Re: [sqlite] Virtual table used to query big external database
On Tue, Apr 01, 2008 at 08:17:54PM -0500, Mike Owens scratched on the wall: > The main reason why SQLite's practical limit is in the 10s of GBs as > opposed to TBs (theoretical) is due to how it tracks dirty pages. > [...] > SQLite tracks dirty pages with a bitmap which is > allocated before each transaction. The size of the bitmap is > proportional to the size (not in rows but in pages) of the database Doesn't the "bitvec" structure introduced in 3.5.7 help with that somewhat? I was under the impression it allowed for sparse bitmaps, which seems like it would be the common case for INSERTs, UPDATEs, and DELETEs. I suppose a huge delete or a DROP [TABLE|INDEX] might still cause the bitmap to get pretty big, but those are more unusual operations. -j -- Jay A. Kreibich < J A Y @ K R E I B I.C H > "'People who live in bamboo houses should not throw pandas.' Jesus said that." - "The Ninja", www.AskANinja.com, "Special Delivery 10: Pop!Tech 2006" ___ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
Re: [sqlite] Virtual table used to query big external database
The main reason why SQLite's practical limit is in the 10s of GBs as opposed to TBs (theoretical) is due to how it tracks dirty pages. This is described in the "Appropriate Uses" page (http://www.sqlite.org/whentouse.html) but I'll rehash it here for convenience. SQLite tracks dirty pages with a bitmap which is allocated before each transaction. The size of the bitmap is proportional to the size (not in rows but in pages) of the database (256 bytes for every 1Mb of database), so as the database grows, the amount of memory allocated before each transaction grows. When you get into the GB range, you are starting to allocate in the MB range of dirty page map memory per transaction, which starts to take its toll on performance. I could be wrong, but from what I know about virtual tables, there is no such correlation between virtual table size and the dirty page bitmap, as SQLite has no idea how big a virtual table is, nor does it manage the data within the vtable. Furthermore, all SQLite really does in a SELECT statement on a vtable is call your code to iterate over it. So really the only performance issue is how long it takes your code to iterate over your vtable. Thus, your table could be in the TB range, and as long as you are fine with iterating over its contents, there is no additional performance issues to speak of. There are ways to implement virtual tables such that you can limit how much of the table is scanned for certain queries, avoiding having to scan the whole thing every time. I wrote an article that touches on this using the match() function. Its available online: http://www.ddj.com/database/202802959 IMO, virtual tables are one of the most powerful and unique features of SQLite. There is a bit of a learning curve, but it's amazing what you can do with them. It sounds like you going to have to iterate over your external table one way or the other. I see no reason why the vtable approach would be any slower than any other approach that iterates over the data. Having said that, while iterating over a large vtable is not a big deal (as your program will just step through it one row at a time), you need to be careful about getting too fancy with your SQL as you may end up triggering a lot of background IO. For example, if you tack on an ORDER BY which sorts one of the columns of your vtable, SQLite will end up essentially copying the vtable contents into a temporary file and sorting it, which may or may not be a strain on your system depending on how big your table is (e.g. your vtable is 30Gb and your /tmp folder is on a 10Gb partition). So think through what you are doing when going beyond a simple SELECT * from big_vtable. -- Mike On Tue, Apr 1, 2008 at 3:12 PM, Aladdin Lampé <[EMAIL PROTECTED]> wrote: > > Hi all! > > Very often, when people ask this list why they have trouble managing in > sqlite a "big table" (50 million lines or more than 10 Go), they are told > that sqlite is an embedded database and is not meant to be used for very big > databases/tables. > > I'm currently in the process of designing a specific, read-only, sqlite > "virtual table" in order to enable sqlite to access data stored in an > external database which is specially designed to handle very big tables. > > My final objective is to be able to easily query a big external table > (stored in another database) through the - excellent - sqlite interface. > > Now I have this terrible doubt: will the existing sqlite "limitations" for > big sqlite tables also apply to my read-only virtual tables? > > Thus... am I currently losing my time developing such a "virtual table" with > this objective in mind? Or is there a better way to achieve my objective? > > Thank you for your help! > > > _ > Votre contact a choisi Hotmail, l'e-mail ultra sécurisé. Créez un compte > gratuitement ! > http://www.windowslive.fr/hotmail/default.asp > ___ > sqlite-users mailing list > sqlite-users@sqlite.org > http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users > ___ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
Re: [sqlite] 2.8.17 --> 3.5.6 performance regression
