Hi all,
Happy to write again in this mailing-list, hope you are all doing well guys !
I am encountering a small problem running my old xquery modules. I guess that it is a classical issue, but I can't remember how to look into BaseX archive :(
Here is the description of the problem : I am running a query, calling module_of_mine.xqm, on two different environments.
The first one is eclipse, linked to the latest interpreter from BaseX80-20140923.183154 (but I tested also against older interpreter). In this way, this query saturated my 12 G Java machine after some minutes.
Then I installed module_of_mine.xqm into BaseX repo, and ran the same query using basex gui with 4G heap size. The query ran perfectly in some 30 sec.
My first penny guess is that the built-in BaseX xquery profiler is doing some nice work there ? If so, is there a way to activate / desactivate this profiler from eclipse or from some configuration files ?
Cheers,
Jean-Marc
Hello Jean-Marc,
I am not sure how you are calling BaseX from within Eclipse, I remember there are different ways. Normally, I would expect it to simply execute basex, so there shouldn't be any difference. The query optimization should be done in both cases. However, it could be for example that your BaseX integration in eclipse uses XQJ, which could possibly introduce some performance bottlenecks.
So if you could shortly describe how you set up BaseX within Eclipse, that would be great.
Cheers, Dirk
On 09/24/2014 06:33 AM, jean-marc Mercier wrote:
Hi all,
Happy to write again in this mailing-list, hope you are all doing well guys !
I am encountering a small problem running my old xquery modules. I guess that it is a classical issue, but I can't remember how to look into BaseX archive :(
Here is the description of the problem : I am running a query, calling module_of_mine.xqm, on two different environments.
The first one is eclipse, linked to the latest interpreter from BaseX80-20140923.183154 (but I tested also against older interpreter). In this way, this query saturated my 12 G Java machine after some minutes.
Then I installed module_of_mine.xqm into BaseX repo, and ran the same query using basex gui with 4G heap size. The query ran perfectly in some 30 sec.
My first penny guess is that the built-in BaseX xquery profiler is doing some nice work there ? If so, is there a way to activate / desactivate this profiler from eclipse or from some configuration files ?
Cheers,
Jean-Marc
Hi Dirk,
Thanks for your answer. I am setting up eclipse indigo, following the step-up procedure http://docs.basex.org/wiki/Integrating_Eclipse. The only difference is that I set using -Xmx12G as Interpreter Arguments.
Cheers
2014-09-26 11:44 GMT+02:00 Dirk Kirsten dk@basex.org:
Hello Jean-Marc,
I am not sure how you are calling BaseX from within Eclipse, I remember there are different ways. Normally, I would expect it to simply execute basex, so there shouldn't be any difference. The query optimization should be done in both cases. However, it could be for example that your BaseX integration in eclipse uses XQJ, which could possibly introduce some performance bottlenecks.
So if you could shortly describe how you set up BaseX within Eclipse, that would be great.
Cheers, Dirk
On 09/24/2014 06:33 AM, jean-marc Mercier wrote:
Hi all,
Happy to write again in this mailing-list, hope you are all doing well
guys
!
I am encountering a small problem running my old xquery modules. I guess that it is a classical issue, but I can't remember how to look into BaseX archive :(
Here is the description of the problem : I am running a query, calling module_of_mine.xqm, on two different environments.
The first one is eclipse, linked to the latest interpreter from BaseX80-20140923.183154 (but I tested also against older interpreter). In this way, this query saturated my 12 G Java machine after some minutes.
Then I installed module_of_mine.xqm into BaseX repo, and ran the same
query
using basex gui with 4G heap size. The query ran perfectly in some 30
sec.
My first penny guess is that the built-in BaseX xquery profiler is doing some nice work there ? If so, is there a way to activate / desactivate
this
profiler from eclipse or from some configuration files ?
Cheers,
Jean-Marc
-- Dirk Kirsten, BaseX GmbH, http://basex.org |-- Firmensitz: Blarerstrasse 56, 78462 Konstanz |-- Registergericht Freiburg, HRB: 708285, Geschäftsführer: | Dr. Christian Grün, Dr. Alexander Holupirek, Michael Seiferle `-- Phone: 0049 7531 28 28 676, Fax: 0049 7531 20 05 22
Hi Jean-Marc,
I guess you set it up using the standalone version? Could you maybe test using the Client/Server set-up. You could then test with a plain BaseXClient and the eclipse environment. As they both should access the same BaseXServer, there shouldn't be any difference in execution.
