Menu
▾
▴
aplaws-devel — Developer Mailing List. Forum postings are forwarded to this list.
You can subscribe to this list here.
| 2003 |
Jan
|
Feb
|
Mar
|
Apr
|
May
|
Jun
|
Jul
|
Aug
|
Sep
|
Oct
|
Nov
|
Dec
(10) |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 2004 |
Jan
|
Feb
(13) |
Mar
(15) |
Apr
(29) |
May
(28) |
Jun
(111) |
Jul
(185) |
Aug
(366) |
Sep
(121) |
Oct
(73) |
Nov
(57) |
Dec
(23) |
| 2005 |
Jan
(30) |
Feb
(49) |
Mar
(51) |
Apr
(47) |
May
(95) |
Jun
(74) |
Jul
(62) |
Aug
(61) |
Sep
(46) |
Oct
(73) |
Nov
(111) |
Dec
(59) |
| 2006 |
Jan
(114) |
Feb
(34) |
Mar
(47) |
Apr
(49) |
May
(106) |
Jun
(47) |
Jul
(78) |
Aug
(31) |
Sep
(35) |
Oct
(39) |
Nov
(63) |
Dec
(17) |
| 2007 |
Jan
(40) |
Feb
(32) |
Mar
(17) |
Apr
(15) |
May
(28) |
Jun
(20) |
Jul
(80) |
Aug
(83) |
Sep
(52) |
Oct
(26) |
Nov
(6) |
Dec
(9) |
| 2008 |
Jan
(22) |
Feb
(11) |
Mar
(45) |
Apr
(5) |
May
(8) |
Jun
|
Jul
(16) |
Aug
(5) |
Sep
(3) |
Oct
(4) |
Nov
(14) |
Dec
(3) |
| 2009 |
Jan
(25) |
Feb
(46) |
Mar
(17) |
Apr
(8) |
May
(74) |
Jun
(48) |
Jul
(11) |
Aug
(9) |
Sep
|
Oct
|
Nov
(2) |
Dec
(2) |
|
From: Aplaws D. L. <apl...@li...> - 2009-12-09 10:17:01
|
Hi Tim, Am Dienstag, den 24.11.2009, 14:02 +0800 schrieb Aplaws Developer List: > we have recently upgraded our local instance of APLAWS+ 1.0.5 to > enable streaming of binary assets to and from the database using SQL > Blobs. I think it is very advantageous to update APLAWS to use BLOBs for binary data as images, binary files, etc.! This should be true for Oracle which seems to have a sophisticated implementation of BLOB. How do you judge the Postgres implementation? I read several documentation which lets me doubt a little bit. It's implementation looks to me a kind of lukewarm, e.g. in terms of security, managing the OID and deleting the last instance, see http://lab.lonerunners.net/blog/binary-data-fetching-through-sqli http://jdbc.postgresql.org/documentation/84/binary-data.html I upcomming 8.5 the project will introduce a new hex format for bytea to avoid all the excaping requirements, see http://developer.postgresql.org/pgdocs/postgres/datatype-binary.html (instead of polishing the BLOB implementation) Or did I misread that stuff? > The bulk of our testing has been with Oracle SE/XE 10.2.0, Profiling > has shown a substantial reduction to memory usage in the JVM when > uploading and downloading large binary assets. However, we have only > undertaken limited testing with PostGres 8.4, so results may vary. Does "limited testing" refer to performance or to basic functions? In the latter case it may be better you send it as a patch so others could test it with postgres without disturbing the trunk. Otherwise you should commit it. By the way, three other related issues: a) Could you resolve the size limitation for filestorage items by using streaming? See https://fedorahosted.org/aplaws/ticket/30 We closed it wontfix because of the amount of work to switch to streaming. b) Could you work with Postgres 8.4 out of the box? If I try to use 8.4 I get an error (see attached listing) when I try to open the default start page (load-bundle does work). c) We had a discussion of the forum some times ago regarding dead lock situations with postgres and a possible patch created by you and your colleagues ( https://sourceforge.net/projects/aplaws/forums/forum/368401/topic/1721796 ) Peter >----------------- Error listing --------------------------------------< 2009-12-06 23:14:41,064 [0:0:1] ERROR rdbms.RDBMSEngine - select t5.privilege as "this.privilege", t5.inherited_p as "this.isInherited", t5.grantee_id as "this.granteeID", t5.user_p as "this.granteeIsUser", t5.name as "this.granteeName", t5.primary_email as "this.granteeEmail", t5.name_id as "this.granteePersonName.id", t5.given_name as "this.granteePersonName.givenN7", t5.family_name as "this.granteePersonName.family8" from ( select acs_permissions.privilege, CASE WHEN acs_permissions.object_id = ? THEN 0 ELSE 1 END as inherited_p, acs_permissions.grantee_id, parties.user_p, parties.name, parties.primary_email, parties.given_name, parties.family_name, parties.name_id from (select p.privilege, p.object_id, p.grantee_id from acs_permissions p, dnm_object_1_granted_context dogc, dnm_granted_context dgc where