You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@mahout.apache.org by Benson Margulies <bi...@gmail.com> on 2010/02/08 03:37:45 UTC

]math[ 1.5, really?

Eclipse thinks that the collections code has about 1000 instance of
@Override of an INTERFACE method, which is supposed to be invalid in
1.5 and valid in 1.6.

Re: ]math[ 1.5, really?

Posted by Dawid Weiss <da...@gmail.com>.
>> was in the pom.xml file.  But it worked. Until like yesterday. Color me
>> confused.

Perhaps you had these classes compiled from the previous runs (when
1.6 flag was on)? I don't see how it could work with javac.

Oh, the third option is to remove @Override; it's not that useful anyway.

D.

Re: ]math[ 1.5, really?

Posted by Dawid Weiss <da...@gmail.com>.
> It's only some @Overrides -- those on interfaces. My vote is to
> eliminate  them  from the 1.5-compatible projects.

+1 from me.

D.

Re: ]math[ 1.5, really?

Posted by Benson Margulies <bi...@gmail.com>.
It's only some @Overrides -- those on interfaces. My vote is to
eliminate  them  from the 1.5-compatible projects.


On Mon, Feb 8, 2010 at 7:47 AM, Sean <sr...@gmail.com> wrote:
> Well javac pretty much defines right behavior. If it works now with
> 1.5/1.5, and there's some logic to it, then OK. It means we'd have to
> flip our IDEs to allow language level 1.6 (but API level 1.5) and all
> is basically fine.
>
> On Mon, Feb 8, 2010 at 12:44 PM, Benson Margulies <bi...@gmail.com> wrote:
>> I don't know from Intellij, but here's my best attempt to explain.
>>
>> Step 1: Sun has a bug. javac should reject these annotations, but does not.
>>
>> Step 2: Various tools do not have this bug. They warn or explode.
>>
>

Re: ]math[ 1.5, really?

Posted by Sean <sr...@gmail.com>.
Well javac pretty much defines right behavior. If it works now with
1.5/1.5, and there's some logic to it, then OK. It means we'd have to
flip our IDEs to allow language level 1.6 (but API level 1.5) and all
is basically fine.

On Mon, Feb 8, 2010 at 12:44 PM, Benson Margulies <bi...@gmail.com> wrote:
> I don't know from Intellij, but here's my best attempt to explain.
>
> Step 1: Sun has a bug. javac should reject these annotations, but does not.
>
> Step 2: Various tools do not have this bug. They warn or explode.
>

Re: ]math[ 1.5, really?

Posted by Benson Margulies <bi...@gmail.com>.
I don't know from Intellij, but here's my best attempt to explain.

Step 1: Sun has a bug. javac should reject these annotations, but does not.

Step 2: Various tools do not have this bug. They warn or explode.



On Mon, Feb 8, 2010 at 7:30 AM, Sean <sr...@gmail.com> wrote:
> But why would it work in Eclipse and not with javac? or rather, what's
> important is what javac does, in the main build. But that was working.
> But it doesn't. And I also just saw my IDE settings flip or something
> and it went haywire (IntelliJ). I am really confused.
>
> On Mon, Feb 8, 2010 at 12:17 PM, Benson Margulies <bi...@gmail.com> wrote:
>> Sorry, you misunderstood me as to the nature of the problem I hit.
>>
>> If you start with a clean tree, and run the maven-eclipse-plugin, it
>> will copy the 1.5 settings from maven/javac to eclipse, and eclipse
>> will go ballistic. You can then manually reset the eclipse settings,
>> and it shuts up. In my personal opinion, this is a metastable
>> situation, and it would be better to kill all the annotations, since
>> some people like to use maven to run the eclipse \compiler/. However,
>> that's just my opinion.
>>
>>
>>
>> On Mon, Feb 8, 2010 at 6:10 AM, Sean <sr...@gmail.com> wrote:
>>> Yah that's what I thought too but Benson says that doesn't work and not what
>>> was in the pom.xml file.  But it worked. Until like yesterday. Color me
>>> confused.
>>>
>>> On Feb 8, 2010 9:00 AM, "Dawid Weiss" <da...@gmail.com> wrote:
>>>
>>> I wrote a post about this a while ago. You need to use the 1.6
>>> compiler, but set the target to 1.5 -- this way you can keep @Override
>>> annotations, but emit valid 1.5 code anyway. I don't know about Maven
>>> (javac), but it definitely works in Eclipse (can be set manually via
>>> project properties).
>>>
>>> D.
>>>
>>> On Mon, Feb 8, 2010 at 3:51 AM, Benson Margulies <bi...@gmail.com>
>>> wrote: > I thought we were...
>>>
>>
>

Re: ]math[ 1.5, really?

