and Rui, > > This might be a duplicate to a bug that I reported, since

Re: [sqlite] Assertion Failed In sqlite3

2019-12-27 Thread Manuel Rigger
Hi Yongheng and Rui, This might be a duplicate to a bug that I reported, since both test cases trigger the same assertion error. See https://sqlite.org/src/tktview?name=37823501c6. Best, Manuel On Fri, Dec 27, 2019 at 6:09 PM Yongheng Chen wrote: > Hi, > > We found an assertion violation bug

[sqlite] Assertion Failed In sqlite3

2019-12-27 Thread Yongheng Chen
Hi, We found an assertion violation bug in sqlite. Here’s the PoC: — CREATE TABLE v0 ( v1 INTEGER PRIMARY KEY ) ; INSERT INTO v0 VALUES ( 10 ) ; SELECT '29' , count () OVER( ORDER BY v1 ) AS m FROM v0 ORDER BY v1 > ( SELECT m ) ; — The bug exists in the latest development code and release