Mike, You guessed correctly! I inserted a "PRAGMA synchronous = NORMAL;" into the SQLite 3 version of my test suite, and obtained the same performance as I observed with SQLite 2. I should have realized that the kind of slowdown I was observing could only be explained by extra disk I/O operations. Thanks for your help! - Richard > I would guess PRAGMA synchronous. Per documentation: > > "In SQLite version 2, the default value is NORMAL. For version 3, the > default was changed to FULL." > > Try setting it to NORMAL for v3 tests and see what that does. > > -- Mike > > On Thu, Mar 27, 2008 at 11:06 PM, Richard Klein > <[EMAIL PROTECTED]> wrote: >> I've recently upgraded from SQLite 2.8.17 >> to 3.5.6. >> >> Upon running a test program that measures >> the execution time of SQL statements typical >> for my application, I've noticed a definite >> performance degradation: >> >> INSERT: 34% slowdown >> UPDATE: 47%" >> DELETE: 50%" >> >> Has anyone else noticed this? I haven't done >> any profiling of the code (yet) to see where >> the time is being spent. >> >> - Richard >> >> ___ >> sqlite-users mailing list >> sqlite-users@sqlite.org >> http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users >> > ___ > sqlite-users mailing list > sqlite-users@sqlite.org > http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users ___ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
Re: [sqlite] Virtual table used to query big external database
The only limitation imposed by SQL that I can think of would be inefficiencies in its query plan builder. That is the part that figures out how to use the indexes available in the database in order to execute the SQL query most efficiently. So it really depends on what type of SQL queries you are going to be running against this huge DB. The dangers are easy to evaluate: Create a quick-and-dirty dummy virtual table mechanism, and respond to the xBestIndex/xFilter functions. You will notice that for certain queries, xBestIndex/xFilter does not get used. That means that the SQLite engine is going to have to walk through your entire table, evaluating the conditions on each field. This is obviously what you wish to avoid. As an example, I noticed briefly (I did not investigate thoroughly) that having an OR condition in a query would prevent the indexes from being used. That was some time ago, and it was before the rewrite of the SQL VM, so I don't know if that still applies. You'll have to investigate your potential queries yourself. A simple query such as "WHERE myvalue > 100" should definitely invoke the use of your own indexes. Ben On Tue, Apr 1, 2008 at 10:12 PM, Aladdin Lampé <[EMAIL PROTECTED]> wrote: > > Hi all! > > Very often, when people ask this list why they have trouble managing in > sqlite a "big table" (50 million lines or more than 10 Go), they are told > that sqlite is an embedded database and is not meant to be used for very big > databases/tables. > > I'm currently in the process of designing a specific, read-only, sqlite > "virtual table" in order to enable sqlite to access data stored in an > external database which is specially designed to handle very big tables. > > My final objective is to be able to easily query a big external table (stored > in another database) through the - excellent - sqlite interface. > > Now I have this terrible doubt: will the existing sqlite "limitations" for > big sqlite tables also apply to my read-only virtual tables? > > Thus... am I currently losing my time developing such a "virtual table" with > this objective in mind? Or is there a better way to achieve my objective? > > Thank you for your help! > > _ > Votre contact a choisi Hotmail, l'e-mail ultra sécurisé. Créez un compte > gratuitement ! > http://www.windowslive.fr/hotmail/default.asp > ___ > sqlite-users mailing list > sqlite-users@sqlite.org > http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users > ___ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
[sqlite] Deadlock with attached databases
Hello All, Is it possible that SQLite deadlocks if it uses attached databases? I am not sure, but if check_all_db_and_lock_all_of_them() is not atomic, it can be... -- Alexander Batyrshin aka bash bash = Biomechanica Artificial Sabotage Humanoid ___ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
[sqlite] Virtual table used to query big external database
Hi all! Very often, when people ask this list why they have trouble managing in sqlite a "big table" (50 million lines or more than 10 Go), they are told that sqlite is an embedded database and is not meant to be used for very big databases/tables. I'm currently in the process of designing a specific, read-only, sqlite "virtual table" in order to enable sqlite to access data stored in an external database which is specially designed to handle very big tables. My final objective is to be able to easily query a big external table (stored in another database) through the - excellent - sqlite interface. Now I have this terrible doubt: will the existing sqlite "limitations" for big sqlite tables also apply to my read-only virtual tables? Thus... am I currently losing my time developing such a "virtual table" with this objective in mind? Or is there a better way to achieve my objective? Thank you for your help! _ Votre contact a choisi Hotmail, l'e-mail ultra sécurisé. Créez un compte gratuitement ! http://www.windowslive.fr/hotmail/default.asp ___ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
Re: [sqlite] insert of string with line breaks
How are you verifying that the db only received one line from the
description? using sqlite3.exe or some other tool, or AIR?