But I don't know what XQDT is doing internally and I would pretty much guess the error is somewhere there (because plain BaseX seems to execute just fine). As the development seems to have pretty much stalled, I am not sure we can do much about that...
On 09/26/2014 11:54 AM, jean-marc Mercier wrote:
Hi Dirk,
Thanks for your answer. I am setting up eclipse indigo, following the step-up procedure http://docs.basex.org/wiki/Integrating_Eclipse. The only difference is that I set using -Xmx12G as Interpreter Arguments.
Cheers
2014-09-26 11:44 GMT+02:00 Dirk Kirsten dk@basex.org:
Hello Jean-Marc,
I am not sure how you are calling BaseX from within Eclipse, I remember there are different ways. Normally, I would expect it to simply execute basex, so there shouldn't be any difference. The query optimization should be done in both cases. However, it could be for example that your BaseX integration in eclipse uses XQJ, which could possibly introduce some performance bottlenecks.
So if you could shortly describe how you set up BaseX within Eclipse, that would be great.
Cheers, Dirk
On 09/24/2014 06:33 AM, jean-marc Mercier wrote:
Hi all,
Happy to write again in this mailing-list, hope you are all doing well
guys
!
I am encountering a small problem running my old xquery modules. I guess that it is a classical issue, but I can't remember how to look into BaseX archive :(
Here is the description of the problem : I am running a query, calling module_of_mine.xqm, on two different environments.
The first one is eclipse, linked to the latest interpreter from BaseX80-20140923.183154 (but I tested also against older interpreter). In this way, this query saturated my 12 G Java machine after some minutes.
Then I installed module_of_mine.xqm into BaseX repo, and ran the same
query
using basex gui with 4G heap size. The query ran perfectly in some 30
sec.
My first penny guess is that the built-in BaseX xquery profiler is doing some nice work there ? If so, is there a way to activate / desactivate
this
profiler from eclipse or from some configuration files ?
Cheers,
Jean-Marc
-- Dirk Kirsten, BaseX GmbH, http://basex.org |-- Firmensitz: Blarerstrasse 56, 78462 Konstanz |-- Registergericht Freiburg, HRB: 708285, Geschäftsführer: | Dr. Christian Grün, Dr. Alexander Holupirek, Michael Seiferle `-- Phone: 0049 7531 28 28 676, Fax: 0049 7531 20 05 22
Hi Dirk,
Ok, thanks for the trick, that should work. The only point is that it should be less user friendly for developing than using the eclipse IDE, since I will have to set up the BaseX repo first of running a query.
2014-09-26 12:11 GMT+02:00 Dirk Kirsten dk@basex.org:
Hi Jean-Marc,
I guess you set it up using the standalone version? Could you maybe test using the Client/Server set-up. You could then test with a plain BaseXClient and the eclipse environment. As they both should access the same BaseXServer, there shouldn't be any difference in execution.
But I don't know what XQDT is doing internally and I would pretty much guess the error is somewhere there (because plain BaseX seems to execute just fine). As the development seems to have pretty much stalled, I am not sure we can do much about that...
On 09/26/2014 11:54 AM, jean-marc Mercier wrote:
Hi Dirk,
Thanks for your answer. I am setting up eclipse indigo, following the step-up procedure http://docs.basex.org/wiki/Integrating_Eclipse. The
only
difference is that I set using -Xmx12G as Interpreter Arguments.
Cheers
2014-09-26 11:44 GMT+02:00 Dirk Kirsten dk@basex.org:
Hello Jean-Marc,
I am not sure how you are calling BaseX from within Eclipse, I remember there are different ways. Normally, I would expect it to simply execute basex, so there shouldn't be any difference. The query optimization
should
be done in both cases. However, it could be for example that your BaseX integration in eclipse uses XQJ, which could possibly introduce some performance bottlenecks.
So if you could shortly describe how you set up BaseX within Eclipse,
that
would be great.
Cheers, Dirk
On 09/24/2014 06:33 AM, jean-marc Mercier wrote:
Hi all,
Happy to write again in this mailing-list, hope you are all doing well
guys
!
