You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@cxf.apache.org by Benson Margulies <bi...@gmail.com> on 2012/01/03 18:45:42 UTC
2.5.1 codegen issues
So I apparently busted some things in the codegen plugin for 2.5.1.
Unfortunately, I've been sick for a week, and that really puts a chomp
into my ability to roll my sleeves up. My strong belief is that we
need proper plugin integration tests, the same way that other maven
plugins deal with this -- the maven-invoker-plugin. Has anyone else
got time to start on this? If not, I'll get onto it when I can.
Re: 2.5.1 codegen issues
Posted by Daniel Kulp <dk...@apache.org>.
On Tuesday, January 03, 2012 5:00:57 PM Benson Margulies wrote:
> On Tue, Jan 3, 2012 at 3:57 PM, Daniel Kulp <dk...@apache.org> wrote:
> >
> > Well, CXF-4002 is not. No idea on that one yet. The fact that it's
> > only in Sonar makes me really question it.
>
> I think that this is a Sonar bug, perhaps exposed by my cleanup of
> that parameter. I'd love to hear from someone sonarial. Sonorus?
It gets even a bit stranger as the CXF build itself at:
https://analysis.apache.org/dashboard/index/5429
and the Camel build there, which both run the CXF codegen plugin, are fine and
run to completion. Thus, it may be a Sonar VERSION issue or maybe and issue
with the version of Maven they have their Sonar build configured to use.
--
Daniel Kulp
dkulp@apache.org - http://dankulp.com/blog
Talend Community Coder - http://coders.talend.com
Re: 2.5.1 codegen issues
Posted by Benson Margulies <bi...@gmail.com>.
On Tue, Jan 3, 2012 at 3:57 PM, Daniel Kulp <dk...@apache.org> wrote:
> On Tuesday, January 03, 2012 3:54:13 PM Daniel Kulp wrote:
>> On Tuesday, January 03, 2012 12:45:42 PM Benson Margulies wrote:
>> > So I apparently busted some things in the codegen plugin for 2.5.1.
>> > Unfortunately, I've been sick for a week, and that really puts a chomp
>> > into my ability to roll my sleeves up. My strong belief is that we
>> > need proper plugin integration tests, the same way that other maven
>> > plugins deal with this -- the maven-invoker-plugin.
>>
>> +1
>>
>> > Has anyone else
>> > got time to start on this? If not, I'll get onto it when I can.
>>
>> I've never looked into using the maven-invoker-plugin to run it tests. I
>> was going to grab olamy to help walk me through it, but he ended up having
>> an additional child suddenly appear yesterday and is now on full time
>> "daddy duty" for a bit. :-) (congrats to him, btw) I may have some
>> time next week, but right now, no.
>>
>> That said, I just looked at the issues and the issues are all due to the
>> exact same issue and I'm testing a fix for it now.
>
> Well, CXF-4002 is not. No idea on that one yet. The fact that it's only in
> Sonar makes me really question it.
I think that this is a Sonar bug, perhaps exposed by my cleanup of
that parameter. I'd love to hear from someone sonarial. Sonorus?
>
> Dan
>
>
>
>> Basically, the
>> defaultOptions were being merged in twice in SOME cases. In the directory
>> scan, it used the defaultOptions to initialize the options, but then a merge
>> was done in mergeOptions and the extraArgs would get doubled.
>>
>> I ended up taking out the passing of the defaultOptions into the various
>> scans and such and letting just the mergeOptions call handle it. That
>> allows it to work for the scan, the explicitly set stuff, and the wsdl dep
>> case.
> --
> Daniel Kulp
> dkulp@apache.org - http://dankulp.com/blog
> Talend Community Coder - http://coders.talend.com
Re: 2.5.1 codegen issues
Posted by Daniel Kulp <dk...@apache.org>.
On Tuesday, January 03, 2012 3:54:13 PM Daniel Kulp wrote:
> On Tuesday, January 03, 2012 12:45:42 PM Benson Margulies wrote:
> > So I apparently busted some things in the codegen plugin for 2.5.1.
> > Unfortunately, I've been sick for a week, and that really puts a chomp
> > into my ability to roll my sleeves up. My strong belief is that we
> > need proper plugin integration tests, the same way that other maven
> > plugins deal with this -- the maven-invoker-plugin.
>
> +1
>
> > Has anyone else
> > got time to start on this? If not, I'll get onto it when I can.
>
> I've never looked into using the maven-invoker-plugin to run it tests. I
> was going to grab olamy to help walk me through it, but he ended up having
> an additional child suddenly appear yesterday and is now on full time
> "daddy duty" for a bit. :-) (congrats to him, btw) I may have some
> time next week, but right now, no.
>
> That said, I just looked at the issues and the issues are all due to the
> exact same issue and I'm testing a fix for it now.
Well, CXF-4002 is not. No idea on that one yet. The fact that it's only in
Sonar makes me really question it.
Dan
> Basically, the
> defaultOptions were being merged in twice in SOME cases. In the directory
> scan, it used the defaultOptions to initialize the options, but then a merge
> was done in mergeOptions and the extraArgs would get doubled.
>
> I ended up taking out the passing of the defaultOptions into the various
> scans and such and letting just the mergeOptions call handle it. That
> allows it to work for the scan, the explicitly set stuff, and the wsdl dep
> case.
--
Daniel Kulp
dkulp@apache.org - http://dankulp.com/blog
Talend Community Coder - http://coders.talend.com
Re: 2.5.1 codegen issues
Posted by Daniel Kulp <dk...@apache.org>.
On Tuesday, January 03, 2012 12:45:42 PM Benson Margulies wrote:
> So I apparently busted some things in the codegen plugin for 2.5.1.
> Unfortunately, I've been sick for a week, and that really puts a chomp
> into my ability to roll my sleeves up. My strong belief is that we
> need proper plugin integration tests, the same way that other maven
> plugins deal with this -- the maven-invoker-plugin.
+1
> Has anyone else
> got time to start on this? If not, I'll get onto it when I can.
I've never looked into using the maven-invoker-plugin to run it tests. I was
going to grab olamy to help walk me through it, but he ended up having an
additional child suddenly appear yesterday and is now on full time "daddy
duty" for a bit. :-) (congrats to him, btw) I may have some time next
week, but right now, no.
That said, I just looked at the issues and the issues are all due to the
exact same issue and I'm testing a fix for it now. Basically, the
defaultOptions were being merged in twice in SOME cases. In the directory
scan, it used the defaultOptions to initialize the options, but then a merge
was done in mergeOptions and the extraArgs would get doubled.
I ended up taking out the passing of the defaultOptions into the various scans
and such and letting just the mergeOptions call handle it. That allows it to
work for the scan, the explicitly set stuff, and the wsdl dep case.
--
Daniel Kulp
dkulp@apache.org - http://dankulp.com/blog
Talend Community Coder - http://coders.talend.com