Re: [sqlite] UNIQUE INDEX behaves differently depending on the order rows are inserted

2016-11-02 Thread Richard Hipp
On 11/2/16, Ben Hearsum wrote: > I've also had a couple of kernel panics lately, and I imagine that could > either directly cause issues, or cause issues the next time Firefox starts > (and tries to recover). SQLite is suppose to recover from a kernel panic without

Re: [sqlite] UNIQUE INDEX behaves differently depending on the order rows are inserted

2016-11-02 Thread Ben Hearsum
Just in case it's not obvious, this is a database that's part of my Firefox profile. The database was likely corrupted during an upgrade to Firefox (I've been speaking with the folks that work on Firefox profiles, too), but I've also had a couple of kernel panics lately, and I imagine that could

Re: [sqlite] UNIQUE INDEX behaves differently depending on the order rows are inserted

2016-11-02 Thread Richard Hipp
On 11/2/16, Ben Hearsum wrote: > > I then did some inserts, and discovered that the order in which the rows are > inserted appears to be affecting the uniqueness of the index: > sqlite> insert into moz_places_bhearsum select * from moz_places order by > guid; > sqlite>

Re: [sqlite] VACUUM causes Out of memory error

2016-11-02 Thread Bernardo Sulzbach
On 11/02/2016 05:18 PM, Christian Czech wrote: It is a fundamental bug. I hope one day it gets fixed. Otherwise 3.15.0 is useless. It has been fixed: http://www.sqlite.org/src/info/3028845329c9b7ac -- Bernardo Sulzbach http://www.mafagafogigante.org/ mafagafogiga...@gmail.com

Re: [sqlite] UNIQUE INDEX behaves differently depending on the order rows are inserted

2016-11-02 Thread Richard Hipp
Can you send me the moz_places.sqlite file you are using to my private email? Or get it to me by some other means. So that we can try to reproduce this? On 11/2/16, Ben Hearsum wrote: > Hi there, > > I recently had an issue with an index in a large-ish (>100,000 rows)

[sqlite] UNIQUE INDEX behaves differently depending on the order rows are inserted

2016-11-02 Thread Ben Hearsum
Hi there, I recently had an issue with an index in a large-ish (>100,000 rows) table becoming corrupted. I tried to issue a REINDEX on it, but it complained of duplicated values. I dug into that a bit, and couldn't find any duplicate entries: sqlite> select count(*) from moz_places; 100575

Re: [sqlite] VACUUM causes Out of memory error

2016-11-02 Thread Christian Czech
I already reported this in this forum a while ago but nobody cared. It is only in the 32bit version when RAM usage gets over 2GB on windows. On 64 bit of course this does not come up if u have enough ram. It is only with 3.15.0. It is easy to reproduce. Problem is that intermediate temp file

Re: [sqlite] VACUUM causes Out of memory error

2016-11-02 Thread Richard Hipp
On 11/2/16, Dan Kennedy wrote: > > We can reproduce it here now too. There will likely be a fix soon... > Dan's fix has been checked in and is available in the "Pre-release Snapshot" on the download page. https://www.sqlite.org/download.html -- D. Richard Hipp

Re: [sqlite] VACUUM causes Out of memory error

2016-11-02 Thread Dan Kennedy
On 11/02/2016 09:36 PM, Ralf Junker wrote: I am the author of the DISQLite3 Delphi wrapper. Assurance to all DISQLite3 users: The leak is not DISQLite3 specific. DISQLite3 5.15.0 behaves exactly like the official SQLite 3.15.0 release. The problem can reliably be reproduced using the official

Re: [sqlite] VACUUM causes Out of memory error

2016-11-02 Thread Ralf Junker
I am the author of the DISQLite3 Delphi wrapper. Assurance to all DISQLite3 users: The leak is not DISQLite3 specific. DISQLite3 5.15.0 behaves exactly like the official SQLite 3.15.0 release. The problem can reliably be reproduced using the official sqlite3.exe 3.15.0 binary on Win32.