Posted by Sean <sr...@gmail.com>.
But why would it work in Eclipse and not with javac? or rather, what's
important is what javac does, in the main build. But that was working.
But it doesn't. And I also just saw my IDE settings flip or something
and it went haywire (IntelliJ). I am really confused.

On Mon, Feb 8, 2010 at 12:17 PM, Benson Margulies <bi...@gmail.com> wrote:
> Sorry, you misunderstood me as to the nature of the problem I hit.
>
> If you start with a clean tree, and run the maven-eclipse-plugin, it
> will copy the 1.5 settings from maven/javac to eclipse, and eclipse
> will go ballistic. You can then manually reset the eclipse settings,
> and it shuts up. In my personal opinion, this is a metastable
> situation, and it would be better to kill all the annotations, since
> some people like to use maven to run the eclipse \compiler/. However,
> that's just my opinion.
>
>
>
> On Mon, Feb 8, 2010 at 6:10 AM, Sean <sr...@gmail.com> wrote:
>> Yah that's what I thought too but Benson says that doesn't work and not what
>> was in the pom.xml file.  But it worked. Until like yesterday. Color me
>> confused.
>>
>> On Feb 8, 2010 9:00 AM, "Dawid Weiss" <da...@gmail.com> wrote:
>>
>> I wrote a post about this a while ago. You need to use the 1.6
>> compiler, but set the target to 1.5 -- this way you can keep @Override
>> annotations, but emit valid 1.5 code anyway. I don't know about Maven
>> (javac), but it definitely works in Eclipse (can be set manually via
>> project properties).
>>
>> D.
>>
>> On Mon, Feb 8, 2010 at 3:51 AM, Benson Margulies <bi...@gmail.com>
>> wrote: > I thought we were...
>>
>

Re: ]math[ 1.5, really?

Posted by Benson Margulies <bi...@gmail.com>.
Sorry, you misunderstood me as to the nature of the problem I hit.

If you start with a clean tree, and run the maven-eclipse-plugin, it
will copy the 1.5 settings from maven/javac to eclipse, and eclipse
will go ballistic. You can then manually reset the eclipse settings,
and it shuts up. In my personal opinion, this is a metastable
situation, and it would be better to kill all the annotations, since
some people like to use maven to run the eclipse \compiler/. However,
that's just my opinion.



On Mon, Feb 8, 2010 at 6:10 AM, Sean <sr...@gmail.com> wrote:
> Yah that's what I thought too but Benson says that doesn't work and not what
> was in the pom.xml file.  But it worked. Until like yesterday. Color me
> confused.
>
> On Feb 8, 2010 9:00 AM, "Dawid Weiss" <da...@gmail.com> wrote:
>
> I wrote a post about this a while ago. You need to use the 1.6
> compiler, but set the target to 1.5 -- this way you can keep @Override
> annotations, but emit valid 1.5 code anyway. I don't know about Maven
> (javac), but it definitely works in Eclipse (can be set manually via
> project properties).
>
> D.
>
> On Mon, Feb 8, 2010 at 3:51 AM, Benson Margulies <bi...@gmail.com>
> wrote: > I thought we were...
>

Re: ]math[ 1.5, really?

Posted by Sean <sr...@gmail.com>.
I did the rest. Anyone care to try the compilation now?

On Mon, Feb 8, 2010 at 11:05 PM, Benson Margulies <bi...@gmail.com> wrote:
> You can leave collections to me.

Re: ]math[ 1.5, really?

Posted by Benson Margulies <bi...@gmail.com>.
You can leave collections to me.

