Re: [sqlite] Multiuser experience under win32 anyone?

2004-03-15 Thread Jay Macaulay
> Hi all,
>
> Has anyone been doing some testing on how sqlite performs in a networked
> multiuser environment?
>
> Cheers

I think what I and a lot of other users do in a situation like this is to
wrap a server around SQLite itself that serializes requests so that you only
ever run the DB from one thread.  While it requires more work, I have found
that SQLite performs quite well like this, depending on what you're trying
to do.  Obviously only one requests can be processed at a time, but since in
most situations SQLite is very fast, you don't notice it unless you are
really trying to do some complex querying.  I guess it just really comes
down to what you're trying to accomplish and what you need the database for.

Jay Macaulay


-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]



RE: [sqlite] Multiuser experience under win32 anyone?

2004-03-15 Thread Fred Williams
In the real world you only need one user to start a transaction, and go to
lunch, then the performance gets, as we say in the south, "mighty poor!'

Fred

-Original Message-
From: Balthasar Indermuehle [mailto:[EMAIL PROTECTED]
Sent: Monday, March 15, 2004 10:34 AM
To: 'Williams, Ken'; [EMAIL PROTECTED]
Subject: RE: [sqlite] Multiuser experience under win32 anyone?


Thanks, I have read that from the FAQ, but how well/not well does it
perform effectively (numbers)? Has anyone run tests on how many selects
and inserts a second can be performned from, say, 5 clients that access
the same DB located on a network drive?

Balthasar Indermuehle
Inside Systems GmbH
http://www.inside.net

"Louis Pasteur's theory of germs is ridiculous fiction."
Pierre Pachet, Professor of Physiology at Toulouse, 1872 .



-Original Message-
From: Williams, Ken [mailto:[EMAIL PROTECTED]
Sent: Monday, March 15, 2004 09:19
To: 'Balthasar Indermuehle'; [EMAIL PROTECTED]
Subject: RE: [sqlite] Multiuser experience under win32 anyone?




> -Original Message-
> From: Balthasar Indermuehle [mailto:[EMAIL PROTECTED]
> Sent: Monday, March 15, 2004 10:03 AM
> To: [EMAIL PROTECTED]
> Subject: [sqlite] Multiuser experience under win32 anyone?
>
>
> Hi all,
>
> Has anyone been doing some testing on how sqlite performs in
> a networked multiuser environment?

If you mean that multiple users will try to connect to the DB at the
same time, I'd say SQLite isn't going to perform well.  Traditional
client-server databases put a lot of work into supporting this model
efficiently (by supporting things like table-level or row-level locking,
rather than database-level like SQLite) and SQLite puts no work into it.

This assumes that multiple users will be reading *and* writing.  If
they're all just reading, SQLite might be fine.

 -Ken

-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]



-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]



-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]



Re: [sqlite] Multiuser experience under win32 anyone?

2004-03-15 Thread D. Richard Hipp
Balthasar Indermuehle wrote:
Has anyone been doing some testing on how sqlite performs in a networked
multiuser environment?
Most network filesystems are *really slow*.  This, in turn, makes
SQLite run really slow when used over a network filesystem.
You should look into a client/server model if you want multi-user
access to a networked database.  Either go with a client/server
RDBMS or put a server wrapper around SQLite.
--
D. Richard Hipp -- [EMAIL PROTECTED] -- 704.948.4565
-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]


[sqlite] Magically turning a memory blob into an SQLite memory db?

2004-03-15 Thread Ng Pheng Siong
Hi,

How difficult is it to have a blob in memory which contains a binary image
of an SQLite disk database, from which to create an instance of an SQLite
memory database?

IOW, instead of creating an empty memory database and then populating it
with data, I want to go from a blob of memory to an application-specific
fully populated memory database.

For those who are familar with OpenSSL programming, I'm talking about
something similar to its BIO_s_mem, into which one can stuff a binary
string representing the serialisation of, say, an X.509 cert, and create (a
pointer to) an X509 struct out of it.

An idiomatic usage in OpenSSL's context is to keep a trusted CA cert or RSA
public key within the application binary, instead of in a disk file which
can be edited.

TIA. Cheers.

-- 
Ng Pheng Siong <[EMAIL PROTECTED]> 

http://firewall.rulemaker.net -+- Firewall Change Management & Version Control
http://sandbox.rulemaker.net/ngps -+- Open Source Python Crypto & SSL

-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]



RE: [sqlite] Multiuser experience under win32 anyone?

2004-03-15 Thread Balthasar Indermuehle
Thanks, I have read that from the FAQ, but how well/not well does it
perform effectively (numbers)? Has anyone run tests on how many selects
and inserts a second can be performned from, say, 5 clients that access
the same DB located on a network drive?

Balthasar Indermuehle
Inside Systems GmbH
http://www.inside.net

"Louis Pasteur's theory of germs is ridiculous fiction." 
Pierre Pachet, Professor of Physiology at Toulouse, 1872 .  



