https://bugzilla.novell.com/show_bug.cgi?id=867887
https://bugzilla.novell.com/show_bug.cgi?id=867887#c1
Jan Matejek changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|ASSIGNED |NEW
CC| |jmatejek@suse.com
AssignedTo|jmatejek@suse.com |Andreas.Stieger@gmx.de
--- Comment #1 from Jan Matejek 2014-03-12 18:47:46 UTC ---
the failure is in test for the progress_handler feature.
failing pseudocode is roughly this:
sqlite3_progress_handler(connection, 1, _progress_handler);
// call progress handler after every instruction
sqlite3_exec(connection, "create table foo(a,b)");
first_count = how_many_times_progress_handler_was_called();
sqlite3_progress_handler(connection, 2, _progress_handler);
// call progress handler after *every two* instructions
sqlite3_exec(connection, "create table bar(a,b)");
second_count = how_many_times_progress_handler_was_called();
assert(first_count > second_count);
// number of instructions should be smaller than (number of instructions)/2
This fails, because first_count == second_count. IOW, the progress handler gets
called *same number of times* for N=1 and N=2
The docs [1] are somewhat vague about its behavior, but this appears to be
wrong in any case. So my guess is that this is a bug in sqlite.
throwing back at you ;)
if this behavior is intentional, or if you think that this is somehow python's
fault (i didn't try to reproduce in plain C, because i don't know sqlite's C
API), feel free to reassign back to me
[1] http://www.sqlite.org/c3ref/progress_handler.html
--
Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.