On Mon, Feb 8, 2010 at 10:04 AM, Sean <sr...@gmail.com> wrote:
> I'll get on this then, kicking out @Override, since I imagine I'm the
> one that put them there.
>
> On Mon, Feb 8, 2010 at 1:38 PM, Benson Margulies <bi...@gmail.com> wrote:
>> Sean,
>>
>> I keep linux in a virtual box on my mac so that I can test things
>> against a reallyo-trullyo sun 1.5.
>>
>> --benson
>

Re: ]math[ 1.5, really?

Posted by Sean <sr...@gmail.com>.
I'll get on this then, kicking out @Override, since I imagine I'm the
one that put them there.

On Mon, Feb 8, 2010 at 1:38 PM, Benson Margulies <bi...@gmail.com> wrote:
> Sean,
>
> I keep linux in a virtual box on my mac so that I can test things
> against a reallyo-trullyo sun 1.5.
>
> --benson

Re: ]math[ 1.5, really?

Posted by Benson Margulies <bi...@gmail.com>.
Sean,

I keep linux in a virtual box on my mac so that I can test things
against a reallyo-trullyo sun 1.5.

--benson


On Mon, Feb 8, 2010 at 8:30 AM, Sean <sr...@gmail.com> wrote:
> I'm using the stock build setup from SVN, Maven + pom.xml. It's
> telling javac to use 1.5.
> Now, the sticky point here is that OS X doesn't have 1.5 by nature.
> However I've installed and linked it up per Grant's instructions. But
> I still maybe suspect this is the issue.
>
> So if someone else has a build failure from clean checkout, there's an
> environment difference, and I'd suspect it's mine that's different. OK
> so pull the @Overrides. Just want to triple-check this is the evidence
> we have.
>
> On Mon, Feb 8, 2010 at 1:20 PM, Benson Margulies <bi...@gmail.com> wrote:
>> On Mon, Feb 8, 2010 at 8:03 AM, Sean <sr...@gmail.com> wrote:
>>> Wait, what? it works perfectly now. mvn clean compile succeeds.
>>
>> Are you really using 1.5, or just telling 1.6 to pretend to be 1.5.
>> Dawid reports failure when really using 1.5.
>>
>>
>>  If I
>>> don't misunderstand, the issue is that some of our tools behave
>>> differently and we have to manually override it once (maybe again if
>>> the pom changes). Compare that to taking a small step backwards and
>>> the work of removing those @Overrides and I'm not sure why we'd do
>>> that.
>>>
>>> On Mon, Feb 8, 2010 at 12:55 PM, Benson Margulies <bi...@gmail.com> wrote:
>>>> In other words, what we have doesn't really work in 1.5. We need to
>>>> fix it or give up and just aim for 1.6.
>>>>
>>>
>>
>

Re: ]math[ 1.5, really?

Posted by Sean <sr...@gmail.com>.
I'm using the stock build setup from SVN, Maven + pom.xml. It's
telling javac to use 1.5.
Now, the sticky point here is that OS X doesn't have 1.5 by nature.
However I've installed and linked it up per Grant's instructions. But
I still maybe suspect this is the issue.

So if someone else has a build failure from clean checkout, there's an
environment difference, and I'd suspect it's mine that's different. OK
so pull the @Overrides. Just want to triple-check this is the evidence
we have.

On Mon, Feb 8, 2010 at 1:20 PM, Benson Margulies <bi...@gmail.com> wrote:
> On Mon, Feb 8, 2010 at 8:03 AM, Sean <sr...@gmail.com> wrote:
>> Wait, what? it works perfectly now. mvn clean compile succeeds.
>
> Are you really using 1.5, or just telling 1.6 to pretend to be 1.5.
> Dawid reports failure when really using 1.5.
>
>
>  If I
>> don't misunderstand, the issue is that some of our tools behave
>> differently and we have to manually override it once (maybe again if
>> the pom changes). Compare that to taking a small step backwards and
>> the work of removing those @Overrides and I'm not sure why we'd do
>> that.
>>
>> On Mon, Feb 8, 2010 at 12:55 PM, Benson Margulies <bi...@gmail.com> wrote:
>>> In other words, what we have doesn't really work in 1.5. We need to
>>> fix it or give up and just aim for 1.6.
>>>
>>
>

Re: ]math[ 1.5, really?