dogc.pd_object_id = ? and dogc.pd_context_id = dgc.pd_object_id and p.object_id = dgc.pd_context_id) acs_permissions, (select groups.group_id as party_id, 0 as user_p, groups.name as name, parties.primary_email as primary_email, '' as given_name, '' as family_name, 0 as name_id from groups, parties where groups.group_id = parties.party_id) parties where acs_permissions.grantee_id = parties.party_id UNION ALL select acs_permissions.privilege, CASE WHEN acs_permissions.object_id = ? THEN 0 ELSE 1 END as inherited_p, acs_permissions.grantee_id, parties.user_p, parties.name, parties.primary_email, parties.given_name, parties.family_name, parties.name_id from (select p.privilege, p.object_id, p.grantee_id from acs_permissions p, dnm_object_1_granted_context dogc, dnm_granted_context dgc where dogc.pd_object_id = ? and dogc.pd_context_id = dgc.pd_object_id and p.object_id = dgc.pd_context_id) acs_permissions, (select users.user_id as party_id, 1 as user_p, '' as name, parties.primary_email as primary_email, person_names.given_name as given_name, person_names.family_name as family_name, users.name_id from users, parties, person_names where users.user_id = parties.party_id and users.name_id = person_names.name_id) parties where acs_permissions.grantee_id = parties.party_id) t5 where t5.user_p = '1' order by t5.grantee_id org.postgresql.util.PSQLException: FEHLER: failed to find conversion function from unknown to character varying at org.postgresql.core.v3.QueryExecutorImpl.receiveErrorResponse(QueryExecutorImpl.java:1548) at org.postgresql.core.v3.QueryExecutorImpl.processResults(QueryExecutorImpl.java:1316) at org.postgresql.core.v3.QueryExecutorImpl.execute(QueryExecutorImpl.java:191) at org.postgresql.jdbc2.AbstractJdbc2Statement.execute(AbstractJdbc2Statement.java:452) at org.postgresql.jdbc2.AbstractJdbc2Statement.executeWithFlags(AbstractJdbc2Statement.java:351) at org.postgresql.jdbc2.AbstractJdbc2Statement.execute(AbstractJdbc2Statement.java:344) at com.redhat.persistence.engine.rdbms.RDBMSEngine.execute(RDBMSEngine.java:525) at com.redhat.persistence.engine.rdbms.RDBMSEngine.execute(RDBMSEngine.java:451) at com.redhat.persistence.engine.rdbms.RDBMSEngine.execute(RDBMSEngine.java:283) at com.redhat.persistence.Cursor.execute(Cursor.java:141) at com.redhat.persistence.Cursor.next(Cursor.java:125) at com.arsdigita.persistence.DataQueryImpl.next(DataQueryImpl.java:663) at com.arsdigita.domain.DomainQuery.next(DomainQuery.java:97) at com.arsdigita.kernel.security.SecurityConfig.getSystemAdministratorEmailAddress(SecurityConfig.java:179) at com.arsdigita.kernel.security.SecurityConfig.getAdminContactEmail(SecurityConfig.java:168) at com.arsdigita.ui.SiteBanner.generateXML(SiteBanner.java:39) at com.arsdigita.bebop.SimpleContainer.generateChildrenXML(SimpleContainer.java:246) at com.arsdigita.bebop.SimpleContainer.generateXML(SimpleContainer.java:263) at com.arsdigita.bebop.Page.generateXML(Page.java:633) at com.arsdigita.bebop.jsp.DefinePage.doEndTag(DefinePage.java:292) at org.apache.jsp.packages.navigation.templates.shp_002ddefault_jsp._jspService(shp_002ddefault_jsp.java:212) -- Dr. Peter Boy University of Bremen Center for Social Policy Research Department for Research Methods, Statistical Analysis, and IT Management pb...@ze... phone: +49 421 218 4374 / 4362 Web: http://www.zes.uni-bremen.de |
|
From: Aplaws D. L. <apl...@li...> - 2009-12-07 10:02:03
|
Hi Tim, Am Dienstag, den 24.11.2009, 13:40 +0800 schrieb Aplaws Developer List: > we have upgraded our local instance of APLAWS+ 1.0.5 to use XSLT 2.0, > based on Saxon HE 9.2.0.2. As to my knowledge Saxon is the only software which implements the 2.0 standard. All the apache projects are still on 1.0. Is there any ongoing work to implement 2.0? Some questions: a) Saxon HE seems to be governed by some kind of open source licence. Do you know if it is compatible with LGPL? b) Could we distibute it? c) And could we distribute it with version 6.5.5 in parallel (without path / access conflicts in the lib dir)? Peter |
|
From: Aplaws D. L. <apl...@li...> - 2009-11-24 06:02:40
|