SQLite itself does not have issues with storing line breaks, but it's
possible the AIR wrapper is losing them somewhere or more likely it's a
display/formatting issue.
One thing to keep in mind is that the Flash Player has funky internal
representations of line breaks in strings. I believe it uses '\r' whereas
windows uses '\r\n' and *nix uses '\n'. I had to make some changes in
Fluorine to handle line breaks properly with AMF and .NET integration. It
could be that the line break is in the db but the tool you're using to view
the data doesn't recognize it as a valid line break.
HTH,
Sam
On Wed, Mar 26, 2008 at 10:33 AM, Raymond Camden <[EMAIL PROTECTED]>
wrote:
> I've got an AIR app that does simple inserts into a SQLite db. The
> insert looks like so:
>
> var insStmt:SQLStatement = new SQLStatement();
> insStmt.sqlConnection = dbconnection;
> sql = "insert into hours(description,projectidfk,clientidfk,hours,date)
> values(:description,:projectidfk,:clientidfk,:hours,:date)";
> insStmt.parameters[":description"] = descriptionfield.text;
> insStmt.parameters[":projectidfk"] = projectfield.selectedItem.id;
> insStmt.parameters[":clientidfk"] = projectfield.selectedItem.clientid;
>
> insStmt.parameters[":hours"] = hoursworkedfield.value;
> insStmt.parameters[":date"] = dateworkedfield.selectedDate;
>
> trace('desc is '+descriptionfield.text);
> insStmt.text = sql;
> insStmt.addEventListener(SQLEvent.RESULT, refreshHandler);
>
>
> Notice I'm using bound parameters for all my values. The description
> field is a text box. When I enter text with line breaks in it, the db
> only gets line one of description.
>
> Any ideas why? A string is a string even if it has line breaks in it -
> afaik.
>
>
> --
>
> ===
> Raymond Camden, Owner of Camden Media, Inc.
>
> Email: [EMAIL PROTECTED]
> Blog : www.coldfusionjedi.com
> AOL IM : cfjedimaster
>
> "My ally is the Force, and a powerful ally it is." - Yoda
> ___
> sqlite-users mailing list
> sqlite-users@sqlite.org
> http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
>
--
-
We're Hiring! Seeking passionate Flex, C#, or C++ (RTSP, H264) developer.
Position is in the Washington D.C. metro area. Contact
[EMAIL PROTECTED]
___
sqlite-users mailing list
sqlite-users@sqlite.org
http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
Re: [sqlite] Trigger's actions and callbacks
> Hi, all.
>
> I have the following problem: I have some processes that access the same db
> file. When one of them would change some data inside the db it must notify
> all other processes about change AND send them changed data (or it's unique
> description).
>
> But I didn't find any method to perform this using db triggers or callbacks.
> On some certain reasons I am not allowed to write wrapper that will send
> notifications using IPC. I would like db to perform this task.
I have a similar problem, which I solved using db triggers, along with a
callback mechanism involving a user-defined function.
> As for triggers and callbacks (hooks): they are fired before the data is
> stored to the db (file), so that I coudn't even read changed data back
> inside hook. sqlite_update_hook() returnes rowID, but I didn't find any API
> to use this value...
Triggers can be defined to happen BEFORE or AFTER the database event.
Either way, you can access the old and new values of a field using the
OLD and NEW keywords. See www.sqlite.org->Documentation->SQL Syntax->CREATE
TRIGGER
for details.
> I would appriciate any ideas or comments.
> Thanks in advance.
Below is a memo I wrote to my boss describing how we could do change
notification
in SQLite. It is written in terms of SQLite 2; to use the same mechanism in
SQLite 3, you would use sqlite3_create_function() instead of
sqlite_create_function()
to register your user-defined function. Note that in SQLite 3, the parameters
to
sqlite(3)_create_function() have changed, as have the parameters passed to your
user-defined function.
If people are interested, I can turn this into documentation (Wiki article?)
for the SQLite website.
Regards,
- Richard Klein
==
Hi Dave,
As you mentioned earlier, the PVR 2.0 Scheduler maintains its own data
structures
that must be kept in sync with the database. Therefore, if the app modifies the
database, the Scheduler needs to be notified of the change.