Posted by Benson Margulies <bi...@gmail.com>.
On Mon, Feb 8, 2010 at 8:03 AM, Sean <sr...@gmail.com> wrote:
> Wait, what? it works perfectly now. mvn clean compile succeeds.

Are you really using 1.5, or just telling 1.6 to pretend to be 1.5.
Dawid reports failure when really using 1.5.


 If I
> don't misunderstand, the issue is that some of our tools behave
> differently and we have to manually override it once (maybe again if
> the pom changes). Compare that to taking a small step backwards and
> the work of removing those @Overrides and I'm not sure why we'd do
> that.
>
> On Mon, Feb 8, 2010 at 12:55 PM, Benson Margulies <bi...@gmail.com> wrote:
>> In other words, what we have doesn't really work in 1.5. We need to
>> fix it or give up and just aim for 1.6.
>>
>

Re: ]math[ 1.5, really?

Posted by Sean <sr...@gmail.com>.
Wait, what? it works perfectly now. mvn clean compile succeeds. If I
don't misunderstand, the issue is that some of our tools behave
differently and we have to manually override it once (maybe again if
the pom changes). Compare that to taking a small step backwards and
the work of removing those @Overrides and I'm not sure why we'd do
that.

On Mon, Feb 8, 2010 at 12:55 PM, Benson Margulies <bi...@gmail.com> wrote:
> In other words, what we have doesn't really work in 1.5. We need to
> fix it or give up and just aim for 1.6.
>

Re: ]math[ 1.5, really?

Posted by Benson Margulies <bi...@gmail.com>.
In other words, what we have doesn't really work in 1.5. We need to
fix it or give up and just aim for 1.6.

Re: ]math[ 1.5, really?

Posted by Dawid Weiss <da...@gmail.com>.
> Your experience is the reverse of mine. In maven, no javac complaints.
> In eclipse, plenty-o-complaints.

Ooops, had the wrong class as the test. Correct, just tried to compile this:

public class OverrideIsHell
{
    public interface A
    {
        public void a();
    }

    public class B implements A
    {
        @Override
        public void a()
        {
        }
    }
}

with ANT:


<project name="test" default="dist">
	<!--
	<property name="build.compiler"
value="org.eclipse.jdt.core.JDTCompilerAdapter"/>
	-->

	<target name="dist">
		<delete dir="tmp/classes" />
		<mkdir dir="tmp/classes" />
		<javac destdir="tmp/classes" source="1.5" target="1.5">
			<src location="src2" />
		</javac>
	</target>
</project>

With javac, this compiles cleanly (1.6). With ecj:

dist:
    [javac] Compiling 1 source file to /home/dweiss/...test/tmp/classes
    [javac] ----------
    [javac] 1. ERROR in /home/dweiss/...test/src2/OverrideIsHell.java
(at line 11)
    [javac] 	public void a()
    [javac] 	            ^^^
    [javac] The method a() of type OverrideIsHell.B must override a
superclass method
    [javac] ----------
    [javac] 1 problem (1 error)

BUILD FAILED
build.xml:8: Compile failed; see the compiler error output for details.

If you go with javac again, but from the 1.5 JDK:

    [javac] Compiling 1 source file to /home/dweiss/... test/tmp/classes
    [javac] ...test/src2/OverrideIsHell.java:10: method does not
override a method from its superclass
    [javac]         @Override
    [javac]          ^
    [javac] 1 error

BUILD FAILED

D.

Re: ]math[ 1.5, really?

Posted by Benson Margulies <bi...@gmail.com>.
Dawid,

Your experience is the reverse of mine. In maven, no javac complaints.
In eclipse, plenty-o-complaints.