Ladies and Gents, we have recently upgraded our local instance of APLAWS+ 1.0.5 to enable streaming of binary assets to and from the database using SQL Blobs. In the current release of APLAWS, all large binary data sets are represented as SQL Blobs in the database and transferred to and from the database as byte arrays. That is, the "content" value of binary assets is of type "byte[]". Our changes enable streaming of binary data for BinaryAsset types using the InputStream and OutputStream available with an SQL Blob instance. That is, the "content" value of binary assets has been changed from type "byte[]" to type "java.sql.Blob". Only BinaryAsset types have changed. All other asset types remain unchanged and continue to transfer binary data using byte arrays. The bulk of our testing has been with Oracle SE/XE 10.2.0, Profiling has shown a substantial reduction to memory usage in the JVM when uploading and downloading large binary assets. However, we have only undertaken limited testing with PostGres 8.4, so results may vary. If other APLAWS community members are interested in these changes, I can post a more comprehensive outline of the changes and/or a send a patch set to the newsgroup. Regards, Tim Telcik -- +++ Tim Telcik tim...@pe... |
|
From: Aplaws D. L. <apl...@li...> - 2009-11-24 05:41:16
|
Ladies and Gents, we have upgraded our local instance of APLAWS+ 1.0.5 to use XSLT 2.0, based on Saxon HE 9.2.0.2. For reference, APLAWS is currently using XSLT 1.0, based on Saxon 6.5.5. Do any other APLAWS community members foresee a need to upgrade to XSLT 2.0 ? If so, I can upload the changes as a patch release. Regards, Tim Telcik -- +++ Tim Telcik tim...@pe... |
|
From: Aplaws D. L. <apl...@li...> - 2009-08-12 16:07:11
|
Read and respond to this message at: https://sourceforge.net/forum/message.php?msg_id=7557814 By: bprucha How is JAAS going to be handled in the JEE deployment? APLAWS uses javax.security.auth.login.Configuration.setConfiguration(getLoginConfig()) to set the JAAS configuration. In JBoss this overides it's way of configuring JAAS and breaks the ability to login to the admin and jmx consoles. The way to package an EAR and configure it in JBoss is described pretty well at http://stuffthathappens.com/blog/2008/05/16/writing-a-custom-jaas-loginmodule/ under the Packaging heading. This is JBoss specific though. I don't know how over app servers handle this. Anyone have any suggestions? ______________________________________________________________________ You are receiving this email because you elected to monitor this forum. To stop monitoring this forum, login to SourceForge.net and visit: https://sourceforge.net/forum/unmonitor.php?forum_id=368401 |
|
From: Aplaws D. L. <apl...@li...> - 2009-08-10 08:04:42
|
Read and respond to this message at: https://sourceforge.net/forum/message.php?msg_id=7553176 By: alan_permeance We've fixed our deadlocking issues (thanks Tim Carpenter). Oracle issues a SHARE lock on child tables when updating parents when there is no index on the FK column. If two transactions have a SHARE lock they will deadlock when DML starts to happen. Once we ensured that all Foreign keys had a single column index the dead locks went away. We haven't tested this on Postgres. We have updated the oracle-se-create.sql and postgres-create.sql to ensure the indexes are created during a load-bundle and we commit them soonish. Reference: http://download.oracle.com/docs/cd/B14117_01/appdev.101/b10795/adfns_sq.htm#1024 837 http://www.oracle-base.com/articles/9i/SQLNewFeatures9i.php#ShareLocksOnUnindexedFKs ______________________________________________________________________ You are receiving this email because you elected to monitor this forum. To stop monitoring this forum, login to SourceForge.net and visit: https://sourceforge.net/forum/unmonitor.php?forum_id=368401 |
|
From: Aplaws D. L. <apl...@li...> - 2009-08-10 06:49:31
|
Read and respond to this message at: https://sourceforge.net/forum/message.php?msg_id=7553065 By: pboy Chris provided a patch for bulk publishing of themes: 1822871: Bulk Theme Publish ( https://sourceforge.net/tracker/?func=detail&aid=1822871&group_id=71505&atid=531527 ) He didn't incorporated it into trunk, unlike most of the other patches he provided via tracker. Unfortunateley I can not get in contact with him. Could someone check it and provide advice, weather to integrate it into trunk? Peter ______________________________________________________________________ You are receiving this email because you elected to monitor this forum. To stop monitoring this forum, login to SourceForge.net and visit: https://sourceforge.net/forum/unmonitor.php?forum_id=368401 |
