Re: [sqlite] apparent column data type affected by numeric operation elsewhere

2014-03-24 Thread mm.w
Hello, attached my build settings/patches if it helps, I added stuff for win Best. On Mon, Mar 24, 2014 at 10:15 AM, mm.w <0xcafef...@gmail.com> wrote: > Hello ok thx, > > I currently use 3.8.3.1 amalgamation, I try to keep on official updates, I > have already mine to deal with > > Best. >

Re: [sqlite] apparent column data type affected by numeric operation elsewhere

2014-03-24 Thread mm.w
Hello ok thx, I currently use 3.8.3.1 amalgamation, I try to keep on official updates, I have already mine to deal with Best. On Mon, Mar 24, 2014 at 10:10 AM, Richard Hipp wrote: > On Mon, Mar 24, 2014 at 1:07 PM, mm.w <0xcafef...@gmail.com> wrote: > > > Hello, > > > > is

Re: [sqlite] apparent column data type affected by numeric operation elsewhere

2014-03-24 Thread Richard Hipp
On Mon, Mar 24, 2014 at 1:07 PM, mm.w <0xcafef...@gmail.com> wrote: > Hello, > > is it fixed in 3.8.4.1 or only in the repo? > is that possible to add the date on the "Download Page" section? > It was fixed earlier this morning on trunk. You can download the latest sources from the website (see

Re: [sqlite] apparent column data type affected by numeric operation elsewhere

2014-03-24 Thread mm.w
Hello, is it fixed in 3.8.4.1 or only in the repo? is that possible to add the date on the "Download Page" section? Best. On Mon, Mar 24, 2014 at 8:35 AM, Richard Hipp wrote: > Also worked correctly in 3.8.2. Problem introduced in 3.8.3. Fixed here: >

Re: [sqlite] apparent column data type affected by numeric operation elsewhere

2014-03-24 Thread Richard Hipp
Also worked correctly in 3.8.2. Problem introduced in 3.8.3. Fixed here: http://www.sqlite.org/src/info/221f8f9447 On Mon, Mar 24, 2014 at 4:52 AM, Clemens Ladisch wrote: > Hi, > > it appears that SQLite thinks it can reuse a register even when > another operation

[sqlite] apparent column data type affected by numeric operation elsewhere

2014-03-24 Thread Clemens Ladisch
Hi, it appears that SQLite thinks it can reuse a register even when another operation changes it by applying numeric affinity: > create table t(x); > insert into t values('1'); > select typeof(x) from t where x=x; text > select typeof(x) from t where x=x+0; integer (first reported here: