harmony-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Pavel Ozhdikhin (JIRA)" <j...@apache.org>
Subject [jira] Commented: (HARMONY-5433) about types
Date Mon, 28 Jan 2008 12:07:38 GMT

    [ https://issues.apache.org/jira/browse/HARMONY-5433?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12563125#action_12563125
] 

Pavel Ozhdikhin commented on HARMONY-5433:
------------------------------------------

Since 574617 JIT uses lazy resolution mode - types are not resolved at compile-time. If the
type of some operand has not yet been resolved before compilation of the method, JIT puts
unresolved type for the operand and adds a call to the resolution helper when type info is
needed in the code. So, actual resolution would be done at run-time. In your example the result
of the call to work() method is not used in the method so it remains unresolved.

You may switch back to the eager resolution mode by using the command line switch: -XX:jit.arg.lazyResolution=off.


> about types
> -----------
>
>                 Key: HARMONY-5433
>                 URL: https://issues.apache.org/jira/browse/HARMONY-5433
>             Project: Harmony
>          Issue Type: Test
>          Components: DRLVM
>         Environment: Harmony-545368 , Harmony-608801
>            Reporter: midstar
>         Attachments: 545368_dump_main.txt, 608801_dump_main.txt, a.java
>
>
> Hi,all! I have run the attached program a.java using opt.emconf in both version 545368
and version 608801,and dumped the HIR insts in both versions.The attached 545368_dump_main.txt
and 608801_dump_main.txt are for the HIR insts of method a::main in the version 545255 and
version 608307 respectively.And I found some differences.
> The fraction of 608307_dump_main.txt is as below:
> I17:call      a::work(t3) ((t9,t10)) -) t11:cls:.Unresolved  bcmap:20 
> and the fraction of 545255_dump_main.txt is as below:
> I17:call      a::work(t3) ((t9,t10)) -) t11:cls:man 
> why the differences exist and why t11 is unresolved in the new version? when it would
be resolved?How can I get it?
> Thanks very much!
>   

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


Mime
View raw message