On Mon, Nov 24, 2008 at 9:30 AM, Michael Schlenker <[EMAIL PROTECTED]> wrote:
> Shane Harrelson schrieb:
> > Sorry for the late reply. I use both the ActiveTCL and the Cygwin
> > distributions for testing on Windows.
> > It should be noted that both versions have
Sorry for the late reply. I use both the ActiveTCL and the Cygwin
distributions for testing on Windows.
It should be noted that both versions have limitations that will cause
"false positives" in the test suite.
Be on the look out especially for problems with 64-integers and floating
point
I'm not certain if there is a publicly available port to the vxWorks
platform, but I know from postings to the mailing list that it has been
done. Was there a particular problem you were encountering that we could
help with? Were you successful at writing a VFS for that target?
-Shane
On
http://www.sqlite.org/cvstrac/timeline should give you and idea of the
number of issues reported against version 3.6.4 and/or which have been
fixed.
HTH.
-Shane
On Tue, Nov 11, 2008 at 4:33 AM, Avinash Mittal <[EMAIL PROTECTED]>wrote:
> Hi,
>
> Can someone tell me how many bugs are present in
http://www.sqlite.org/releaselog/3_6_0.html
-Shane
On Tue, Nov 11, 2008 at 6:10 AM, Mauricio <[EMAIL PROTECTED]> wrote:
> Hi,
>
> I'm having a linking problem and I suspect that the fault goes to the
> version of sqlite3 that comes with my distribution. It's 3.5.9. These
> functions seems
Check the third party processes running on the respective machines. Virus
scanners, file indexing utilities, TortoiseSVN, etc. have been shown to
adversely affect performance on Windows machines by locking the journal file
at inopportune times.
-Shane
On Wed, Nov 5, 2008 at 6:36 AM, Darko
http://www.sqlite.org/omitted.html
Not a "roadmap", but it does give you a hint about the order of potential
new features.
HTH.
-Shane
On Tue, Nov 4, 2008 at 11:06 PM, Avinash Mittal <[EMAIL PROTECTED]>wrote:
> Hi All,
>
> Is there any roadmap made for SQLite?
> I couldn't found any specific
On Thu, Oct 30, 2008 at 1:02 PM, Shane Harrelson
<[EMAIL PROTECTED]> wrote:
> The sizes that I mentioned (315KB vs 205KB) are for the final .dll and
That should read (235KB and 205KB).
___
sqlite-users mailing list
sqlite-users@sqlite
The sizes that I mentioned (315KB vs 205KB) are for the final .dll and
.so size. You might try linking your object files into a lib to see
how that affects size. You could also try running the "strip"
command on the object files to ensure all the debugging symbols are
removed.
Any OMIT options
Are you compiling the SQLite shell utility, or just the library? The
Windows and Linux versions of the lib available from the download page
which are compiled with all the optional features are 235kb and 205kb
respectively. It's surprising that the compiled version for AVR32 is
more than twice
Thanks for the report. The problem has existed in SQLite since version 3.6.1.
This has already been reported in multiple tickets, see tickets #3420 and #3448.
This was caused by check-ins [5450] and [5441] which removed dead code
and exposed the undefined symbols.
This was (hopefully) corrected
I was able to duplicate the issue, and I've updated the configure
scripts to correct the problem.
If you are generating your makefile using configure, you can pass it
options like this:
configure CFLAGS=-DSQLITE_ENABLE_UPDATE_DELETE_LIMIT=1
This will generate a makefile which passes the
(and parse.c, etc.), into a single source file,
sqlite3.c.
You should be able to check your generate sqlite3.c file for a comment
of the form "Begin file parse.h".
HTH.
-Shane
On Wed, Oct 22, 2008 at 8:41 AM, Shane Harrelson <[EMAIL PROTECTED]> wrote:
> What version are you tryi
What version are you trying to build? 3.6.4?
Where did you get your sources? The SQLite website?
Are you building from the amalgamation? or the individual sources?
-Shane
On Wed, Oct 22, 2008 at 8:07 AM, Brent Austin <[EMAIL PROTECTED]> wrote:
> I've been trying to get sqlite3 to build on my
-DSQLITE_ENABLE_UPDATE_DELETE_LIMIT is not being passed in make to
your object compiles...
I'll try to duplicate this here.
-Shane
On Fri, Oct 17, 2008 at 5:40 PM, John Belli <[EMAIL PROTECTED]> wrote:
> OK, I should have all the required utils, and configure apparently
> worked properly, but
You can add it as an option to either configure or make.