--benson


On Mon, Feb 8, 2010 at 7:43 AM, Dawid Weiss <da...@gmail.com> wrote:
> Confirmed, javac will complain about it, JDT does not. Interesting. Workarounds:
>
> - preprocess sources from ant and substitute @Override -> <empty>
> - compile with ejc (Eclipse compiler).
>
> The second worked for me if you compile from ant, for example:
>
>        <property name="build.compiler"
> value="org.eclipse.jdt.core.JDTCompilerAdapter"/>
>
>        <target name="dist">
>                <echo>${eclipse.home}</echo>
>                <javac destdir="tmp/classes" source="1.6" target="1.5">
>                        <src location="src2" />
>                        <compilerarg value="-version" />
>                </javac>
>        </target>
>
> it's just the javac that is picky about @Override.
>
> D.
>
> On Mon, Feb 8, 2010 at 12:10 PM, Sean <sr...@gmail.com> wrote:
>> Yah that's what I thought too but Benson says that doesn't work and not what
>> was in the pom.xml file.  But it worked. Until like yesterday. Color me
>> confused.
>>
>> On Feb 8, 2010 9:00 AM, "Dawid Weiss" <da...@gmail.com> wrote:
>>
>> I wrote a post about this a while ago. You need to use the 1.6
>> compiler, but set the target to 1.5 -- this way you can keep @Override
>> annotations, but emit valid 1.5 code anyway. I don't know about Maven
>> (javac), but it definitely works in Eclipse (can be set manually via
>> project properties).
>>
>> D.
>>
>> On Mon, Feb 8, 2010 at 3:51 AM, Benson Margulies <bi...@gmail.com>
>> wrote: > I thought we were...
>>
>

Re: ]math[ 1.5, really?

Posted by Dawid Weiss <da...@gmail.com>.
Confirmed, javac will complain about it, JDT does not. Interesting. Workarounds:

- preprocess sources from ant and substitute @Override -> <empty>
- compile with ejc (Eclipse compiler).

The second worked for me if you compile from ant, for example:

	<property name="build.compiler"
value="org.eclipse.jdt.core.JDTCompilerAdapter"/>

	<target name="dist">
		<echo>${eclipse.home}</echo>
		<javac destdir="tmp/classes" source="1.6" target="1.5">
			<src location="src2" />
			<compilerarg value="-version" />
		</javac>
	</target>

it's just the javac that is picky about @Override.

D.

On Mon, Feb 8, 2010 at 12:10 PM, Sean <sr...@gmail.com> wrote:
> Yah that's what I thought too but Benson says that doesn't work and not what
> was in the pom.xml file.  But it worked. Until like yesterday. Color me
> confused.
>
> On Feb 8, 2010 9:00 AM, "Dawid Weiss" <da...@gmail.com> wrote:
>
> I wrote a post about this a while ago. You need to use the 1.6
> compiler, but set the target to 1.5 -- this way you can keep @Override
> annotations, but emit valid 1.5 code anyway. I don't know about Maven
> (javac), but it definitely works in Eclipse (can be set manually via
> project properties).
>
> D.
>
> On Mon, Feb 8, 2010 at 3:51 AM, Benson Margulies <bi...@gmail.com>
> wrote: > I thought we were...
>

Re: ]math[ 1.5, really?

Posted by Sean <sr...@gmail.com>.
Yah that's what I thought too but Benson says that doesn't work and not what
was in the pom.xml file.  But it worked. Until like yesterday. Color me
confused.

On Feb 8, 2010 9:00 AM, "Dawid Weiss" <da...@gmail.com> wrote:

I wrote a post about this a while ago. You need to use the 1.6
compiler, but set the target to 1.5 -- this way you can keep @Override
annotations, but emit valid 1.5 code anyway. I don't know about Maven
(javac), but it definitely works in Eclipse (can be set manually via
project properties).

D.

On Mon, Feb 8, 2010 at 3:51 AM, Benson Margulies <bi...@gmail.com>
wrote: > I thought we were...

Re: ]math[ 1.5, really?

Posted by Dawid Weiss <da...@gmail.com>.
I wrote a post about this a while ago. You need to use the 1.6
compiler, but set the target to 1.5 -- this way you can keep @Override
annotations, but emit valid 1.5 code anyway. I don't know about Maven
(javac), but it definitely works in Eclipse (can be set manually via
project properties).

D.

On Mon, Feb 8, 2010 at 3:51 AM, Benson Margulies <bi...@gmail.com> wrote:
> I thought we were trying for 1.5 compat after all. If we are, someone
> (me) should fix all of these. If not, we should change the compiler
> plugin in the POM to say 1.6.
>
> Eclipse treats this as a fatal compilation error if the setting is '1.5.'.
>
> On Sun, Feb 7, 2010 at 9:46 PM, Sean <sr...@gmail.com> wrote:
>> Yes, that's all correct, what's the issue?
>>
>> On Mon, Feb 8, 2010 at 2:37 AM, Benson Margulies <bi...@gmail.com> wrote:
>>> Eclipse thinks that the collections code has about 1000 instance of
>>> @Override of an INTERFACE method, which is supposed to be invalid in
>>> 1.5 and valid in 1.6.
>>>
>>
>

Re: ]math[ 1.5, really?