|
From: Aplaws D. L. <apl...@li...> - 2009-08-10 06:31:13
|
Read and respond to this message at: https://sourceforge.net/forum/message.php?msg_id=7553039 By: pboy Hi Terry, in 2007 you provided 2 patches: 1831466: Selecting section of mparticle as Related Link causes NPE 1828052 : Poller causes PooledConnectionSource to grow At least the first one is not in trunk (didn't check the second). Could you check weather they are still relevant and weather it is favourable to include them? I would like to do a little bit housekeeping :-) Thanks Peter ______________________________________________________________________ You are receiving this email because you elected to monitor this forum. To stop monitoring this forum, login to SourceForge.net and visit: https://sourceforge.net/forum/unmonitor.php?forum_id=368401 |
|
From: Aplaws D. L. <apl...@li...> - 2009-08-07 22:23:19
|
Read and respond to this message at: https://sourceforge.net/forum/message.php?msg_id=7550536 By: shawnlane We are using Postgres 8. something. It can happen when people try to load very large files too. ______________________________________________________________________ You are receiving this email because you elected to monitor this forum. To stop monitoring this forum, login to SourceForge.net and visit: https://sourceforge.net/forum/unmonitor.php?forum_id=368401 |
|
From: Aplaws D. L. <apl...@li...> - 2009-08-07 22:16:52
|
Read and respond to this message at: https://sourceforge.net/forum/message.php?msg_id=7550534 By: pboy Terry, I'm just working on an more extensive portal administration guide, play a bit around with portal and consider to use it for our site. 2 questions a) what is exactly the difference between navigation directory portlet and the content directory portlet? Both provide a view of the navigation tree. Perhaps it is a dump question, but I'm unfamiliar with all the LAW categorization machinery and may not see the obvious. - Under which circumstances you should use one or the other porftlet? - is navigation directory perhaps a superset of content directory so the latter could be removed? b) is it possible to relocate navigation directory portlet to c.a.london.navigation.portlet, where some other navigation related portlets reside? We would like to reduce horizontal dependencies and de-couple modules. Portal should be installable independently from Navigation and vice versa. Thanks Peter ______________________________________________________________________ You are receiving this email because you elected to monitor this forum. To stop monitoring this forum, login to SourceForge.net and visit: https://sourceforge.net/forum/unmonitor.php?forum_id=368401 |
|
From: Aplaws D. L. <apl...@li...> - 2009-08-07 22:05:08
|
Read and respond to this message at: https://sourceforge.net/forum/message.php?msg_id=7550523 By: pboy Hm, we seem to have the same or a similiar problem here. From time to time the system is very very slow and it is practically impossible to do any work in content-center. The public site shows a very slow performance as well. The prblem vanishes after some time. We are using postgresql, so the patch provided by Chris doesn't affect us. What db are you using? Peter ______________________________________________________________________ You are receiving this email because you elected to monitor this forum. To stop monitoring this forum, login to SourceForge.net and visit: https://sourceforge.net/forum/unmonitor.php?forum_id=368401 |
|
From: Aplaws D. L. <apl...@li...> - 2009-08-05 06:37:55
|
Read and respond to this message at: https://sourceforge.net/forum/message.php?msg_id=7545274 By: shawnlane We have noticed the deadlocks. Database performance drops off, then you look in the database and clear the deadlocks. We have never got to the bottom of the actual cause, but we do notice it when people are editing and publishing. I suppose in normal terms the software should not allow deadlocks, so it is a real bug.. our workaround is to check the database of deadlocks and delete these when they are found. ______________________________________________________________________ You are receiving this email because you elected to monitor this forum. To stop monitoring this forum, login to SourceForge.net and visit: https://sourceforge.net/forum/unmonitor.php?forum_id=368401 |