For a makefile generated from configure, you would do something like:
make OPTS=-DSQLITE_ENABLE_UPDATE_DELETE_LIMIT=1
TCL is needed to build the amalgamation, as TCl scripts are used to
generate the "amalgamated" file. ActiveTCL works.
Possibly HW (memory, disk, etc.) failures?
On Wed, Oct 15, 2008 at 2:54 AM, Dami Laurent (PJ) <
[EMAIL PROTECTED]> wrote:
> Hi all,
>
> I have a Perl Catalyst app using SQLite as storage. It has been running
> for several months without any problem.
> Now in the last few days we had several
Version 3.6.1 (IIRC) had changes to improve the error detection and retry
logic for this condition (typically caused when a 3rd party application
opens the SQLite journal file.)
HTH.
-Shane
On Mon, Oct 13, 2008 at 12:52 PM, Doug <[EMAIL PROTECTED]> wrote:
> I'm using SQLite 3.5.6 on Windows
Try
http://sqlite.org/releaselog/3_6_4.html
On Wed, Oct 15, 2008 at 8:17 AM, Alexey Pechnikov <[EMAIL PROTECTED]>wrote:
> Hello!
>
> В сообщении от Wednesday 15 October 2008 16:11:47 D. Richard Hipp
> написал(а):
> > For a summary of changes and enhancements
> > that have occurred in version
In the SQLite CVS repository, look at the test subdirectory. There are a
set of TCL based files and scripts which are fed to the SQLite test harness
to implement much of the regression tests for SQLite.
You can probably grep through the *.test files to get examples of many, many
SQL statements
This was my fault. http://www.sqlite.org/cvstrac/chngview?cn=5654
strcasecmp() isn't available on all platforms, and I naively assumed
sqlite3StrICmp() would be (it's not in this case do to the way you
are compiling/linking). I'll review the issue and see what I can do.
On Mon, Sep 22, 2008
There may be one... if not, you could always use .dump from the CLI on both
and then run the text based dumps though a merge utility (like WinMerge) or
through a diff utility. This would probably give you a reasonably good
starting point.
___
Windows XP SP2, SQLite 3.6.1, Intel T2400
(1.83GHZ) Dual Core, 2Gb RAM, 5000RPM Drive
SQL error near line 112: no such table: TEST1
SQL error near line 127: no such table: TEST1
0|performance.txt,v 1.0|1.0|0|0.0K Rows/Second
1|Trivial Inserts|21.17|10077696|476.0K Rows/Second
2|Trivial
This was previously reported, and couldn't be reproduced.
http://www.sqlite.org/cvstrac/tktview?tn=3262
Any help in recreating it would be appreciated.
-Shane
On 8/6/08, Peter Holmes <[EMAIL PROTECTED]> wrote:
>
> FWIW, on a Dell Inspiron 530 running Ubuntu 7.10 with Tcl8.4 installed,
>
>
would appreciate any confirmation that it works for you :)
On 7/30/08, Jeremy Spiegel <[EMAIL PROTECTED]> wrote:
>
> Shane,
>
> Just saw the checkin for the fix. Thanks!
>
> :) Jeremy
>
> -Original Message-
> From: [EMAIL PROTECTED]
> [mailto:
GetFileAttributes() returns INVALID_FILE_ATTRIBUTES with an error of
ERROR_ACCESS_DENIED if the file is in a "pending delete" state. I'll update
the retry logic in os_win.c in winDelete() to add this additional check. I
believe that should improve the situation (as long as the other application
If you're really serious about storing your data in a "native" form to
reduce read/write overhead for handling endian-ness (and I think this is
going to be more trouble than it's worth) then you can look at replacing the
following functions in vdbeaux.c:
** sqlite3VdbeSerialType()
**
on the suggested lemon changes (--help support and usage
updates) -- perhaps open a ticket and see what DRH says.
-Shane
On 7/23/08, Dennis Cote <[EMAIL PROTECTED]> wrote:
>
> Shane Harrelson wrote:
> > I checked in some updates to the "configure" support that will hopefully
>
On 7/22/08, Steve Friedman <[EMAIL PROTECTED]> wrote:
>
>
>
> Actually, I have both --enable-tempstore=always and -DTEMP_STORE=2
> (belts and suspenders).
Then you probably want to use -DSQLITE_TEMP_STORE=2 instead.
___
sqlite-users mailing list
I checked in some updates to the "configure" support that will hopefully do
the right thing and pass any OMIT options to lemon and mkkeywordhash.
