harmony-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Geir Magnusson Jr." <g...@pobox.com>
Subject Re: [jira] Updated: (HARMONY-3222) assertion fires with DaCapo eclipse (jitrino built in debug)
Date Fri, 02 Mar 2007 18:54:51 GMT
I think that if it's interesting (non trivial), it's not a bad idea  
to mention it on the dev list.  More eyes...

geir


On Feb 26, 2007, at 11:50 PM, Alexey Varlamov wrote:

> I believe it depends on issue priority/severity: normally it is enough
> to set "Patch available flag", but if you need the patch reviewed &
> integrated urgently, feel free to alert on the dev-list.
> --
> Alexey
>
> 2007/2/27, Naveen Neelakantam <neelakan@uiuc.edu>:
>> Fix provided.
>>
>> Should I be emailing the dev list when I fix something?  Or should
>> the issue just be marked as resolved?
>>
>> I'm not really clear on the protocol (and yes, I read http://
>> harmony.apache.org/issue_resolution_guideline.html).  :-)
>>
>> Thanks,
>> Naveen
>>
>> Begin forwarded message:
>>
>> > From: "Naveen Neelakantam (JIRA)" <jira@apache.org>
>> > Date: February 26, 2007 6:41:05 PM CST
>> > To: neelakan@uiuc.edu
>> > Subject: [jira] Updated: (HARMONY-3222) assertion fires with DaCapo
>> > eclipse (jitrino built in debug)
>> >
>> >
>> >      [ https://issues.apache.org/jira/browse/HARMONY-3222?
>> > page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
>> >
>> > Naveen Neelakantam updated HARMONY-3222:
>> > ----------------------------------------
>> >
>> >     Attachment: imm_cast_bug.patch
>> >
>> > This patch (imm_cast_bug) fixes the issue.  The problem is with
>> > casting.  Basically the following was occuring:
>> >
>> > in32 cast to uint32 cast to int64
>> >
>> > As a result, negative constants would become large positive  
>> constants:
>> >
>> > -1 cast to 4294967295 cast to 4294967295
>> >
>> > If this casted value was used to dereference an array it would be
>> > multipled by the array element size (let's say 4),  which would
>> > exceed the maximum value that could fit in a 32-bit integer:
>> >
>> > 4294967295* 4 = 17179869180
>> >
>> >> assertion fires with DaCapo eclipse (jitrino built in debug)
>> >> ------------------------------------------------------------
>> >>
>> >>                 Key: HARMONY-3222
>> >>                 URL: https://issues.apache.org/jira/browse/
>> >> HARMONY-3222
>> >>             Project: Harmony
>> >>          Issue Type: Bug
>> >>          Components: DRLVM
>> >>         Environment: RHEL4 update 4, core 2 duo, gcc 4.1.0
>> >>            Reporter: Naveen Neelakantam
>> >>         Attachments: imm_cast_bug.patch
>> >>
>> >>
>> >> To see the assert, jitrino must be built in debug mode and the -
>> >> Xem:server option must be used
>> >>> java -showversion -Xem:server -jar dacapo-2006-10-MR1.jar eclipse
>> >> Apache Harmony Launcher : (c) Copyright 1991, 2006 The Apache
>> >> Software Foundation or its licensors, as applicable.
>> >> java version "1.5.0"
>> >> pre-alpha : not complete or compatible
>> >> svn = r510120, (Feb 21 2007), Linux/ia32/gcc 4.1.0, debug build
>> >> http://incubator.apache.org/harmony
>> >> ===== DaCapo eclipse starting =====
>> >> <setting up workspace...>
>> >> <creating
>> >>  
>> projects.............................................................
>> >> .>
>> >> <running tests at level 0...>
>> >> <performing build tests...>
>> >>         org.apache.ant (not open) opening cleaning building
>> >>         org.junit (not open) opening cleaning building
>> >> java: /home/zilles/neelakan/Sandbox/Harmony/test/working_vm/vm/
>> >> jitrino/src/codegenerator/ia32/Ia32Inst.cpp:92: void
>> >> Jitrino::Ia32::Opnd::setMemOpndSubOpnd
>> >> (Jitrino::Ia32::MemOpndSubOpndKind, Jitrino::Ia32::Opnd*):
>> >> Assertion `(so != MemOpndSubOpndKind_Displacement) || (2147483647
>> >> >= opnd->getImmValue() && (-2147483647 - 1) <= opnd->getImmValue
>> >> ())' failed.
>> >
>> > --
>> > 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