You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@commons.apache.org by "Gary D. Gregory" <gg...@apache.org> on 2022/11/17 13:00:21 UTC
[BCEL] Invalid test or bug?
Hi All & Mark Roberts:
I added JavaMathTestCase as a disabled test as it fails.
Is this a legal test to try or do we have a bug?
Gary
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
For additional commands, e-mail: dev-help@commons.apache.org
Re: [BCEL] Invalid test or bug?
Posted by "Gary D. Gregory" <gg...@apache.org>.
Mark,
What I had in mind (compared to your diff) file, is something like this:
https://paste.apache.org/oxn8x
Note that this fixed throwing the exception but there are still many JRE classes that fail to verify.
Gary
On 2022/11/19 15:32:04 Gary Gregory wrote:
> Here is where we stand today:
>
> - I fixed bugs where the JVM specification types data as "u2"
> (unsigned 2 bytes, range 0 to 65535) but we used to read these items
> as signed shorts so we "lost" values from 32768 to 65535 or we read
> them as negative values. Amazingly this did not cause tests to fail,
> which does not say much for our test coverage but these would be hard
> to find without specially crafted class files.
> - On Java8, the previously failing test
> VerifyJavaMathTestCase.testBigDecimal() passes but fails differently
> than it did when run on Java 11 or above. I only enabled the test on
> Java 8.
> - VerifyJavaUtilTestCase still fails to verify all its classes.
> - VerifyJavaHomesTestCase still has tons of failures
>
> I might not hold up releasing 6.7.0 too long since these issues must
> have existed for a long time.
>
> Gary
>
> On Thu, Nov 17, 2022 at 10:53 AM Mark Roberts <ma...@cs.washington.edu> wrote:
> >
> > I will try to take a look at this later today.
> >
> > Mark
> >
> > -----Original Message-----
> > From: Gary D. Gregory [mailto:ggregory@apache.org]
> > Sent: Thursday, November 17, 2022 7:14 AM
> > To: dev@commons.apache.org
> > Subject: Re: [BCEL] Invalid test or bug?
> >
> > More specifically, javap says:
> >
> > 21: invokevirtual #68 // Method
> > "[B".clone:()Ljava/lang/Object;
> >
> > So calling a method on an array with invokevirtual is ok and we have a bug.
> >
> > Thoughts?
> >
> > Gary
> >
> > On 2022/11/17 14:45:41 "Gary D. Gregory" wrote:
> > > Hm, I'm thinking bug when I see javap output like:
> > >
> > > #68 = Methodref #901.#902 //
> > > "[B".clone:()Ljava/lang/Object;
> > >
> > > Thoughts?
> > >
> > > Gary
> > >
> > > On 2022/11/17 13:04:32 "Gary D. Gregory" wrote:
> > > > Actually: VerifyJavaMathTestCase and VerifyJavaUtilTestCase
> > > >
> > > > Gary
> > > >
> > > > On 2022/11/17 13:00:21 "Gary D. Gregory" wrote:
> > > > > Hi All & Mark Roberts:
> > > > >
> > > > > I added JavaMathTestCase as a disabled test as it fails.
> > > > >
> > > > > Is this a legal test to try or do we have a bug?
> > > > >
> > > > > Gary
> > > > >
> > > > > ------------------------------------------------------------------
> > > > > --- To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
> > > > > For additional commands, e-mail: dev-help@commons.apache.org
> > > > >
> > > > >
> > > >
> > > > --------------------------------------------------------------------
> > > > - To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
> > > > For additional commands, e-mail: dev-help@commons.apache.org
> > > >
> > > >
> > >
> > > ---------------------------------------------------------------------
> > > To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
> > > For additional commands, e-mail: dev-help@commons.apache.org
> > >
> > >
> >
> > ---------------------------------------------------------------------
> > To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
> > For additional commands, e-mail: dev-help@commons.apache.org
> >
> > ---------------------------------------------------------------------
> > To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
> > For additional commands, e-mail: dev-help@commons.apache.org
> >
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
> For additional commands, e-mail: dev-help@commons.apache.org
>
>
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
For additional commands, e-mail: dev-help@commons.apache.org
Re: [BCEL] Invalid test or bug?