There was also a minor fix to handle SQLITE_OMIT_VIEW being defined while
SQLITE_OMIT_SUBQUERY is undefined in select.c (something you probably rand
Without creating a .DEF file for MSVC to use, you need to tell it which
functions to "export".
The easiest way to do this is with the __declspec(dllexport).
You should modify your source file and add the following before each public
function:
__declspec(dllexport)
So for instance:
int
I'm not a very knowledgeable on Debian distributions, but
http://packages.qa.debian.org/g/gcc-4.3.html
indicates that gcc (Debian 4.3.1-4) which you posted that you are using was
only made available 2008-07-01,
is marked as "unstable", and hasn't even made it to their testing stage
yet. Is there
Try
make CFLAGS="-g -O2 -fno-fast-math"
and see if that fails like your test case 2. I *think* this will turn on
the -O2 optimizations and disable fast-math.
Other than that, I don't have any other suggestions.
HTH.
-Shane
On 7/8/08, Tom Epperly <[EMAIL PROTECTED]> wrote:
>
You can find the same issue reported for Fedora from a few weeks ago:
http://www.sqlite.org/cvstrac/tktview?tn=3186
Here's the original thread from the mailing list discussion:
http://sqlite.org:8080/cgi-bin/mailman/private/sqlite-dev/2008-June/000172.html
-ffast_math was the culprit in this
Make sure SQLite isn't being compiled with -ffast_math on the the Debian
side. That might cause problems.
On 7/3/08, Tom Epperly <[EMAIL PROTECTED]> wrote:
>
> When I tried reproducing this behavior on a Red Hat box, the 3.5.9
> version gave the expected results, so I guess it is a Debian
I was able to reproduce this by setting by TZ to GMT +10:00. It's a
floating point rounding issue in the julian date functions. We're
investigating how to best correct it, but I don't have a "fix" for you now.
On 6/12/08, BareFeet <[EMAIL PROTECTED]> wrote:
>
> Hi Shane,
>
> >> This:
Tom-
Can you provide some details of your test setup? What version of SQLite?
What platform (compiler, O/S, processor, 32bit vs 64bit, etc.)?
I updated the date testscripts in CVS to add tests for you cases below, and
they worked correctly for version 3.5.9 of SQLite compiled with both GCC and
On 6/11/08, Christophe Leske <[EMAIL PROTECTED]> wrote:
>
> Shane Harrelson schrieb:
> > Were you able to try this Christophe? I expect it to only be a very
> slight
> > performance improvement, but I'm still curious as to how much.
> >
> Shane,
>
> tha
Were you able to try this Christophe? I expect it to only be a very slight
performance improvement, but I'm still curious as to how much.
On 6/9/08, Shane Harrelson <[EMAIL PROTECTED]> wrote:
>
> Each pair of columns (min/max) represents one dimension. So for
> latitude/longit
If you just need the rtree extension, I believe it will be included by
default in the next version of the core sqlite library.
-Shane
On Tue, Jun 10, 2008 at 12:39 AM, Stephen Woodbridge <
[EMAIL PROTECTED]> wrote:
> Hi all,
>
> I was reading through the sqlite source and noticed that there
Each pair of columns (min/max) represents one dimension. So for
latitude/longitude coordinates you would have 5 columns: 1 id column and 2
columns each for latitude and longitude. 5 "columns" equates to 2
dimensions of data. 7 "columns" equates to 3 dimensions of data. Etc.
HTH.
-Shane
On
You can improve performance (space/speed) a little by changing the
RTREE_MAX_DIMENSIONS at the top of rtree.c to match you data set. It
defaults to 5 dimensions, but you could reduce to this to 2, or 3 with city
size.
#define RTREE_MAX_DIMENSIONS 5
-Shane
On 6/9/08, Dennis Cote <[EMAIL
With the recent discussions on R-Tree's and spherical coordinates (i.e.
latitude and longitude), I was wondering how boundary conditions would be
handled.
If I choose a location on the equator as far from the prime meridian as
possible, and try to query for all the locations "near" that,
within a
In addition to playing with the splitting algorithms, since you are
compiling your own DLL, you can customize the R-Tree module for 2 dimensions
-- i.e. hard-code it for 5 columns of data.
This would *roughly* be accomplished by replacing all occurrences of
pRtree->nDim
with
(2)
and
The default R-Tree code uses the "R*-tree algorithm" for splitting. While
this should typically perform better than the other two variants (Guttman
Quadratic and Linear splitting), you may want to test them with your data
set and queries to see if either performs better. You will need to
Dennis-
Your last "simplification":
> -- a further simplification of the general case that removes
> -- redundant terms
> select * from City
> where id in
> (
> select id from CityLoc
> where (lat_min < :max_lat and lat_max > :min_lat)
> and
Once you get it working with your data, you may want to play around with the
defines at the top of rtree.c.