Posted by Robin Anil <ro...@gmail.com>.
Benson, there are many places where FindBugs and PMD plugin scream of
issues. I saw some 5K issues on math+collections alone in PMD. I helped
remove some from the set and map templates. Do check it out

Robin


On Mon, Feb 8, 2010 at 9:10 AM, Sean <sr...@gmail.com> wrote:

> Hum, I guess that's the final answer if it's from javac. Though in
> theory there's not a reason this must be so. I forget if we already
> went through this.
>
> But what I don't get is why it's stopped working. I also saw this
> problem today. The annotations and pom have been that way for some
> time. Did something become an interface?
>
> No I don't think the whole project should use 1.5 -- we can't in fact
> since Hadoop uses 1.6.
>
> Worst case we remove those annotations but want to figure out what
> changed first.
>
> On Mon, Feb 8, 2010 at 3:03 AM, Benson Margulies <bi...@gmail.com>
> wrote:
> > Oh, well, maven doesn't think this possible:
> >
> > [INFO] Compilation failure
> > Failure executing javac, but could not parse the error:
> > javac: source release 1.6 requires target release 1.6
> >
> >
> >
> > On Sun, Feb 7, 2010 at 10:00 PM, Benson Margulies <bi...@gmail.com>
> wrote:
> >> OK, that's not what we have in the POM:
> >>
> >> <configuration>
> >>          <source>1.5</source>
> >>          <target>1.5</target>
> >>        </configuration>
> >>
> >> I guess I'll make a patch for that.
> >>
> >> On Sun, Feb 7, 2010 at 9:53 PM, Sean <sr...@gmail.com> wrote:
> >>> Yes, runtime compatibility, not compile time. Compiler should treat
> >>> the code as 1.6 and output 1.5 bytecode.
> >>>
> >>> On Mon, Feb 8, 2010 at 2:51 AM, Benson Margulies <
> bimargulies@gmail.com> wrote:
> >>>> I thought we were trying for 1.5 compat after all. If we are, someone
> >>>> (me) should fix all of these. If not, we should change the compiler
> >>>> plugin in the POM to say 1.6.
> >>>>
> >>>> Eclipse treats this as a fatal compilation error if the setting is
> '1.5.'.
> >>>>
> >>>
> >>
> >
>

Re: ]math[ 1.5, really?

Posted by Sean <sr...@gmail.com>.
Hum, I guess that's the final answer if it's from javac. Though in
theory there's not a reason this must be so. I forget if we already
went through this.

But what I don't get is why it's stopped working. I also saw this
problem today. The annotations and pom have been that way for some
time. Did something become an interface?

No I don't think the whole project should use 1.5 -- we can't in fact
since Hadoop uses 1.6.

Worst case we remove those annotations but want to figure out what
changed first.

On Mon, Feb 8, 2010 at 3:03 AM, Benson Margulies <bi...@gmail.com> wrote:
> Oh, well, maven doesn't think this possible:
>
> [INFO] Compilation failure
> Failure executing javac, but could not parse the error:
> javac: source release 1.6 requires target release 1.6
>
>
>
> On Sun, Feb 7, 2010 at 10:00 PM, Benson Margulies <bi...@gmail.com> wrote:
>> OK, that's not what we have in the POM:
>>
>> <configuration>
>>          <source>1.5</source>
>>          <target>1.5</target>
>>        </configuration>
>>
>> I guess I'll make a patch for that.
>>
>> On Sun, Feb 7, 2010 at 9:53 PM, Sean <sr...@gmail.com> wrote:
>>> Yes, runtime compatibility, not compile time. Compiler should treat
>>> the code as 1.6 and output 1.5 bytecode.
>>>
>>> On Mon, Feb 8, 2010 at 2:51 AM, Benson Margulies <bi...@gmail.com> wrote:
>>>> I thought we were trying for 1.5 compat after all. If we are, someone
>>>> (me) should fix all of these. If not, we should change the compiler
>>>> plugin in the POM to say 1.6.
>>>>
>>>> Eclipse treats this as a fatal compilation error if the setting is '1.5.'.
>>>>
>>>
>>
>

Re: ]math[ 1.5, really?