|
From: Aplaws D. L. <apl...@li...> - 2009-08-05 03:51:13
|
Read and respond to this message at: https://sourceforge.net/forum/message.php?msg_id=7545130 By: alan_permeance Hi all, We are currently having the same deadlocking issues. It occurs in all of our environments regardless of the size of the DB. Its even happening on the latest fedora trunk. We've noticed that when republishing two different items that do not have any related items all is ok. However as soon as one item has references to another (related link, metadata etc) it deadlocks every time. I'm not convinced that the deadlock is related to content Items as such. Being to think that maybe the persistence layer handling of associations is not fully concurrent? Has any one out there experience this, have any suggestions? Cheers Al ______________________________________________________________________ You are receiving this email because you elected to monitor this forum. To stop monitoring this forum, login to SourceForge.net and visit: https://sourceforge.net/forum/unmonitor.php?forum_id=368401 |
|
From: Aplaws D. L. <apl...@li...> - 2009-07-28 09:29:25
|
Read and respond to this message at: https://sourceforge.net/forum/message.php?msg_id=7531219 By: pboy I started to update our sorceforge site. APLAWS+ release 1.0.4 is now available here, including the long outstanding release notes. Finishing the 1.0.4 release notes I started work for release notes 1.0.5. I included less than half of the modifications and it is already 25 pages long! There are so many new features, enhancements, and bug fixes - let us get it out the dooor! You may have a look at it: https://fedorahosted.org/aplaws/ticket/35 Peter ______________________________________________________________________ You are receiving this email because you elected to monitor this forum. To stop monitoring this forum, login to SourceForge.net and visit: https://sourceforge.net/forum/unmonitor.php?forum_id=368401 |
|
From: Aplaws D. L. <apl...@li...> - 2009-07-21 17:23:08
|
Read and respond to this message at: https://sourceforge.net/forum/message.php?msg_id=7519350 By: pfosterlbc I was wondering whether someone might be able to help with an issue I have encountered with our Aplaws installations? This concerns form submitted URLs, for example with the `quick links' URL one is able to create in /ccm/navigation/admin, or the related links associated with content items in /ccm/content/admin. The issue appears to involve non http URLs -- when I attempt to submit an https URL on the live server, the /ccm/content/admin `related links' form complains that the URL is invalid. For the quick links form, the URL is submitted successfully, however appears as http://https://test.com in the edit dialogue. Therefore, I presume that in both cases, the URL pattern is not being recognised correctly. Are there any obvious configuration options which affect the URL processing behaviour and might be related to the observed issue? So far, I have been comparing ccm get configurations between two servers, as the issue is perplexingly not present on our staging environment. A comparison of installed packages and ccm parameters reveals that the recorded aplaws module versions are identical and that there is no striking difference between the two configurations. Best wishes, Peter ______________________________________________________________________ You are receiving this email because you elected to monitor this forum. To stop monitoring this forum, login to SourceForge.net and visit: https://sourceforge.net/forum/unmonitor.php?forum_id=243715 |
|
From: Aplaws D. L. <apl...@li...> - 2009-07-09 09:47:25
|