> /* Either, both or none of the following may be set to activate
> ** r*tree variant algorithms.
> */
> #define VARIANT_RSTARTREE_CHOOSESUBTREE 0
> #define VARIANT_RSTARTREE_REINSERT 1
>
You can read a very good overview of R-Trees at Wikipedia.
http://en.wikipedia.org/wiki/R-tree which includes some coordinate based
examples.
On 6/4/08, Jay A. Kreibich <[EMAIL PROTECTED]> wrote:
>
> On Wed, Jun 04, 2008 at 06:18:22PM +0200, Christophe Leske scratched on the
> wall:
> >
> > >
Without creating a .DEF file for MSVC to use, you need to tell it which
functions to "export".
The easiest way to do this is with the __declspec(dllexport).
You should modify the rtree.c source file and add the following before each
public function:
__declspec(dllexport)
So for instance, line
This wiki page (http://www.sqlite.org/cvstrac/wiki?p=LoadableExtensions)
talks about
SQLite's loadable extension functionality.
If you can tell me what platform you're compiling for (processor, O/S
version, etc.), and what build tools
(cygwin/gcc, mingw, MSVC, etc.) you're using, I will try and
The format of the varints in the DB were not changed.For Varint32s, the
macro support was cleaned up and changes made to use it consistently. The
MACROs inline the single byte case, so code for this was disabled in the
actual function. The actual functions were re-implemented to utilize
This may be fixed by ticket 3030:
http://www.sqlite.org/cvstrac/tktview?tn=3030
On 5/6/08, Hans Guijt <[EMAIL PROTECTED]> wrote:
>
> Hi all,
>
>
> I have a simple question: on my SQLite database I set
> sqlite3_busy_timeout() to a generous 1 milliseconds, but sometimes
> it doesn't wait _at
If you have TCL installed, you can build the 'testfixture' which is used to
run the tests.
;# Unpack the source tree into "sqlite"
mkdir sqlite
cd sqlite
tar xzf sqlite.tar.gz
;# Build will occur in a sibling directory
cd ..
mkdir bld
;# Change to the build directory
cd bld
;# Run the
Is the limit of 64 tables in a join changed now that it's using "bit
vectors" instead of "bit maps"? Similar on number of attached databases?
Or were the changes only to the bitmaps used for page tracking?
http://www.sqlite.org/limits.html
-Shane
___
Embedded Database Vendors Face Challenges:
http://www.eweek.com/article2/0,1895,2243406,00.asp
How could an article like this not mention SQLite?
Now thats a good idea! Thanks!
On 8/15/07, [EMAIL PROTECTED] <[EMAIL PROTECTED]> wrote:
>
> "Shane Harrelson" <[EMAIL PROTECTED]> wrote:
> >
> > Other than the normal caveats for using customized versions of the
> SQLite
> > code, does this sou
On 8/15/07, [EMAIL PROTECTED] <[EMAIL PROTECTED]> wrote:
>
> "Shane Harrelson" <[EMAIL PROTECTED]> wrote:
> > I have a table that has the rows sorted by using a column which is
> filled
> > with values from random(*).
> >
> > Is there
I have a table that has the rows sorted by using a column which is filled
with values from random(*).
Is there a PRAGMA or other API which allows me to set the "seed" used by
random(*) such that
I can reproduce the same random sequence each time?
Thanks.
-Shane
I have a query of the following form using SQLite's built in GLOB function:
SELECT * FROM foo WHERE bar GLOB '*'
I implemented a custom glob(x,y) function to replace the built-in GLOB,
registering it with:
sqlite3_create_function(db, "glob", 2, SQLITE_ANY, NULL, my_glob, NULL,
NULL);
This all
To use pragmas from code, do I simply prepare them as a regular SQL
statement and then execute them?
And when can they/should they be done? As the first statement after an
open?
Are the pragma values stored with the database? Or do they have to be
issued after each open?
Thanks.
-Shane
Good stuff! Is this in the wiki somewhere?
On 6/12/07, [EMAIL PROTECTED] <[EMAIL PROTECTED]> wrote:
"Adler, Eliedaat" <[EMAIL PROTECTED]> wrote:
> Hi,
>
> I want to add an index on a sparsely populated column -
> i.e. for most rows the column is null.