In playing around with SQLite, I have come up with a change notification
mechanism
that makes use of a user-defined function is conjunction with a SQL entity
known as
a "trigger".
A trigger is a SQL statement that is associated with a specified table, and
with a
specified action (INSERT, UPDATE, or DELETE) on that table. This SQL statement
is
automatically executed when the associated action is performed on the associated
table.
To create an INSERT trigger on the 'requests' table, you would execute the
following
SQL statement:
CREATE TRIGGER InsertedRequest AFTER INSERT ON requests
BEGIN
SELECT ChangeNotify('requests', 'INSERT', NEW.ulRequestId);
END;
UPDATE and DELETE triggers would be created in a similar fashion.
The above statement defines a trigger, named 'InsertedRequest', that will fire
*after*
an INSERT operation is performed on the 'requests' table. This trigger will
call a
user-defined function named 'ChangeNotify'. ChangeNotify takes three
parameters: The
name of the affected table, the operation (INSERT, UPDATE, or DELETE) that was
performed,
and the Request ID (ulRequestId) of the affected row.
The ChangeNotify function is defined as follows:
static void changeNotify(sqlite_func* context, int argc, const char** argv)
{
const char *table, *action, *rowid;
assert(argc == 3);
table = argv[0];
action = argv[1];
rowid = argv[2];
dprintf("Database changed: table = %s, action = %s, rowid = %s\n", table,
action, rowid);
}
Like all user-defined functions, changeNotify() is called with three
parameters: An opaque
pointer 'context', the argument count 'argc' (which will be '3' in this case),
and the three
arguments (table, action, and row in this case).
This simple version of changeNotify() simply prints out the calling parameters.
The production
version will invoke a callback that has been registered by the Scheduler.
Of course, like all user-defined functions, changeNotify() must be registered
with SQLite:
status = sqlite_create_function(pDb, "ChangeNotify", 3, changeNotify,
NULL);
if (status != SQLITE_OK) {
/* error */
}
where
pDbis an opaque pointer;
"ChangeNotify" is the name of the user-defined function;
3 is the number of parameters taken by the user-defined function;
and
changeNotify is the address of the user-defined function.
___
sqlite-users mailing list
sqlite-users@sqlite.org
http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
Re: [sqlite] mem5 ("buddy system") usable?
On Apr 1, 2008, at 2:37 PM, Richard Klein wrote: >> > Fair enough. But can I assume that mem5 *does* currently > work, as far as you know? (I'd really like to use it, as > I'm using SQLite on an embedded system.) > mem5 works as far as I know. D. Richard Hipp [EMAIL PROTECTED] ___ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
[sqlite] Exception running javasqlite on kaffe for arm architecture
Hi, I'm trying to run sqlite using the java wrapper at http://www.ch-werner.de/javasqlite/ on an arm linux target using kaffe 1.1.7 and jamvm. It appears that I'm having a problem with the JNI layer. I am able to run $sqlite3 from the command line, creating/editing database files with no problems. When I run it as a java app: kaffe -jar sqlite.jar data.db I get an exception: kaffe-bin: exception.c:106: vmExcept_setJNIFrame: Assertion `fp != (JNIFrameAddress)0' failed. Oddly, it fails the first time when it tries to open the database, but the database file is created. the next time it fails when I try to do a select on a table on the database, with the same exception. I followed the instructions to cross compile sqlite and the java wrappers for our arm target, and don't see anything obvious that might be causing this. Just wondering if anyone has come across this issue before. Lastly, the versions I built for the x86 target run fine on the x86 kaffe and jamvm. gcc version is 3.3.1 kernel is Linux version 2.4.20__mvlcee31-omap730_gsm_gprs ([EMAIL PROTECTED]) (gcc version 3.3.1 (MontaVista 3.3.1-7.0.2.0401382 2004-10-08)) #1 Thu Nov 29 16:32:50 HKT 2007 kernel headers are Thanks, Charlie Hamilton D2 Technologies [EMAIL PROTECTED] ___ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
Re: [sqlite] mem5 ("buddy system") usable?