I am encountering a small problem running my old xquery modules. I
guess
that it is a classical issue, but I can't remember how to look into
BaseX
archive :(
Here is the description of the problem : I am running a query, calling module_of_mine.xqm, on two different environments.
The first one is eclipse, linked to the latest interpreter from BaseX80-20140923.183154 (but I tested also against older
interpreter). In
this way, this query saturated my 12 G Java machine after some
minutes.
Then I installed module_of_mine.xqm into BaseX repo, and ran the same
query
using basex gui with 4G heap size. The query ran perfectly in some 30
sec.
My first penny guess is that the built-in BaseX xquery profiler is
doing
some nice work there ? If so, is there a way to activate / desactivate
this
profiler from eclipse or from some configuration files ?
Cheers,
Jean-Marc
-- Dirk Kirsten, BaseX GmbH, http://basex.org |-- Firmensitz: Blarerstrasse 56, 78462 Konstanz |-- Registergericht Freiburg, HRB: 708285, Geschäftsführer: | Dr. Christian Grün, Dr. Alexander Holupirek, Michael Seiferle `-- Phone: 0049 7531 28 28 676, Fax: 0049 7531 20 05 22
-- Dirk Kirsten, BaseX GmbH, http://basex.org |-- Firmensitz: Blarerstrasse 56, 78462 Konstanz |-- Registergericht Freiburg, HRB: 708285, Geschäftsführer: | Dr. Christian Grün, Dr. Alexander Holupirek, Michael Seiferle `-- Phone: 0049 7531 28 28 676, Fax: 0049 7531 20 05 22
Dirk,
Hello ! Sorry, the problem seems to be on BaseX side.
1- I have tried to use eclipse IDE setting the interpreter as a client. Doesn't work, it saturated my Java machine. 2- I have tried to sollicitate directly the BaseX client module through client:connect and client:query. Doesn't work, it saturated my Java machine. 3- Then I toggled totally the eclipse IDE : I opened a local server instance basexserver.bat, I opened a dos windows basexclient.bat, and tried to RUN the query file directly on the client console. Doesn't work, it still saturated my JAVA Machine. 4- I finally tested that the same query ran perfectly using the BaseX Gui.
My first penny guess is that this problem could be linked to this closed bug, that we encountered some 7 months ago : https://github.com/BaseXdb/basex/commit/5a926e0ea74b3e4a09f9a735ef7f19006413... . This impression is motivated because I have the same symptoms than for this bug : as I terminate the query, the eclipse console IDE freeze and I have to kill the BAseX server process to retrieve.
5- Thus I tried also to force non verbose mode using Basex client interpreter (setting -v options). Doesn't work, it still saturated my JAVA Machine.
Do we have any way to check if a query run in verbose mode ?
Cheeres,
Jean-Marc
2014-09-26 16:20 GMT+02:00 jean-marc Mercier jeanmarc.mercier@gmail.com:
Hi Dirk,
Ok, thanks for the trick, that should work. The only point is that it should be less user friendly for developing than using the eclipse IDE, since I will have to set up the BaseX repo first of running a query.
2014-09-26 12:11 GMT+02:00 Dirk Kirsten dk@basex.org:
Hi Jean-Marc,
I guess you set it up using the standalone version? Could you maybe test using the Client/Server set-up. You could then test with a plain BaseXClient and the eclipse environment. As they both should access the same BaseXServer, there shouldn't be any difference in execution.
But I don't know what XQDT is doing internally and I would pretty much guess the error is somewhere there (because plain BaseX seems to execute just fine). As the development seems to have pretty much stalled, I am not sure we can do much about that...
On 09/26/2014 11:54 AM, jean-marc Mercier wrote:
Hi Dirk,
Thanks for your answer. I am setting up eclipse indigo, following the step-up procedure http://docs.basex.org/wiki/Integrating_Eclipse. The
only
difference is that I set using -Xmx12G as Interpreter Arguments.
Cheers
2014-09-26 11:44 GMT+02:00 Dirk Kirsten dk@basex.org:
Hello Jean-Marc,
I am not sure how you are calling BaseX from within Eclipse, I
remember
there are different ways. Normally, I would expect it to simply
execute
basex, so there shouldn't be any difference. The query optimization
should
be done in both cases. However, it could be for example that your
BaseX
integration in eclipse uses XQJ, which could possibly introduce some performance bottlenecks.
