[JIRA] Assigned: (GF_CORBA_V3_MIRROR-3) Codegen incorrectly computes max stack size for generated methods

  • From: "kcavanaugh (JIRA)" <jira-no-reply@kenai.com>
  • To: issues@gf-corba-v3-mirror.kenai.com
  • Subject: [JIRA] Assigned: (GF_CORBA_V3_MIRROR-3) Codegen incorrectly computes max stack size for generated methods
  • Date: Fri, 13 Mar 2009 19:03:17 +0000 (GMT+00:00)


     [ 
http://kenai.com/jira/browse/GF_CORBA_V3_MIRROR-3?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

kcavanaugh reassigned GF_CORBA_V3_MIRROR-3:
-------------------------------------------

    Assignee: kcavanaugh

> Codegen incorrectly computes max stack size for generated methods
> -----------------------------------------------------------------
>
>                 Key: GF_CORBA_V3_MIRROR-3
>                 URL: http://kenai.com/jira/browse/GF_CORBA_V3_MIRROR-3
>             Project: gf-corba-v3-mirror
>          Issue Type: Bug
>         Environment: GFv3
>            Reporter: kcavanaugh
>            Assignee: kcavanaugh
>
> The current version of codegen in the ORB (VERSION-3.0.0-b011 and earlier) 
> has a bug where occasionally
> the maximum stack size is incorrectly computed for methods with a large 
> number of catch blocks.
> Apparently when codegen generates bytecode through ASTByteCodeVisitor, 
> somehow the handler list
> of catch handlers in the ASM ClassWriter is either not correctly 
> initialized, or is corrupted after it is
> generated.  This results in maxStackSize=0, which causes verifier errors 
> when the generated class is loaded
> in the Java VM.
> Temporary fix: use ClassReader -> ClassWriter with a special visitor that 
> simply forces ASM to re-compute the stack size.
> This works, probably because the ClassReader does not do whatever ASM is 
> doing that screws up the handler list.
> Available in 3.0.0-b012.
> A better fix: compute stack size in ASMByteCodeVisitor (simple recursive 
> decompostion algorithm).  Should be simpler
> at the codegen level because the shape of the flow graph is much more 
> restricted.
> Another possibility is to really fix whatever is wrong with the ASM usage, 
> but the recursive decomposition may
> be a better choice.

-- 
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators: 
http://kenai.com/jira/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira

        


[JIRA] Created: (GF_CORBA_V3_MIRROR-3) Codegen incorrectly computes max stack size for generated methods

kcavanaugh (JIRA) 03/13/2009

[JIRA] Assigned: (GF_CORBA_V3_MIRROR-3) Codegen incorrectly computes max stack size for generated methods

kcavanaugh (JIRA) 03/13/2009
  • Mysql
  • Glassfish
  • Jruby
  • Rails
  • Nblogo
Terms of Use; Privacy Policy;
© 2010, Oracle Corporation and/or its affiliates
(revision 20120518.3c65429)
 
 
Close
loading
Please Confirm
Close