Bugzilla – Full Text Bug Listing
| Summary: | Basic Client & Agent compatibility on Linux IA32 | ||
|---|---|---|---|
| Product: | z_Archived | Reporter: | Karla Callaghan <karla.callaghan> |
| Component: | TPTP | Assignee: | Igor Alelekov <igor.alelekov> |
| Status: | CLOSED FIXED | QA Contact: | |
| Severity: | enhancement | ||
| Priority: | P1 | CC: | christophe.telep, jkubasta, kevin.p.o'leary, randy.d.smith |
| Version: | unspecified | Keywords: | plan |
| Target Milestone: | --- | ||
| Hardware: | PC | ||
| OS: | Linux | ||
| URL: | http://www.eclipse.org/tptp/groups/Architecture/documents/features/hf_125103.html | ||
| Whiteboard: | closed460 | ||
| Bug Depends on: | 154587, 159677, 161096, 161285, 161289 | ||
| Bug Blocks: | 108584, 119571, 121658, 159817 | ||
|
Description
Karla Callaghan
Indications are this needs a couple of makefile changes, changes to the top-level build script, and a few fixes where the Linux compiler balks at code let through by MSDEV. src-native-new/build/build_tptp_ac.script adds new library builds (and removes TPTPConfigBinding which is long gone). src-native-new/src/transport/TPTPAgentCompTL/TPTPAgentCompTL.make -- Fixed library name to shortened tptpACTL form src-native-new/src/transport/TPTPAgentCompTL/AgentCmdHandlers.c -- removed casts which made for invalid lvalues src-native-new/src/transport/TPTPClientCompTL/TPTPClientCompTL.make -- Fixed library name to shortened tptpCCTL form src-native-new/src/transport/TPTPClientCompTL/ClientCTL.c -- expanded += to avoid having cast make for invalid lvalue src-native-new/src/transport/TPTPClientCompTL/RACmdHandler.c -- added static to function declaration to match definition Took about half the time estimated. Great job backward compatibility guys to make this so easy! Reopening this feature so can track it for 4.3 planning. Could not get a few pieces completed and tested on the Linux platforms in time for 4.2, so this feature was deferred for official completion in the next release. Making this an enhancement - needs tracking as such for 4.3 planning. Proposed for 4.3 plan. Reviewed and approved by the AG on July 13 I have just completed my checkins to allow BC to work on linux. This included loading plugins and setting the environment correctly. To test this work I was able to correctly profile java applications using a an AC running under linux-ia32. This code is specific to linux, I also believe this will complete our AC/BC solution on linux-ipf and linux-em64t. The only question is if we should close this bug right now or wait for a complete test pass run. Karla, does this need to be assigned to Kevin? Still, with the 200608230100 nightly build, getting the NoClassDefFound exception. Logfile still has "startProcesses Thread: Error waiting for a startProcess cmd" in servicelog.log when DEBUG on. Thus still can't profile out of the box. Note that I've also added a dependency on bug 154587 ... the SetConfig/security side of things has moved out from under us since the 4.2.0 release. The JVM loading *is* working now... if Jrockit's jre/lib/i386 and jre/lib/i386/jrockit paths are added to LD_LIBRARY_PATH. This will allow the file server testing... but even with the JVM loaded, the profiling fails. I was just able to successfully profile using 0823 build This was running ACServer on sw-opt007 ... Launching java (jrockit) from the command-line as controlled. I then do a remote profile and it seems to work fine. This needs a complete test pass to claim full functionality... but it would be helpful to know if someone else could duplicate my success in this... our env's are different... maybe there is a dependency we are missing. But this is missing the "top priority use case" ... I predict that much higher than attaching external to a running Java process, the most likely use case is straightforward: Start the AC (or not, or RAC), Start the workbench, Profile That is *the* use case in my mind. If we can't do that, I don't think we can say with a straight face to users that "we can profile". are you saying the RAC is not working for you as well? I'm not saying that the use case that you have specified isn't the main use case. We will do a complete test pass which includes all profiling. From the perspective of the AC the messages should be the same for each use case. I was just trying to see if anyone else had been able to get any profiling to work. (if there was something unique to my env or the machine I was running on) Moving over to Kevin, since he's been doing all the work on this. Reassign to Kevin. I have passed the automated test suites using the AC/BC. I have also verified the latest 4.3 builds passed as well. I am still waiting for indepedent confirmation of the tests passes from Thay, but I am closing this bug. If there are further issues with Linux AC/BC I will address as new defects. Re-opening as this feature will not be a part of the 4.3 GA. Intermitent connection problems were discovered when a client tries to attach to the AC when it was started in Secure mode. The code for this support was not platform-specific and worked fine for Windows, but now shows failures on Linux and the reasons are not clear. Discovery occurred late in 4.3 due to the 4.2.1 schedule overlap causing resource conflicts. In order to reduce risk to the 4.3 schedule which has had to be pushed out, have choosen to rollback to using the RAC on Linux IA32 rather than the new AC's backwards compatibility as it has had too little air time. Note that the code for this work is not being removed, the only change is in putting the RAC pkg as the Linux IA32 agent controller download link vs. the new AC pkg. When 4.4 commences and the security connection issues are resolved, the new AC pkg should be put back so that it is tested throughout all the iterations giving time to find/fix other issues that might be lurking. The secure connection issues have been added as dependencies to this bug (159677, 161096, 161285, 161289). Assigning AC backwards compatibility Linux enhancemet to Igor. resolving as fixed As of TPTP 4.6.0, TPTP is in maintenance mode and focusing on improving quality by resolving relevant enhancements/defects and increasing test coverage through test creation, automation, Build Verification Tests (BVTs), and expanded run-time execution. As part of the TPTP Bugzilla housecleaning process (see http://wiki.eclipse.org/Bugzilla_Housecleaning_Processes), this enhancement/defect is verified/closed by the Project Lead since this enhancement/defect has been resolved and unverified for more than 1 year and considered to be fixed. If this enhancement/defect is still unresolved and reproducible in the latest TPTP release (http://www.eclipse.org/tptp/home/downloads/), please re-open. |