So if you could shortly describe how you set up BaseX within Eclipse,
that
would be great.
Cheers, Dirk
On 09/24/2014 06:33 AM, jean-marc Mercier wrote:
Hi all,
Happy to write again in this mailing-list, hope you are all doing
well
guys
!
I am encountering a small problem running my old xquery modules. I
guess
that it is a classical issue, but I can't remember how to look into
BaseX
archive :(
Here is the description of the problem : I am running a query,
calling
module_of_mine.xqm, on two different environments.
The first one is eclipse, linked to the latest interpreter from BaseX80-20140923.183154 (but I tested also against older
interpreter). In
this way, this query saturated my 12 G Java machine after some
minutes.
Then I installed module_of_mine.xqm into BaseX repo, and ran the same
query
using basex gui with 4G heap size. The query ran perfectly in some 30
sec.
My first penny guess is that the built-in BaseX xquery profiler is
doing
some nice work there ? If so, is there a way to activate /
desactivate
this
profiler from eclipse or from some configuration files ?
Cheers,
Jean-Marc
-- Dirk Kirsten, BaseX GmbH, http://basex.org |-- Firmensitz: Blarerstrasse 56, 78462 Konstanz |-- Registergericht Freiburg, HRB: 708285, Geschäftsführer: | Dr. Christian Grün, Dr. Alexander Holupirek, Michael Seiferle `-- Phone: 0049 7531 28 28 676, Fax: 0049 7531 20 05 22
-- Dirk Kirsten, BaseX GmbH, http://basex.org |-- Firmensitz: Blarerstrasse 56, 78462 Konstanz |-- Registergericht Freiburg, HRB: 708285, Geschäftsführer: | Dr. Christian Grün, Dr. Alexander Holupirek, Michael Seiferle `-- Phone: 0049 7531 28 28 676, Fax: 0049 7531 20 05 22
Dirk,
Hello. I also tried to run the query on a Standalone BaseX version. It does not work neither, saturating my JAVA machine. Note that the verbose mode is off, here the result of INFO run on the standalone console :
BaseX 8.0 beta 5910246 [Standalone] Try help to get more information.
OPEN SOPHISIMPORTDB
Database 'SOPHISIMPORTDB' was opened in 244.02 ms.
help
Try HELP [...] to get info on a specific command.
ADD (TO [path]) [input] Add resource to database. ALTER [DATABASE|USER] [...] Alter database name or user password. CHECK [input] Open or create database. CLOSE Close current database. COPY [name] [newname] Copy database. CREATE [BACKUP|DATABASE|EVENT|INDEX|USER] [...] Create database, index, user, backup or event. CS [query] Run XQuery and set result as new context set. DELETE [path] Delete resources from database. DROP [BACKUP|DATABASE|EVENT|INDEX|USER] [...] Drop database, index, user, backup or event. EXIT Exit application. EXPORT [path] Export database to XML. FIND [keywords] Run a keyword query. FLUSH Flush database. GET [option] Show global option. GRANT [NONE|READ|WRITE|CREATE|ADMIN] (ON [database]) TO [user] Grant user permissions. HELP ([command]) Get help on BaseX commands. INFO ([DATABASE|INDEX|STORAGE]) Show information on current database. INSPECT Perform integrity checks. KILL [name] Kill user sessions. LIST ([name] ([path])) List databases or resources of database. OPEN [name] ([path]) Open database. OPTIMIZE (ALL) Optimize the database. PASSWORD ([password]) Change password. QUIT Exit application. RENAME [path] [newpath] Rename resources in database. REPLACE [path] [input] Replace resources in database. REPO [DELETE|INSTALL|LIST] Install, delete or list packages. RESTORE [name-(date)] Restore database. RETRIEVE [path] Retrieve raw data. RUN [path] Run query or command script. EXECUTE [input] Execute command script. SET [option] ([value]) Set global options. SHOW [BACKUPS|EVENTS|SESSIONS|USERS] Show server information. STORE (TO [path]) [input] Store raw data. TEST [path] Run XQUnit tests. XQUERY [query] Run XQuery.