> On Mar 31, 2008, at 10:00 PM, Richard Klein wrote: >> I just downloaded 3.5.7, and noticed that there is a >> new memory allocator mem5.c, which apparently uses the >> "buddy system" to allocate power-of-two-sized chunks >> of memory from a static pool. This allocator is used >> by defining SQLITE_POW2_MEMORY_SIZE. >> >> Is it okay to use this allocator, or is it only exper- >> imental (and therefore liable to disappear in a future >> release)? >> > > I'm not making any promises about any of the current five > memory allocators. I might decide to replace them all tomorrow. > > But mem5 is high on the list of memory allocators to keep > since it can, under some circumstances, guarantee not to > fragment memory, which is a desirable property for > embedded systems. > > D. Richard Hipp > [EMAIL PROTECTED] > Fair enough. But can I assume that mem5 *does* currently work, as far as you know? (I'd really like to use it, as I'm using SQLite on an embedded system.) - Richard Klein ___ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
Re: [sqlite] Trigger's actions and callbacks
"Vladimir Volkov" <[EMAIL PROTECTED]> wrote in message news:[EMAIL PROTECTED] > I have the following problem: I have some processes that access the > same db file. When one of them would change some data inside the db > it must notify all other processes about change AND send them changed > data (or it's unique description). > > But I didn't find any method to perform this using db triggers or > callbacks. Correct. SQLite is not an interprocess communication mechanism. You will have to come up with your own scheme, outside SQLite. > On some certain reasons I am not allowed to write wrapper > that will send notifications using IPC. You mean, you need to perform IPC but you are not allowed to write code that performs IPC? That's quite a corner you are painted into. > I would like db to perform > this task. Then I guess SQLite is not suitable for your problem (though I can't think of any other DBMS that can do that, either). > As for triggers and callbacks (hooks): they are fired before the data > is stored to the db (file), so that I coudn't even read changed data > back inside hook. As long as you use the same connection to access the database as the one that caused the trigger or callback to fire in the first place, you should see the new data. It doesn't matter that the data has not been written to the physical file yet. If you are trying to use a different connection, you would only see changes once the first connection committed its transaction - so don't do that. > sqlite_update_hook() returnes rowID, but I didn't > find any API to use this value... You just run a statement along the lines of select * from dbName.tableName where rowid=? You need to run this statement on the same connection the hook is installed on. Igor Tandetnik ___ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
[sqlite] Trigger's actions and callbacks
Hi, all. I have the following problem: I have some processes that access the same db file. When one of them would change some data inside the db it must notify all other processes about change AND send them changed data (or it's unique description). But I didn't find any method to perform this using db triggers or callbacks. On some certain reasons I am not allowed to write wrapper that will send notifications using IPC. I would like db to perform this task. As for triggers and callbacks (hooks): they are fired before the data is stored to the db (file), so that I coudn't even read changed data back inside hook. sqlite_update_hook() returnes rowID, but I didn't find any API to use this value... I would appriciate any ideas or comments. Thanks in advance. -- View this message in context: http://www.nabble.com/Trigger%27s-actions-and-callbacks-tp16418413p16418413.html Sent from the SQLite mailing list archive at Nabble.com. ___ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
Re: [sqlite] sqlite-users Digest, Vol 3, Issue 99
Hi Dennis, may be I understand now. I have enabled mail delivery in my sqlite users account options (which was not enabled before), now I should receive all messages so that I can reply to one of them. May be it was my problem, because I have never received any message from sqlite user mailing list. Thanks for your patience and your help, LB [EMAIL PROTECTED] wrote: Send sqlite-users mailing list submissions to sqlite-users@sqlite.org To subscribe or unsubscribe via the World Wide Web, visit http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users or, via email, send a message with subject or body 'help' to [EMAIL PROTECTED] You can reach the person managing the list at [EMAIL PROTECTED] When replying, please edit your Subject line so it is more specific than "Re: Contents of sqlite-users digest..." Today's Topics: 1. Re: Scope of sqlite3_changes(), how to reset it? (Igor Tandetnik) 2. Re: DBD::SQLite::Amalgamation? (rahed) 3. Re: DBD::SQLite::Amalgamation? (Jim Dodgen) 4. Re: DBD::SQLite::Amalgamation? (P Kishor) 5. Re: DBD::SQLite::Amalgamation? (Jim Dodgen) 6. Re: DBD::SQLite::Amalgamation? (P Kishor) 7. Re: Indexing and Search speed ([EMAIL PROTECTED]) 8. When I try to .read I get a "can't open" message (Douglas McCarroll) 9. Re: When I try to .read I get a "can't open" message (P Kishor) 10. Re: When I try to .read I get a "can't open" message (Steven Fisher) 11. Re: When I try to .read I get a "can't open" message (Clark Christensen) -- Message: 1 Date: Mon, 31 Mar 2008 12:03:32 -0400 From: "Igor Tandetnik" Subject: Re: [sqlite] Scope of sqlite3_changes(), how to reset it? To: sqlite-users@sqlite.org Message-ID: Michael Schlenker wrote: > 1. Is sqlite3_changes() reset to 0 when a new sqlite3_exec() is > started? May or may not. E.g. if you execute a SELECT statement it doesn't affect change count. > 2. Does a PRAGMA index_list() manipulate sqlite3_changes() in any way? I haven't tried it, but I don't see why it should. > 3. How can i reset sqlite3_changes() manually if its not done on > sqlite3_exec()? "Reset" as in set to zero? Run a statement that is guaranteed not to touch any rows, e.g. delete from sometable where 0; Igor Tandetnik -- Message: 2 Date: Mon, 31 Mar 2008 19:31:20 + From: rahed Subject: Re: [sqlite] DBD::SQLite::Amalgamation? To: "General Discussion of SQLite Database" Message-ID: Content-Type: text/plain; charset=UTF-8 On 3/31/08, Jim Dodgen wrote: > Any Perl people out there using this yet (version 3.5.7)? > > I'm continually having problems when doing "make test" Today I tried and installed on Solaris successfully. Both the module 1.14 and sqlite3 amalgamation. -- Radek -- Message: 3 Date: Mon, 31 Mar 2008 12:59:58 -0700 From: "Jim Dodgen" Subject: Re: [sqlite] DBD::SQLite::Amalgamation? To: sqlite-users@sqlite.org Message-ID: <[EMAIL PROTECTED]> Content-Type: text/plain; charset=ISO-8859-1 Humm... I'll keep looking. thanks On 3/30/08, Jim Dodgen wrote: > Any Perl people out there using this yet (version 3.5.7)? > > I'm continually having problems when doing "make test" > > It hangs forever and never completes > > I have tried this on multiple boxes running both Fedora and Redhat ES, > all 64 bit. > -- Jim Dodgen [EMAIL PROTECTED] -- Message: 4 Date: Mon, 31 Mar 2008 15:14:07 -0500 From: "P Kishor" Subject: Re: [sqlite] DBD::SQLite::Amalgamation? To: "General Discussion of SQLite Database" Message-ID: Content-Type: text/plain; charset=ISO-8859-1 On 3/30/08, Jim Dodgen wrote: > Any Perl people out there using this yet (version 3.5.7)? > > I'm continually having problems when doing "make test" > > It hangs forever and never completes > > I have tried this on multiple boxes running both Fedora and Redhat ES, > all 64 bit. Jim, I am using 3.5.6 (using the amalgamation... see the instructions I put on the SQLite wiki) and that worked well for me on both a Mac Leopard and a RHEL ES3 box. I see no reason to upgrade to 3.5.7 yet. That said, I also not working with 64 bit system. You might get better luck in getting an answer if you post at least the tail portion of the make output, otherwise no one has any idea where your process hangs. -- Puneet Kishor http://punkish.eidesis.org/ Nelson Institute for Environmental Studies http://www.nelson.wisc.edu/ Open Source Geospatial Foundation (OSGeo) http://www.osgeo.org/ -- Message: 5 Date: Mon, 31 Mar 2008 13:34:42 -0700 From: "Jim Dodgen" Subject: Re: [sqlite] DBD::SQLite::Amalgamation? To: [EMAIL PROTECTED], "General Discussion of SQLite Database" Message-ID: <[EMAIL PROTECTED]> Content-Type: text/plain; charset=ISO-8859-1 On 3/31/08, P Kishor wrote: > On 3/30/08, Jim Dodgen wrote: > > Any Perl people out there using this yet (version 3.5.7)? > > > > I'm continually
Re: [sqlite] DBD::SQLite::Amalgamation?
On 3/31/08, Jim Dodgen <[EMAIL PROTECTED]> wrote: > It makes ok, here is the output from "make test". the hang is after > "t/06error" it never finishes > > DBD-SQLite-Amalgamation-3.5.7 > > # make test > PERL_DL_NONLAZY=1 /usr/bin/perl "-MExtUtils::Command::MM" "-e" > "test_harness(0, 'blib/lib', 'blib/arch')" t/*.t > t/00basic...ok > t/01logon...ok > t/02cr_tableok > t/03insert..ok > t/04select..ok > t/05tranok > t/06error...ok 1/2 > > (this never finishes) In the beginnig there are unlink and connect commands in t/06 which were successful in previous tests. So something may go wrong with eval of the do test which should throw an error in [EMAIL PROTECTED] You could step in with the debugger to see where's the fault. -- Radek ___ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