Re: [sqlite] VACUUM causes Out of memory error

2016-11-02 Thread Donald Griggs
Stephan, I don't have the skills to fix your difficulty, but FWIW, I did reproduce your symptom using the database you provided in case that helps others diagnose it properly. I used: -- SQLite version 3.15.0 -- official command line utility -- Windows 7 -- 16GB installed ram, plenty of hard

Re: [sqlite] Partial Indexes and use of LIKE

2016-11-02 Thread nomad
On Tue Nov 01, 2016 at 10:19:24PM -0400, Richard Hipp wrote: > On 11/1/16, Mark Lawrence wrote: > > Hello all, > > > > The documentation for partial indexes (section 3.2) lists "=, <, >, <=, > > >=, <>, or IN" as operators that will trigger the use of an > > index WHERE c IS NOT

Re: [sqlite] sqlite3_table_column_metadata with nil parameters bug ?

2016-11-02 Thread Richard Hipp
On 11/2/16, LacaK wrote: > Hi *, > > when I pass to sqlite3_table_column_metadata() function 2,3,4 parameters > as null i receive SIGSEGV. > Is this expected behavior ? Yes. The interface requires a non-NULL value for the 3rd parameter. (The 2nd and 4th parameters may be NULL.)

Re: [sqlite] Misspelling of 'quanity' in Row Values documentation section 3.3

2016-11-02 Thread Richard Hipp
On 11/1/16, Omar Qadri wrote: > In the first sentence of section 3.3 of the Row Values documentation ( > https://sqlite.org/rowvalue.html#search_against_multi_column_keys ) the > word "quantity" is spelt as "quanity". Fixed now. Thanks. -- D. Richard Hipp d...@sqlite.org

[sqlite] sqlite3_table_column_metadata with nil parameters bug ?

2016-11-02 Thread LacaK
Hi *, when I pass to sqlite3_table_column_metadata() function 2,3,4 parameters as null i receive SIGSEGV. Is this expected behavior ? I understand that it does not give big sense call this function with all null parameters, but when I use in generic case as 2,3,4 parameters:

[sqlite] Misspelling of 'quanity' in Row Values documentation section 3.3

2016-11-02 Thread Omar Qadri
In the first sentence of section 3.3 of the Row Values documentation ( https://sqlite.org/rowvalue.html#search_against_multi_column_keys ) the word "quantity" is spelt as "quanity". ___ sqlite-users mailing list sqlite-users@mailinglists.sqlite.org

[sqlite] VACUUM causes Out of memory error

2016-11-02 Thread Hennekens, Stephan
Hello, Since last version of sqlite (3.15.0) VACUUM causes an 'Out of memory' error, at least on my bigger databases (> 1 Gb). I came across this error when I installed the latest DISQLite3 version (http://www.yunqa.de/delphi/products/sqlite3/index). I was also able to reproduce the error

Re: [sqlite] undefined symbol: Tcl_TranslateFileName

2016-11-02 Thread nomad
On Wed Nov 02, 2016 at 02:56:01PM +0530, ravi.shan...@cellworksgroup.com wrote: > load a tclsqlite3.5.3.so file it throws an error like this. Probably released on 2007-11-27 - nearly 9 years ago. > http://search.cpan.org/~vkon/Tcl-1.05/Tcl.pm Released on 2016-06-28 - 4 months ago > Perl

[sqlite] undefined symbol: Tcl_TranslateFileName

2016-11-02 Thread ravi.shan...@cellworksgroup.com
Hi guys, I am using a module called Tcl-1.05 in perl. But when i try to load a tclsqlite3.5.3.so file it throws an error like this. Module Name - Tcl-1.05 Link : http://search.cpan.org/~vkon/Tcl-1.05/Tcl.pm Perl verison - 5.8.8 Tcl version - 8.5 The tclsqlite3.5.3.so file exists