INFO
General Information: Version: 8.0 beta 5910246 Used Memory: 126 MB
Global options: DBPATH: C:/Jiheme/informatique/basex8.0/data DEBUG: false EVENTPORT: 1985 GLOBALLOCK: false HOST: localhost HTTPLOCAL: false KEEPALIVE: 600 LANG: English LANGKEYS: false LOG: true LOGMSGMAXLEN: 1000 NONPROXYHOSTS: PARALLEL: 8 PASSWORD: PORT: 1984 PROXYHOST: PROXYPORT: 80 REPOPATH: C:/Jiheme/informatique/basex8.0/repo RESTXQPATH: SERVERHOST: SERVERPORT: 1984 STOPPORT: 8985 TIMEOUT: 30 USER: WEBPATH: C:/Jiheme/informatique/basex8.0/webapp
Local options ADDARCHIVES: true ADDCACHE: false ADDRAW: false ATTRINDEX: true AUTOFLUSH: true BINDINGS: CACHEQUERY: false CASESENS: false CATFILE: CHECKSTRINGS: true CHOP: true COMPPLAN: true CREATEFILTER: *.xml CREATEONLY: false CSVPARSER: DEFAULTDB: false DIACRITICS: false DOTCOMPACT: false DOTPLAN: false DTD: false EXPORTER: FORCECREATE: false FTINDEX: false FTINDEXSPLITSIZE: 0 HTMLPARSER: INDEXSPLITSIZE: 0 INLINELIMIT: 100 INTPARSE: false JSONPARSER: LANGUAGE: en LSERROR: 0 MAINMEM: false MAXCATS: 100 MAXHITS: -1 MAXLEN: 96 MAXSTAT: 30 MIXUPDATES: false PARSER: xml QUERYINFO: false QUERYPATH: RUNQUERY: true RUNS: 1 SERIALIZE: true SERIALIZER: SKIPCORRUPT: false STEMMING: false STOPWORDS: STRIPNS: false TAILCALLS: 256 TEXTINDEX: true TEXTPARSER: UPDINDEX: false WRITEBACK: false XINCLUDE: true XMLPLAN: false
2014-09-27 11:34 GMT+02:00 jean-marc Mercier jeanmarc.mercier@gmail.com:
Dirk,
Hello ! Sorry, the problem seems to be on BaseX side.
1- I have tried to use eclipse IDE setting the interpreter as a client. Doesn't work, it saturated my Java machine. 2- I have tried to sollicitate directly the BaseX client module through client:connect and client:query. Doesn't work, it saturated my Java machine. 3- Then I toggled totally the eclipse IDE : I opened a local server instance basexserver.bat, I opened a dos windows basexclient.bat, and tried to RUN the query file directly on the client console. Doesn't work, it still saturated my JAVA Machine. 4- I finally tested that the same query ran perfectly using the BaseX Gui.
My first penny guess is that this problem could be linked to this closed bug, that we encountered some 7 months ago : https://github.com/BaseXdb/basex/commit/5a926e0ea74b3e4a09f9a735ef7f19006413... . This impression is motivated because I have the same symptoms than for this bug : as I terminate the query, the eclipse console IDE freeze and I have to kill the BAseX server process to retrieve.
5- Thus I tried also to force non verbose mode using Basex client interpreter (setting -v options). Doesn't work, it still saturated my JAVA Machine.
Do we have any way to check if a query run in verbose mode ?
Cheeres,
Jean-Marc
2014-09-26 16:20 GMT+02:00 jean-marc Mercier jeanmarc.mercier@gmail.com:
Hi Dirk,
Ok, thanks for the trick, that should work. The only point is that it should be less user friendly for developing than using the eclipse IDE, since I will have to set up the BaseX repo first of running a query.
2014-09-26 12:11 GMT+02:00 Dirk Kirsten dk@basex.org:
Hi Jean-Marc,
I guess you set it up using the standalone version? Could you maybe test using the Client/Server set-up. You could then test with a plain BaseXClient and the eclipse environment. As they both should access the same BaseXServer, there shouldn't be any difference in execution.
But I don't know what XQDT is doing internally and I would pretty much guess the error is somewhere there (because plain BaseX seems to execute just fine). As the development seems to have pretty much stalled, I am not sure we can do much about that...
On 09/26/2014 11:54 AM, jean-marc Mercier wrote:
Hi Dirk,
Thanks for your answer. I am setting up eclipse indigo, following the step-up procedure http://docs.basex.org/wiki/Integrating_Eclipse. The
only
difference is that I set using -Xmx12G as Interpreter Arguments.