Read and respond to this message at: https://sourceforge.net/forum/message.php?msg_id=7490138 By: pboy Hi Terry, Steve Pugh did some work on maven and created a set of pom files which basically work. I'll sent them to you via PM later. It's on my todo list for a while, but I had to do too much other thins. Would be fine if you could spend some work on it. Peter ______________________________________________________________________ You are receiving this email because you elected to monitor this forum. To stop monitoring this forum, login to SourceForge.net and visit: https://sourceforge.net/forum/unmonitor.php?forum_id=368401 |
|
From: Aplaws D. L. <apl...@li...> - 2009-07-09 08:55:53
|
Read and respond to this message at: https://sourceforge.net/forum/message.php?msg_id=7490040 By: terry_permeance A while back someone mentioned they had looked at maven. I've had a bit of a play and it looks reasonably straight forward. I've got ccm-core compiling but have yet to generate DDL or get the tests passing. I don't want to reinvent the wheel if someone else has already done the migration. Anybody got maven scripts to contribute? Thanks, Terry ______________________________________________________________________ You are receiving this email because you elected to monitor this forum. To stop monitoring this forum, login to SourceForge.net and visit: https://sourceforge.net/forum/unmonitor.php?forum_id=368401 |
|
From: Aplaws D. L. <apl...@li...> - 2009-07-09 06:56:26
|
Read and respond to this message at: https://sourceforge.net/forum/message.php?msg_id=7489682 By: terry_permeance Done. See r1953 NEW USAGE: ccm set com.arsdigita.cms.contenttypes.glossaryitem.definition.editor=text (default) ccm set com.arsdigita.cms.contenttypes.glossaryitem.definition.editor=wysiwyg ______________________________________________________________________ You are receiving this email because you elected to monitor this forum. To stop monitoring this forum, login to SourceForge.net and visit: https://sourceforge.net/forum/unmonitor.php?forum_id=368401 |
|
From: Aplaws D. L. <apl...@li...> - 2009-07-07 01:23:26
|
Read and respond to this message at: https://sourceforge.net/forum/message.php?msg_id=7483820 By: terry_permeance Done. See https://fedorahosted.org/aplaws/changeset/1952 ______________________________________________________________________ You are receiving this email because you elected to monitor this forum. To stop monitoring this forum, login to SourceForge.net and visit: https://sourceforge.net/forum/unmonitor.php?forum_id=368401 |
|
From: Aplaws D. L. <apl...@li...> - 2009-07-06 13:49:10
|
Read and respond to this message at: https://sourceforge.net/forum/message.php?msg_id=7482621 By: pboy Hi Terrry, I would prefer to wait, until we have a clear plan for the user installation tool and at least a tentative implementation. And we should have a tool to keep the dist dir uptodate, something like post commit hook or a nightly build. But I think we should create such kind of directory and use it to create the user distribution and to add it to devel environment so you need not compile all modules each time. Peter ______________________________________________________________________ You are receiving this email because you elected to monitor this forum. To stop monitoring this forum, login to SourceForge.net and visit: https://sourceforge.net/forum/unmonitor.php?forum_id=368401 |
|
From: Aplaws D. L. <apl...@li...> - 2009-07-06 13:32:19
|
Read and respond to this message at: https://sourceforge.net/forum/message.php?msg_id=7482624 By: pboy Yes, it is old stuff and not needed anymore. Would you remove it? Peter ______________________________________________________________________ You are receiving this email because you elected to monitor this forum. To stop monitoring this forum, login to SourceForge.net and visit: https://sourceforge.net/forum/unmonitor.php?forum_id=368401 |
|
From: Aplaws D. L. <apl...@li...> - 2009-07-06 13:09:45
|
Read and respond to this message at: https://sourceforge.net/forum/message.php?msg_id=7482589 By: terry_permeance Can we create the war and package in a sub-folder called "dist"? This seems to be standard practice or is it just maven? ______________________________________________________________________ You are receiving this email because you elected to monitor this forum. To stop monitoring this forum, login to SourceForge.net and visit: https://sourceforge.net/forum/unmonitor.php?forum_id=368401 |
|