-Original Message-
From: Williams, Ken [mailto:[EMAIL PROTECTED] 
Sent: Monday, March 15, 2004 09:19
To: 'Balthasar Indermuehle'; [EMAIL PROTECTED]
Subject: RE: [sqlite] Multiuser experience under win32 anyone?




> -Original Message-
> From: Balthasar Indermuehle [mailto:[EMAIL PROTECTED]
> Sent: Monday, March 15, 2004 10:03 AM
> To: [EMAIL PROTECTED]
> Subject: [sqlite] Multiuser experience under win32 anyone?
> 
> 
> Hi all,
> 
> Has anyone been doing some testing on how sqlite performs in
> a networked multiuser environment?

If you mean that multiple users will try to connect to the DB at the
same time, I'd say SQLite isn't going to perform well.  Traditional
client-server databases put a lot of work into supporting this model
efficiently (by supporting things like table-level or row-level locking,
rather than database-level like SQLite) and SQLite puts no work into it.

This assumes that multiple users will be reading *and* writing.  If
they're all just reading, SQLite might be fine.

 -Ken

-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]



-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]



RE: [sqlite] Multiuser experience under win32 anyone?

2004-03-15 Thread Williams, Ken


> -Original Message-
> From: Balthasar Indermuehle [mailto:[EMAIL PROTECTED]
> Sent: Monday, March 15, 2004 10:03 AM
> To: [EMAIL PROTECTED]
> Subject: [sqlite] Multiuser experience under win32 anyone?
> 
> 
> Hi all,
> 
> Has anyone been doing some testing on how sqlite performs in 
> a networked multiuser environment?

If you mean that multiple users will try to connect to the DB at the same
time, I'd say SQLite isn't going to perform well.  Traditional client-server
databases put a lot of work into supporting this model efficiently (by
supporting things like table-level or row-level locking, rather than
database-level like SQLite) and SQLite puts no work into it.

This assumes that multiple users will be reading *and* writing.  If they're
all just reading, SQLite might be fine.

 -Ken

-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]



[sqlite] Multiuser experience under win32 anyone?

2004-03-15 Thread Balthasar Indermuehle
Hi all,

Has anyone been doing some testing on how sqlite performs in a networked
multiuser environment?

Cheers


Balthasar Indermuehle
Inside Systems GmbH
http://www.inside.net

"Louis Pasteur's theory of germs is ridiculous fiction." 
Pierre Pachet, Professor of Physiology at Toulouse, 1872 .  


-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]



Re: [sqlite] Borland C++ Builder 6

2004-03-15 Thread D. Richard Hipp
eno wrote:
Drew, Stephen wrote:

Further to my original question, does anyone know why the SQLite library
compiles when (and only when, as far as I can tell) the following 
lines in
SQLiteInt.h :

#include 
#include 
#include 
#include 
are replaced with the original C equivalents:

#include 
#include 
#include 
#include 
Thanks,
Steve


seems that Borland Builder 6 doesn't have state-of-the-art C++ header 
files, or is holding these in separate C++ include directories, thus not 
finding them when compiling a ".c" C source file.

BTW: As of 2.8.12 the C header files are included in sqliteint.h, at 
least in my copy.

cvs annotate shows me that the standard C header files have been unchanged
in the sqliteInt.h file since before the first public release in May of
2000.  Those four lines have never been different.
I don't know where Mr. Drew got his "" changes.  But he didn't
get them from the SQLite source tree.
--
D. Richard Hipp -- [EMAIL PROTECTED] -- 704.948.4565
-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]


Re: [sqlite] Borland C++ Builder 6

2004-03-15 Thread eno
Drew, Stephen wrote:
Further to my original question, does anyone know why the SQLite library
compiles when (and only when, as far as I can tell) the following lines in
SQLiteInt.h :
#include 
#include 
#include 
#include 
are replaced with the original C equivalents:

#include 
#include 
#include 
#include 
Thanks,
Steve
seems that Borland Builder 6 doesn't have state-of-the-art C++ header 
files, or is holding these in separate C++ include directories, thus not 
finding them when compiling a ".c" C source file.

BTW: As of 2.8.12 the C header files are included in sqliteint.h, at 
least in my copy.

/eno

-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]


RE: [sqlite] Borland C++ Builder 6

2004-03-15 Thread Drew, Stephen
Further to my original question, does anyone know why the SQLite library
compiles when (and only when, as far as I can tell) the following lines in
SQLiteInt.h :

#include 
#include 
#include 
#include 

are replaced with the original C equivalents:

#include 
#include 
#include 
#include 

Thanks,
Steve

-Original Message-
From: Drew, Stephen [mailto:[EMAIL PROTECTED]
Sent: Monday, March 15, 2004 11:36 AM
To: '[EMAIL PROTECTED]'
Subject: [sqlite] Borland C++ Builder 6


Has anyone any experience of compiling the SQLite library with Borland C++
Builder 6 and it's shipped version of STLPort?
 
Regards,
Steve

-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]



[sqlite] Borland C++ Builder 6

2004-03-15 Thread Drew, Stephen
Has anyone any experience of compiling the SQLite library with Borland C++
Builder 6 and it's shipped version of STLPort?
 
Regards,
Steve