Cheers
2014-09-26 11:44 GMT+02:00 Dirk Kirsten dk@basex.org:
Hello Jean-Marc,
I am not sure how you are calling BaseX from within Eclipse, I
remember
there are different ways. Normally, I would expect it to simply
execute
basex, so there shouldn't be any difference. The query optimization
should
be done in both cases. However, it could be for example that your
BaseX
integration in eclipse uses XQJ, which could possibly introduce some performance bottlenecks.
So if you could shortly describe how you set up BaseX within
Eclipse, that
would be great.
Cheers, Dirk
On 09/24/2014 06:33 AM, jean-marc Mercier wrote:
Hi all,
Happy to write again in this mailing-list, hope you are all doing
well
guys
!
I am encountering a small problem running my old xquery modules. I
guess
that it is a classical issue, but I can't remember how to look into
BaseX
archive :(
Here is the description of the problem : I am running a query,
calling
module_of_mine.xqm, on two different environments.
The first one is eclipse, linked to the latest interpreter from BaseX80-20140923.183154 (but I tested also against older
interpreter). In
this way, this query saturated my 12 G Java machine after some
minutes.
Then I installed module_of_mine.xqm into BaseX repo, and ran the
same
query
using basex gui with 4G heap size. The query ran perfectly in some
30
sec.
My first penny guess is that the built-in BaseX xquery profiler is
doing
some nice work there ? If so, is there a way to activate /
desactivate
this
profiler from eclipse or from some configuration files ?
Cheers,
Jean-Marc
-- Dirk Kirsten, BaseX GmbH, http://basex.org |-- Firmensitz: Blarerstrasse 56, 78462 Konstanz |-- Registergericht Freiburg, HRB: 708285, Geschäftsführer: | Dr. Christian Grün, Dr. Alexander Holupirek, Michael Seiferle `-- Phone: 0049 7531 28 28 676, Fax: 0049 7531 20 05 22
-- Dirk Kirsten, BaseX GmbH, http://basex.org |-- Firmensitz: Blarerstrasse 56, 78462 Konstanz |-- Registergericht Freiburg, HRB: 708285, Geschäftsführer: | Dr. Christian Grün, Dr. Alexander Holupirek, Michael Seiferle `-- Phone: 0049 7531 28 28 676, Fax: 0049 7531 20 05 22
Hi Jean-Marc,
back again. Could you possibly send us the problematic query files (provided that these can be run without too much initialization overhead)?
Cheers, Christian
On Sat, Sep 27, 2014 at 11:34 AM, jean-marc Mercier jeanmarc.mercier@gmail.com wrote:
Dirk,
Hello ! Sorry, the problem seems to be on BaseX side.
1- I have tried to use eclipse IDE setting the interpreter as a client. Doesn't work, it saturated my Java machine. 2- I have tried to sollicitate directly the BaseX client module through client:connect and client:query. Doesn't work, it saturated my Java machine. 3- Then I toggled totally the eclipse IDE : I opened a local server instance basexserver.bat, I opened a dos windows basexclient.bat, and tried to RUN the query file directly on the client console. Doesn't work, it still saturated my JAVA Machine. 4- I finally tested that the same query ran perfectly using the BaseX Gui.
My first penny guess is that this problem could be linked to this closed bug, that we encountered some 7 months ago : https://github.com/BaseXdb/basex/commit/5a926e0ea74b3e4a09f9a735ef7f19006413... . This impression is motivated because I have the same symptoms than for this bug : as I terminate the query, the eclipse console IDE freeze and I have to kill the BAseX server process to retrieve.
5- Thus I tried also to force non verbose mode using Basex client interpreter (setting -v options). Doesn't work, it still saturated my JAVA Machine.
Do we have any way to check if a query run in verbose mode ?
Cheeres,
Jean-Marc
2014-09-26 16:20 GMT+02:00 jean-marc Mercier jeanmarc.mercier@gmail.com:
Hi Dirk,
Ok, thanks for the trick, that should work. The only point is that it should be less user friendly for developing than using the eclipse IDE, since I will have to set up the BaseX repo first of running a query.