Posted by Benson Margulies <bi...@gmail.com>.
Oh, well, maven doesn't think this possible:

[INFO] Compilation failure
Failure executing javac, but could not parse the error:
javac: source release 1.6 requires target release 1.6



On Sun, Feb 7, 2010 at 10:00 PM, Benson Margulies <bi...@gmail.com> wrote:
> OK, that's not what we have in the POM:
>
> <configuration>
>          <source>1.5</source>
>          <target>1.5</target>
>        </configuration>
>
> I guess I'll make a patch for that.
>
> On Sun, Feb 7, 2010 at 9:53 PM, Sean <sr...@gmail.com> wrote:
>> Yes, runtime compatibility, not compile time. Compiler should treat
>> the code as 1.6 and output 1.5 bytecode.
>>
>> On Mon, Feb 8, 2010 at 2:51 AM, Benson Margulies <bi...@gmail.com> wrote:
>>> I thought we were trying for 1.5 compat after all. If we are, someone
>>> (me) should fix all of these. If not, we should change the compiler
>>> plugin in the POM to say 1.6.
>>>
>>> Eclipse treats this as a fatal compilation error if the setting is '1.5.'.
>>>
>>
>

Re: ]math[ 1.5, really?

Posted by Benson Margulies <bi...@gmail.com>.
OK, that's not what we have in the POM:

<configuration>
          <source>1.5</source>
          <target>1.5</target>
        </configuration>

I guess I'll make a patch for that.

On Sun, Feb 7, 2010 at 9:53 PM, Sean <sr...@gmail.com> wrote:
> Yes, runtime compatibility, not compile time. Compiler should treat
> the code as 1.6 and output 1.5 bytecode.
>
> On Mon, Feb 8, 2010 at 2:51 AM, Benson Margulies <bi...@gmail.com> wrote:
>> I thought we were trying for 1.5 compat after all. If we are, someone
>> (me) should fix all of these. If not, we should change the compiler
>> plugin in the POM to say 1.6.
>>
>> Eclipse treats this as a fatal compilation error if the setting is '1.5.'.
>>
>

Re: ]math[ 1.5, really?

Posted by Sean <sr...@gmail.com>.
Yes, runtime compatibility, not compile time. Compiler should treat
the code as 1.6 and output 1.5 bytecode.

On Mon, Feb 8, 2010 at 2:51 AM, Benson Margulies <bi...@gmail.com> wrote:
> I thought we were trying for 1.5 compat after all. If we are, someone
> (me) should fix all of these. If not, we should change the compiler
> plugin in the POM to say 1.6.
>
> Eclipse treats this as a fatal compilation error if the setting is '1.5.'.
>

Re: ]math[ 1.5, really?

Posted by Benson Margulies <bi...@gmail.com>.
I thought we were trying for 1.5 compat after all. If we are, someone
(me) should fix all of these. If not, we should change the compiler
plugin in the POM to say 1.6.

Eclipse treats this as a fatal compilation error if the setting is '1.5.'.

On Sun, Feb 7, 2010 at 9:46 PM, Sean <sr...@gmail.com> wrote:
> Yes, that's all correct, what's the issue?
>
> On Mon, Feb 8, 2010 at 2:37 AM, Benson Margulies <bi...@gmail.com> wrote:
>> Eclipse thinks that the collections code has about 1000 instance of
>> @Override of an INTERFACE method, which is supposed to be invalid in
>> 1.5 and valid in 1.6.
>>
>

Re: ]math[ 1.5, really?

Posted by Sean <sr...@gmail.com>.
Yes, that's all correct, what's the issue?

On Mon, Feb 8, 2010 at 2:37 AM, Benson Margulies <bi...@gmail.com> wrote:
> Eclipse thinks that the collections code has about 1000 instance of
> @Override of an INTERFACE method, which is supposed to be invalid in
> 1.5 and valid in 1.6.
>