From: Aplaws D. L. <apl...@li...> - 2009-07-06 13:03:53
|
Read and respond to this message at: https://sourceforge.net/forum/message.php?msg_id=7482575 By: terry_permeance Can this be removed from README.txt. It doesn't seem to be needed? As a temporary measure: copy ${CATALINA_HOME}/lib/system/ccm-core* to ${JAVA_HOME}/jre/lib/ext otherwise you will get a malformedURL exception during server startup Regards, Terry ______________________________________________________________________ You are receiving this email because you elected to monitor this forum. To stop monitoring this forum, login to SourceForge.net and visit: https://sourceforge.net/forum/unmonitor.php?forum_id=368401 |
|
From: Aplaws D. L. <apl...@li...> - 2009-07-02 13:02:41
|
Read and respond to this message at: https://sourceforge.net/forum/message.php?msg_id=7475944 By: bprucha > Is it really true, that each module should be able to run in its own context? I don't see modules necessarily as being an extension to the existing functionality. A module can provide completely new and independent functionality. Developers should be able to use the core as a lower level framework for their own applications. This is what I use the code for. I tap into the user/group/permission functionality and dispatcher and templating mechanism to create multiple themable instances of my applications. Beyond that, the software I've developed stands on its own. Having each module in it's own context also makes it truly modular. You don't have to merge the moduals web.xml in with the core web.xml. This also allows you to undeploy/redeploy individual moduals without having to undeploy/redeploy the entire system. Those are the main benefits I see in having modules in their own context. I don't think each module has to be in its own context but when a module provides new application services that simply rely on the core for lower level API it's nice of have that capability. > a theme is very specific to a web application, no one wants to mix themes between several web applications. Why not? In an enterprise environment you want all of you applications to have the same look and feel. > As to my knowledge version 5 of APLAWS was a JEE application, were all modules were installed into one > context (ccm). Do you know, WHY that had been changed? I suppose I never understood this design decision. I'm not sure why. The code used the c.a.sitenode.SiteNodeDispatcher to locate the package dispatcher to handle the request. That was deprecated in favor of the c.a.web.DispatcherServlet which tries to forward the request to a servlet defined in c.a.web.Application. The new dispatcher mechanism is more flexible and allows the application to reside in a different context. My guess is that it makes the moduals more plug and play. > Just another topic: If each module is dependent from resources in core, I suppose it is an example of > (very) strong coupling. So it should not divided into different modules (at least according to theory :-) ) I see core as being a lower level web framework that should be in a shared jar. I think much of whats in their can/should be replaced with other competing technologies that are maintained as independent projects. What should stay are the things that make APLAWS+ truly unique like the way it handles users/groups/permissions and dispatches modules in a way that facilitates multiple instances of an application. ______________________________________________________________________ You are receiving this email because you elected to monitor this forum. To stop monitoring this forum, login to SourceForge.net and visit: https://sourceforge.net/forum/unmonitor.php?forum_id=368401 |
|
From: Aplaws D. L. <apl...@li...> - 2009-06-30 06:35:13
|
Read and respond to this message at: https://sourceforge.net/forum/message.php?msg_id=7471028 By: pboy Terry, your proposal is much better and we should do that way. Peter ______________________________________________________________________ You are receiving this email because you elected to monitor this forum. To stop monitoring this forum, login to SourceForge.net and visit: https://sourceforge.net/forum/unmonitor.php?forum_id=368401 |
7 messages has been excluded from this view by a project administrator.