2014-09-26 12:11 GMT+02:00 Dirk Kirsten dk@basex.org:
Hi Jean-Marc,
I guess you set it up using the standalone version? Could you maybe test using the Client/Server set-up. You could then test with a plain BaseXClient and the eclipse environment. As they both should access the same BaseXServer, there shouldn't be any difference in execution.
But I don't know what XQDT is doing internally and I would pretty much guess the error is somewhere there (because plain BaseX seems to execute just fine). As the development seems to have pretty much stalled, I am not sure we can do much about that...
On 09/26/2014 11:54 AM, jean-marc Mercier wrote:
Hi Dirk,
Thanks for your answer. I am setting up eclipse indigo, following the step-up procedure http://docs.basex.org/wiki/Integrating_Eclipse. The only difference is that I set using -Xmx12G as Interpreter Arguments.
Cheers
2014-09-26 11:44 GMT+02:00 Dirk Kirsten dk@basex.org:
Hello Jean-Marc,
I am not sure how you are calling BaseX from within Eclipse, I remember there are different ways. Normally, I would expect it to simply execute basex, so there shouldn't be any difference. The query optimization should be done in both cases. However, it could be for example that your BaseX integration in eclipse uses XQJ, which could possibly introduce some performance bottlenecks.
So if you could shortly describe how you set up BaseX within Eclipse, that would be great.
Cheers, Dirk
On 09/24/2014 06:33 AM, jean-marc Mercier wrote:
Hi all,
Happy to write again in this mailing-list, hope you are all doing well
guys
!
I am encountering a small problem running my old xquery modules. I guess that it is a classical issue, but I can't remember how to look into BaseX archive :(
Here is the description of the problem : I am running a query, calling module_of_mine.xqm, on two different environments.
The first one is eclipse, linked to the latest interpreter from BaseX80-20140923.183154 (but I tested also against older interpreter). In this way, this query saturated my 12 G Java machine after some minutes.
Then I installed module_of_mine.xqm into BaseX repo, and ran the same
query
using basex gui with 4G heap size. The query ran perfectly in some 30
sec.
My first penny guess is that the built-in BaseX xquery profiler is doing some nice work there ? If so, is there a way to activate / desactivate
this
profiler from eclipse or from some configuration files ?
Cheers,
Jean-Marc
-- Dirk Kirsten, BaseX GmbH, http://basex.org |-- Firmensitz: Blarerstrasse 56, 78462 Konstanz |-- Registergericht Freiburg, HRB: 708285, Geschäftsführer: | Dr. Christian Grün, Dr. Alexander Holupirek, Michael Seiferle `-- Phone: 0049 7531 28 28 676, Fax: 0049 7531 20 05 22
-- Dirk Kirsten, BaseX GmbH, http://basex.org |-- Firmensitz: Blarerstrasse 56, 78462 Konstanz |-- Registergericht Freiburg, HRB: 708285, Geschäftsführer: | Dr. Christian Grün, Dr. Alexander Holupirek, Michael Seiferle `-- Phone: 0049 7531 28 28 676, Fax: 0049 7531 20 05 22
Hi christian,
Hope your holidays went finely !
I can send you the XQUERY files, that consists mainly the graph database V0.0 open sourced files. However, you won't be able to reproduce the issue without a database sample. Actually, I can't send you my database, that has been built with my client data, thus containing confidential informations.
Don't you have another idea to go ahead with this issue ? If not, I'll have to build a database from public data, like a wikipedia dump, but it will take some time.
Cheers
2014-10-06 13:53 GMT+02:00 Christian Grün christian.gruen@gmail.com:
Hi Jean-Marc,
back again. Could you possibly send us the problematic query files (provided that these can be run without too much initialization overhead)?
Cheers, Christian
On Sat, Sep 27, 2014 at 11:34 AM, jean-marc Mercier jeanmarc.mercier@gmail.com wrote:
Dirk,
Hello ! Sorry, the problem seems to be on BaseX side.
1- I have tried to use eclipse IDE setting the interpreter as a client. Doesn't work, it saturated my Java machine. 2- I have tried to sollicitate directly the BaseX client module through client:connect and client:query. Doesn't work, it saturated my Java
machine.
3- Then I toggled totally the eclipse IDE : I opened a local server
instance
basexserver.bat, I opened a dos windows basexclient.bat, and tried to RUN the query file directly on the client console. Doesn't work, it still saturated my JAVA Machine. 4- I finally tested that the same query ran perfectly using the BaseX
Gui.