>
> 1. Would this index be very compact?
It appears that I can *NOT* create a custom collation function for a column
declared as an INTEGER. Well, I can create it, and prepare a statement to
use it, but my custom function is *NEVER* called. Is this by design? The
exact same code works for a TEXT column.
Should the
On 2/6/07, Joe Wilson <[EMAIL PROTECTED]> wrote:
--- Shane Harrelson <[EMAIL PROTECTED]> wrote:
> On 2/5/07, Joe Wilson <[EMAIL PROTECTED]> wrote:
> > > Yes, it's typical. Each database instance is typically composed of
> > > around 50k records, all in
On 2/5/07, Joe Wilson <[EMAIL PROTECTED]> wrote:
--- Shane Harrelson <[EMAIL PROTECTED]> wrote:
> On 2/5/07, Dennis Cote <[EMAIL PROTECTED]> wrote:
> > Shane Harrelson wrote:
> > > Perform 50 transactions of 1000 inserts each (5 total).
> > &g
On 2/5/07, Dennis Cote <[EMAIL PROTECTED]> wrote:
Shane Harrelson wrote:
> Perform 50 transactions of 1000 inserts each (5 total).
>
Shane,
Is this your normal usage pattern? Inserting records in blocks of around
1000 per transaction. Or would you be more likely to inser
On 2/5/07, A. Pagaltzis <[EMAIL PROTECTED]> wrote:
* [EMAIL PROTECTED] <[EMAIL PROTECTED]> [2007-02-04 13:35]:
> "A. Pagaltzis" <[EMAIL PROTECTED]> wrote:
> > It's a pity that INSERT OR IGNORE (apparently?) does not set
> > last_insert_id properly regardless of outcome,
>
> Consider this case:
>
On 2/2/07, Dennis Cote <[EMAIL PROTECTED]> wrote:
Shane Harrelson wrote:
> On 2/2/07, A. Pagaltzis <[EMAIL PROTECTED]> wrote:
>> * Nicolas Williams <[EMAIL PROTECTED]> [2007-02-01 00:10]:
>> > Sounds like you should want to use INSERT OR IGNORE ... INTO
&g
On 2/2/07, A. Pagaltzis <[EMAIL PROTECTED]> wrote:
* Nicolas Williams <[EMAIL PROTECTED]> [2007-02-01 00:10]:
> Sounds like you should want to use INSERT OR IGNORE ... INTO
> Strings and then SELECT the rowid of the string for use in
> INSERTing INTO Object.
That was my first thought as well.
"Shane Harrelson" <[EMAIL PROTECTED]> wrote:
>
> I have two tables, an "Objects" table with a foreign key into a second
> "Strings" table which is composed of unique values. It is a many to
> one relationship, that is, several Objects may refere
On 1/31/07, Dennis Cote <[EMAIL PROTECTED]> wrote:
Shane Harrelson wrote:
> On 1/31/07, [EMAIL PROTECTED] <[EMAIL PROTECTED]> wrote:
>>
>> The official way to find the conflicting entry is to do a query.
>>
>> SELECT rowid FROM table WHERE uniquecolumn=?
On 1/31/07, [EMAIL PROTECTED] <[EMAIL PROTECTED]> wrote:
"Shane Harrelson" <[EMAIL PROTECTED]> wrote:
> when i try to insert a row into a table that has a UNIQUE constraint
> on a column, and I get the SQLITE_CONSTRAINT result code because i'm
> inserting a d
when i try to insert a row into a table that has a UNIQUE constraint
on a column, and I get the SQLITE_CONSTRAINT result code because i'm
inserting a duplicate value, is there anyway to determine the rowid of
the conflict?
looking at the internals of the VDBE, i found that the rowid of the
when i try to insert a row into a table that has a UNIQUE constraint
on a column, and I get the SQLITE_CONSTRAINT result code because i'm
inserting a duplicate value, is there anyway to determine the rowid of
the conflict?
looking at the internals of the VDBE, i found that the rowid of the
On 1/30/07, [EMAIL PROTECTED] <[EMAIL PROTECTED]> wrote:
"Shane Harrelson" <[EMAIL PROTECTED]> wrote:
> what's the most reliable method for checking the size of the jounal
> file for a database? is there anyway to flush that or sync it during
> testing? i tried u
what's the most reliable method for checking the size of the jounal
file for a database? is there anyway to flush that or sync it during
testing? i tried using stat() on the jounral file but under windows
it sometimes reports the jounral file having size 0 even though I know
it's not empty.
101 - 175 of 175 matches
Mail list logo