-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Fred Williams wrote:
> I say there is no known translation that
> would allow the three SQLite, "Small, Fast, Reliable" adjectives to
> translate into any regurgitated language output, with the exception of
> compiling SQLite source with a C++
Having had the unfortunate opportunity to use a couple of language
translators as well as spending about six fruitless months developing one
which in the end was no better, I say there is no known translation that
would allow the three SQLite, "Small, Fast, Reliable" adjectives to
translate into
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Nikolaus Rath wrote:
> Is it normal that UPDATE still takes more than 11 times as much time as
> SELECT, or should I be able to get better performance?
The UPDATE time also includes the time waiting to acquire file locks. If
you have concurrent
"Igor Tandetnik" writes:
> Nikolaus Rath wrote:
>> When running apswtrace on my Python program, I got the following
>> output:
>>
>> LONGEST RUNNING - AGGREGATE
>>
>> 16638 11.041 UPDATE inodes SET size=MAX(size,?), ctime=?, mtime=?
>> WHERE id=? 16638 0.938 SELECT s3key
Nikolaus Rath wrote:
> When running apswtrace on my Python program, I got the following
> output:
>
> LONGEST RUNNING - AGGREGATE
>
> 16638 11.041 UPDATE inodes SET size=MAX(size,?), ctime=?, mtime=?
> WHERE id=? 16638 0.938 SELECT s3key FROM inode_s3key WHERE inode=?
> AND offset=?
>
> i.e.
Hi,
When running apswtrace on my Python program, I got the following output:
LONGEST RUNNING - AGGREGATE
16638 11.041 UPDATE inodes SET size=MAX(size,?), ctime=?, mtime=? WHERE id=?
16638 0.938 SELECT s3key FROM inode_s3key WHERE inode=? AND offset=?
i.e. both statements were executed
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Alexander Kitaev wrote:
> Not to depend on native SQLite binaries or
> opaque NestedVM code,
As a matter of interest what problem exactly do you have with NestedVM?
It's output is indeed opaque (not human comprehensible) but the same is true
of Java
Hi,
There has been a bit more discussion about this on
http://code.google.com/p/apsw/issues/detail?id=59. I guesss Roger would
be quite interested in the test case you mentioned. I have now decided
to simply deactivate shared cache again. It seems the benefits aren't
worth it in my case.
As a
Hello All,
I'd like to invite those users who program in Java to take a look at our
new project - SQLJet.
SQLJet was started as part of another projects that has to work with
SQLite database (SVNKit). Not to depend on native SQLite binaries or
opaque NestedVM code, we've implemented part of
This isn't directly related to SQLite, but rather to how I am converting the
source code into a DLL. I hope someone here can help me understand what's
happening.
It appears that my previous problem was caused by some problem within the
ancient, unbelievably slow and ugly library we have been
D. Richard Hipp , On 8/11/2009 16:39:
> program is using LinuxThreads or NPTL for its threading. (SQLite has
> to know which is used because there are serious bugs in LinuxThreads
> that SQLite has to work around.) So pthreads gets used once, by
> SQLite, even if you don't do any threading in
On Aug 11, 2009, at 4:28 PM, Mark Richards wrote:
> D. Richard Hipp , On 8/11/2009 16:02:
>>
>> Perhaps pthreads is going goofy. Please recompile with -
>> DSQLITE_THREADSAFE=0 and see if that helps.
>>
> BINGO.
>
> env CC="gcc-cris -mlinux -isystem $EROOT/include"CPP="gcc-cris
> -mlinux
D. Richard Hipp , On 8/11/2009 16:02:
>
> Perhaps pthreads is going goofy. Please recompile with -
> DSQLITE_THREADSAFE=0 and see if that helps.
>
BINGO.
env CC="gcc-cris -mlinux -isystem $EROOT/include"CPP="gcc-cris
-mlinux -E -isystem $EROOT/include" CXX="g++-cris -mlinux -xc++
On Aug 11, 2009, at 2:53 PM, Mark Richards wrote:
>
> When sqlite3_open() is called as above, got returns -1 (forever). -1
> returning from a NONBLOCK read is accepted behaviour, but pretty
> quickly
> the read() should give some data. Instead it returns -1 each time.
> (If
> I allow fd[0]
Yes, tried that. No change in result.
Assuming sqlite does a |F_GETFD| then restores when done.
I built in some testing to see if the file descriptor was being mangled,
and cannot detect any difference.
Test:
flags = fcntl(fd[0], F_GETFL, 0);
char *cFlagMsg;
On Aug 11, 2009, at 2:53 PM, Mark Richards wrote:
> Environment:
> Linux axis 2.6.19 #9 PREEMPT Mon Apr 6 15:44:03 EDT 2009 cris
> unknown
>
> Sqlite:
> Sqlite: sqlite-3.6.14
> ./configure --host=cris-axis-linux-gnu
> --prefix=/AEMDEV/83+/devboard-R2_10/target/cris-axis-linux-gnu
>
I also (just) tried 3.6.17. Same issue.
/m
Mark Richards , On 8/11/2009 14:53:
> Environment:
> Linux axis 2.6.19 #9 PREEMPT Mon Apr 6 15:44:03 EDT 2009 cris unknown
>
> Sqlite:
> Sqlite: sqlite-3.6.14
> ./configure --host=cris-axis-linux-gnu
>
Environment:
Linux axis 2.6.19 #9 PREEMPT Mon Apr 6 15:44:03 EDT 2009 cris unknown
Sqlite:
Sqlite: sqlite-3.6.14
./configure --host=cris-axis-linux-gnu
--prefix=/AEMDEV/83+/devboard-R2_10/target/cris-axis-linux-gnu
--enable-static=yes --enable-shared=yes
On 11 Aug 2009, at 4:50pm, Radcon Entec wrote:
> At startup, the application creates three tables. If the file
> previously existed, the create table queries fail. My code checks
> the error message, and if it indicates that the table previously
> existed, it ignores the error.
Use the
--- On Tue, 8/11/09, Radcon Entec wrote:
> From: Radcon Entec
> Subject: [sqlite] Cannot insert records into a table after dropping and
> recreating it
> To: sqlite-users@sqlite.org
> Date: Tuesday, August 11, 2009, 10:50 AM
> Greetings!
>
> I
Greetings!
I have an application that uses an SQLite database file that may or may not
exist when the application starts. At startup, the application creates three
tables. If the file previously existed, the create table queries fail. My
code checks the error message, and if it indicates
On 11 Aug 2009, at 11:25am, Gilles Ganault wrote:
> I'd like to fill a ComboBox widget with the list of tables in a
> database file, using the free Aducom wrapper.
>
> Would someone have some code handy on how to extract this information
> from a SQLite database, ie. the equivalent for the CLI
Hello
I'd like to fill a ComboBox widget with the list of tables in a
database file, using the free Aducom wrapper.
Would someone have some code handy on how to extract this information
from a SQLite database, ie. the equivalent for the CLI command
".tables".
Thank you.
More detailed testing revealed that it is not a reader but the writer who
observes a long wait time. The writer appears to get unlocked after all
reader-threads have completed. This must be a case of writer starvation. I had
not realized that shared cache mode is, by default, prone to this. The
24 matches
Mail list logo