This Bugzilla instance is a read-only archive of historic NetBeans bug reports. To report a bug in NetBeans please follow the project's instructions for reporting issues.
Summary: | Deploying and running an application that uses table generation and a SEQUENCE table for PK generation gives phantom deployment failures | ||
---|---|---|---|
Product: | serverplugins | Reporter: | ievans <ievans> |
Component: | GlassFish | Assignee: | Vince Kraemer <vkraemer> |
Status: | VERIFIED FIXED | ||
Severity: | normal | CC: | musilt2, phejl, pjiricka, vkraemer |
Priority: | P1 | ||
Version: | 7.0 | ||
Hardware: | All | ||
OS: | All | ||
Issue Type: | DEFECT | Exception Reporter: | |
Attachments: |
server.log file from GlassFish 3.1.1 b10
Project that shows bug diff that applies to release701 branch |
Description
ievans
2011-07-07 19:38:12 UTC
Created attachment 109301 [details]
server.log file from GlassFish 3.1.1 b10
Got a sample project that you could attach? Created attachment 109302 [details]
Project that shows bug
Steps to reproduce using attached project: 1. Open rsvp project in NetBeans 7.x configured with GlassFish 3.1.1. 2. Run rsvp project. 3. Clean project (this undeploys rsvp but keeps the SEQUENCE table). 4. Run rsvp project again. This will trigger the deployment error, but if you look in GlassFish 3.1.1's Applications list you'll see rsvp deployed. I can replicate. it is a plugin or server issue... not related to the javaee/persistence support of the ide... it looks like there was a change in the way the deploy 'commands' return value is computed in 3.1 vs 3.1.1... in the 3.1 FCS, the return from the second deploy was "SUCCESS" and in 3.1.1 it is "WARNING". The fix is pretty trivial for 7.1. forgot to mark as started this is more significant than p3... Looks serious, should be fixed for 7.0.1. Already fixed in trunk: http://hg.netbeans.org/web-main/rev/4977b072464c I'm not really sure how well the equalsIgnoreCase() deals with equality in diferrent Locales. Perhaps this would be more safe: if (outputCode == null || !"FAILURE".equals(outputCode.toUpperCase(Locale.ENGLISH))) { // NOI18N Otherwise fix looks good. i'd like to understand this a little bit: - deployment failed because of sql exception caused by existing table, which was by server notified as WARNING (previously by SUCCESS)? - do i understand correctly that this will happend every time user runs some app in Glassfish, then cleans project, and tries to deploy again? Is this specific only to this project type and library combination? (I'm trying to figure out how many users would be impact by this) Integrated into 'main-golden' Changeset: http://hg.netbeans.org/main-golden/rev/4977b072464c User: Vince Kraemer <vkraemer@netbeans.org> Log: #199956 : server started sending WARNING messages... surprise. (In reply to comment #11) > i'd like to understand this a little bit: > - deployment failed because of sql exception caused by existing table, which > was by server notified as WARNING (previously by SUCCESS)? The deployment was successful on the server, but the IDE claims that it failed, since the IDE did not get "SUCCESS" as the exit code. > - do i understand correctly that this will happend every time user runs some > app in Glassfish, then cleans project, and tries to deploy again? This will happen when the user does multiple deployments of an app that uses generated PKs and the like... not every app ever, but a fairly sizable subset. > Is this > specific only to this project type and library combination? (I'm trying to > figure out how many users would be impact by this) This would affect any Java EE project type that deploys onto GlassFish 3.1.1 that uses generated IDs (In reply to comment #10) > I'm not really sure how well the equalsIgnoreCase() deals with equality in > diferrent Locales. Perhaps this would be more safe: > > if (outputCode == null || > !"FAILURE".equals(outputCode.toUpperCase(Locale.ENGLISH))) { // NOI18N > > Otherwise fix looks good. great catch. integrated http://hg.netbeans.org/web-main/rev/e4ca1213b3ca Created attachment 109402 [details]
diff that applies to release701 branch
here is what would be applied to release701 branch in the releases repo...
Integrated into 'main-golden' Changeset: http://hg.netbeans.org/main-golden/rev/e4ca1213b3ca User: Vince Kraemer <vkraemer@netbeans.org> Log: #199956 : server started sending WARNING messages... surprise. Now l10n safe for folks running in a Turkish locale. I can verify this, I have tried to reproduce it the way described in comment #4 and it works fine. Product Version: NetBeans IDE Dev (Build 201107130600) Java: 1.7.0; Java HotSpot(TM) Client VM 21.0-b17 System: Windows XP version 5.1 running on x86; Cp1252; en_US (nb) thanks Lado! Please go on with integration to releases 701 branch. http://hg.netbeans.org/releases/rev/04b8324fff00 fix into release701 branch Integrated into 'releases' Changeset: http://hg.netbeans.org/releases/rev/04b8324fff00 User: Vince Kraemer <vkraemer@netbeans.org> Log: #199956 : server warning triggers false deployment failed result in IDE |