Posted by Gary Gregory <ga...@gmail.com>.
Here is where we stand today:
- I fixed bugs where the JVM specification types data as "u2"
(unsigned 2 bytes, range 0 to 65535) but we used to read these items
as signed shorts so we "lost" values from 32768 to 65535 or we read
them as negative values. Amazingly this did not cause tests to fail,
which does not say much for our test coverage but these would be hard
to find without specially crafted class files.
- On Java8, the previously failing test
VerifyJavaMathTestCase.testBigDecimal() passes but fails differently
than it did when run on Java 11 or above. I only enabled the test on
Java 8.
- VerifyJavaUtilTestCase still fails to verify all its classes.
- VerifyJavaHomesTestCase still has tons of failures
I might not hold up releasing 6.7.0 too long since these issues must
have existed for a long time.
Gary
On Thu, Nov 17, 2022 at 10:53 AM Mark Roberts <ma...@cs.washington.edu> wrote:
>
> I will try to take a look at this later today.
>
> Mark
>
> -----Original Message-----
> From: Gary D. Gregory [mailto:ggregory@apache.org]
> Sent: Thursday, November 17, 2022 7:14 AM
> To: dev@commons.apache.org
> Subject: Re: [BCEL] Invalid test or bug?
>
> More specifically, javap says:
>
> 21: invokevirtual #68 // Method
> "[B".clone:()Ljava/lang/Object;
>
> So calling a method on an array with invokevirtual is ok and we have a bug.
>
> Thoughts?
>
> Gary
>
> On 2022/11/17 14:45:41 "Gary D. Gregory" wrote:
> > Hm, I'm thinking bug when I see javap output like:
> >
> > #68 = Methodref #901.#902 //
> > "[B".clone:()Ljava/lang/Object;
> >
> > Thoughts?
> >
> > Gary
> >
> > On 2022/11/17 13:04:32 "Gary D. Gregory" wrote:
> > > Actually: VerifyJavaMathTestCase and VerifyJavaUtilTestCase
> > >
> > > Gary
> > >
> > > On 2022/11/17 13:00:21 "Gary D. Gregory" wrote:
> > > > Hi All & Mark Roberts:
> > > >
> > > > I added JavaMathTestCase as a disabled test as it fails.
> > > >
> > > > Is this a legal test to try or do we have a bug?
> > > >
> > > > Gary
> > > >
> > > > ------------------------------------------------------------------
> > > > --- To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
> > > > For additional commands, e-mail: dev-help@commons.apache.org
> > > >
> > > >
> > >
> > > --------------------------------------------------------------------
> > > - To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
> > > For additional commands, e-mail: dev-help@commons.apache.org
> > >
> > >
> >
> > ---------------------------------------------------------------------
> > To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
> > For additional commands, e-mail: dev-help@commons.apache.org
> >
> >
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
> For additional commands, e-mail: dev-help@commons.apache.org
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
> For additional commands, e-mail: dev-help@commons.apache.org
>
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
For additional commands, e-mail: dev-help@commons.apache.org
RE: [BCEL] Invalid test or bug?
Posted by Mark Roberts <ma...@cs.washington.edu>.
I will try to take a look at this later today.
Mark
-----Original Message-----
From: Gary D. Gregory [mailto:ggregory@apache.org]
Sent: Thursday, November 17, 2022 7:14 AM
To: dev@commons.apache.org
Subject: Re: [BCEL] Invalid test or bug?
More specifically, javap says:
21: invokevirtual #68 // Method
"[B".clone:()Ljava/lang/Object;
So calling a method on an array with invokevirtual is ok and we have a bug.
Thoughts?
Gary
On 2022/11/17 14:45:41 "Gary D. Gregory" wrote:
> Hm, I'm thinking bug when I see javap output like:
>
> #68 = Methodref #901.#902 //
> "[B".clone:()Ljava/lang/Object;
>
> Thoughts?