My first penny guess is that this problem could be linked to this closed bug, that we encountered some 7 months ago :
https://github.com/BaseXdb/basex/commit/5a926e0ea74b3e4a09f9a735ef7f19006413...
. This impression is motivated because I have the same symptoms than for this bug : as I terminate the query, the eclipse console IDE freeze and I have to kill the BAseX server process to retrieve.
5- Thus I tried also to force non verbose mode using Basex client interpreter (setting -v options). Doesn't work, it still saturated my
JAVA
Machine.
Do we have any way to check if a query run in verbose mode ?
Cheeres,
Jean-Marc
2014-09-26 16:20 GMT+02:00 jean-marc Mercier <jeanmarc.mercier@gmail.com :
Hi Dirk,
Ok, thanks for the trick, that should work. The only point is that it should be less user friendly for developing than using the eclipse IDE, since I will have to set up the BaseX repo first of running a query.
2014-09-26 12:11 GMT+02:00 Dirk Kirsten dk@basex.org:
Hi Jean-Marc,
I guess you set it up using the standalone version? Could you maybe
test
using the Client/Server set-up. You could then test with a plain
BaseXClient
and the eclipse environment. As they both should access the same BaseXServer, there shouldn't be any difference in execution.
But I don't know what XQDT is doing internally and I would pretty much guess the error is somewhere there (because plain BaseX seems to
execute
just fine). As the development seems to have pretty much stalled, I am
not
sure we can do much about that...
On 09/26/2014 11:54 AM, jean-marc Mercier wrote:
Hi Dirk,
Thanks for your answer. I am setting up eclipse indigo, following the step-up procedure http://docs.basex.org/wiki/Integrating_Eclipse.
The
only difference is that I set using -Xmx12G as Interpreter Arguments.
Cheers
2014-09-26 11:44 GMT+02:00 Dirk Kirsten dk@basex.org:
Hello Jean-Marc,
I am not sure how you are calling BaseX from within Eclipse, I remember there are different ways. Normally, I would expect it to simply execute basex, so there shouldn't be any difference. The query optimization should be done in both cases. However, it could be for example that your BaseX integration in eclipse uses XQJ, which could possibly introduce
some
performance bottlenecks.
So if you could shortly describe how you set up BaseX within
Eclipse,
that would be great.
Cheers, Dirk
On 09/24/2014 06:33 AM, jean-marc Mercier wrote: > Hi all, > > Happy to write again in this mailing-list, hope you are all doing > well guys > ! > > I am encountering a small problem running my old xquery modules. I > guess > that it is a classical issue, but I can't remember how to look
into
> BaseX > archive :( > > Here is the description of the problem : I am running a query, > calling > module_of_mine.xqm, on two different environments. > > The first one is eclipse, linked to the latest interpreter from > BaseX80-20140923.183154 (but I tested also against older > interpreter). In > this way, this query saturated my 12 G Java machine after some > minutes. > > Then I installed module_of_mine.xqm into BaseX repo, and ran the > same query > using basex gui with 4G heap size. The query ran perfectly in some > 30 sec. > > My first penny guess is that the built-in BaseX xquery profiler is > doing > some nice work there ? If so, is there a way to activate / > desactivate this > profiler from eclipse or from some configuration files ? > > Cheers, > > Jean-Marc >
-- Dirk Kirsten, BaseX GmbH, http://basex.org |-- Firmensitz: Blarerstrasse 56, 78462 Konstanz |-- Registergericht Freiburg, HRB: 708285, Geschäftsführer: | Dr. Christian Grün, Dr. Alexander Holupirek, Michael Seiferle `-- Phone: 0049 7531 28 28 676, Fax: 0049 7531 20 05 22
-- Dirk Kirsten, BaseX GmbH, http://basex.org |-- Firmensitz: Blarerstrasse 56, 78462 Konstanz |-- Registergericht Freiburg, HRB: 708285, Geschäftsführer: | Dr. Christian Grün, Dr. Alexander Holupirek, Michael Seiferle `-- Phone: 0049 7531 28 28 676, Fax: 0049 7531 20 05 22
basex-talk@mailman.uni-konstanz.de
-
Christian Grün -
Dirk Kirsten -
jean-marc Mercier