>
> Gary
>
> On 2022/11/17 13:04:32 "Gary D. Gregory" wrote:
> > Actually: VerifyJavaMathTestCase and VerifyJavaUtilTestCase
> >
> > Gary
> >
> > On 2022/11/17 13:00:21 "Gary D. Gregory" wrote:
> > > Hi All & Mark Roberts:
> > >
> > > I added JavaMathTestCase as a disabled test as it fails.
> > >
> > > Is this a legal test to try or do we have a bug?
> > >
> > > Gary
> > >
> > > ------------------------------------------------------------------
> > > --- To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
> > > For additional commands, e-mail: dev-help@commons.apache.org
> > >
> > >
> >
> > --------------------------------------------------------------------
> > - To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
> > For additional commands, e-mail: dev-help@commons.apache.org
> >
> >
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
> For additional commands, e-mail: dev-help@commons.apache.org
>
>
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
For additional commands, e-mail: dev-help@commons.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
For additional commands, e-mail: dev-help@commons.apache.org
Re: [BCEL] Invalid test or bug?
Posted by "Gary D. Gregory" <gg...@apache.org>.
More specifically, javap says:
21: invokevirtual #68 // Method "[B".clone:()Ljava/lang/Object;
So calling a method on an array with invokevirtual is ok and we have a bug.
Thoughts?
Gary
On 2022/11/17 14:45:41 "Gary D. Gregory" wrote:
> Hm, I'm thinking bug when I see javap output like:
>
> #68 = Methodref #901.#902 // "[B".clone:()Ljava/lang/Object;
>
> Thoughts?
>
> Gary
>
> On 2022/11/17 13:04:32 "Gary D. Gregory" wrote:
> > Actually: VerifyJavaMathTestCase and VerifyJavaUtilTestCase
> >
> > Gary
> >
> > On 2022/11/17 13:00:21 "Gary D. Gregory" wrote:
> > > Hi All & Mark Roberts:
> > >
> > > I added JavaMathTestCase as a disabled test as it fails.
> > >
> > > Is this a legal test to try or do we have a bug?
> > >
> > > Gary
> > >
> > > ---------------------------------------------------------------------
> > > To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
> > > For additional commands, e-mail: dev-help@commons.apache.org
> > >
> > >
> >
> > ---------------------------------------------------------------------
> > To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
> > For additional commands, e-mail: dev-help@commons.apache.org
> >
> >
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
> For additional commands, e-mail: dev-help@commons.apache.org
>
>
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
For additional commands, e-mail: dev-help@commons.apache.org
Re: [BCEL] Invalid test or bug?
Posted by "Gary D. Gregory" <gg...@apache.org>.
Hm, I'm thinking bug when I see javap output like:
#68 = Methodref #901.#902 // "[B".clone:()Ljava/lang/Object;
Thoughts?
Gary
On 2022/11/17 13:04:32 "Gary D. Gregory" wrote:
> Actually: VerifyJavaMathTestCase and VerifyJavaUtilTestCase
>
> Gary
>
> On 2022/11/17 13:00:21 "Gary D. Gregory" wrote:
> > Hi All & Mark Roberts:
> >
> > I added JavaMathTestCase as a disabled test as it fails.
> >
> > Is this a legal test to try or do we have a bug?
> >
> > Gary
> >
> > ---------------------------------------------------------------------
> > To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
> > For additional commands, e-mail: dev-help@commons.apache.org
> >
> >
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
> For additional commands, e-mail: dev-help@commons.apache.org
>
>
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
For additional commands, e-mail: dev-help@commons.apache.org
Re: [BCEL] Invalid test or bug?
Posted by "Gary D. Gregory" <gg...@apache.org>.
Actually: VerifyJavaMathTestCase and VerifyJavaUtilTestCase
Gary
On 2022/11/17 13:00:21 "Gary D. Gregory" wrote:
> Hi All & Mark Roberts:
>
> I added JavaMathTestCase as a disabled test as it fails.
>
> Is this a legal test to try or do we have a bug?
>
> Gary
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
> For additional commands, e-mail: dev-help@commons.apache.org
>
>
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
For additional commands, e-mail: dev-help@commons.apache.org