Glad to hear that.
Remko
On Wed, May 23, 2018 at 1:55 Russel Winder wrote:
> Hi,
>
> All Gant master/HEAD tests pass with Groovy 2.5.0-rc-3.
>
> :-)
>
> --
> Russel.
> ==
> Dr Russel Winder t: +44 20 7585 2200
> 41 Buckmaster Roadm: +44 7770 465 0
Hi,
All Gant master/HEAD tests pass with Groovy 2.5.0-rc-3.
:-)
--
Russel.
==
Dr Russel Winder t: +44 20 7585 2200
41 Buckmaster Roadm: +44 7770 465 077
London SW11 1EN, UK w: www.russel.org.uk
signature.asc
Description: This is a digitally s
On Fri, 2018-05-18 at 22:29 +0900, Remko Popma wrote:
[…]
> Just out of interest, will Gant migrate to groovy.cli.commons.CliBuilder or
> groovy.cli.picocli.CliBuilder?
Well on the one hand, Gant is a dead project that no-one cares about. It
hasn't had a release since Codehaus died, mainly because
Paul changed `groovy.util.CliBuilder` to delegate to
`groovy.cli.commons.CliBuilder` just now.
That should get rid of the reported error.
groovy.util.CliBuilder is now deprecated.
Just out of interest, will Gant migrate to groovy.cli.commons.CliBuilder or
groovy.cli.picocli.CliBuilder?
Remko
So do you want to change this back for RC3? I know it's late to change
this, but arguably it was late to change it in RC2 too.
-Keegan
On Thu, May 10, 2018, 11:08 AM Remko Popma wrote:
> I updated the PR to make the signature of `setPosix` accept Boolean
> instead of boolean and added tests t
I updated the PR to make the signature of `setPosix` accept Boolean instead
of boolean and added tests to ensure that nulls are handled correctly.
Thanks for pointing that out!
The picocli version of CliBuilder has tests to verify that Java-like
properties such as -Dkey=value are processed correct
Here's all the differences between 2.4 and 2.5
- setFormatter(org.apache.commons.cli.HelpFormatter) (in the PR)
- setParser(org.apache.commons.cli.CommandLineParser) (in the PR)
- setOptions(org.apache.commons.cli.Options)
- setPosix(java.lang.Boolean) (changed from boolean to Boo
Short answer is we can have as many RCs as needed to get it right but we
don't really want to be chopping and changing too much. Big changes are
supposed to be settling down by now - fate just brought this change late in
the game compared to what we'd like. I'd really like to get +ve feedback
from
I realize now that simply removing the `parser` and `formatter` properties
was a mistake. While the picocli version of CliBuilder cannot use
commons-cli classes I should have anticipated that some applications are
setting these properties.
I believe that I have a solution that provides a smoother
Both CliBuilder implementations are there via their full package names, so
encouraging people to use the non-delegated implementation of their choice
is a good option. They will need to get used to such changes in Groovy 3
anyway.
For Groovy 3, we might have a way for the compiler to translate fro
I'll take a look. I'll also see if there are other places that can break.
If there are several ways it can break, then maybe we should not use a
delegate to picocli and instead have folks switch the CliBuilder instance
they instantiate.
-Keegan
On Sun, May 6, 2018, 7:56 PM Remko Popma wrote:
I think I found a way to fix this.
See https://github.com/apache/groovy/pull/696
This PR adds CliBuilder.setParser and CliBuilder.setFormatter methods that
ignore the specified value and print a warning to stderr.
On Sun, May 6, 2018 at 9:14 PM, Remko Popma wrote:
> In 2.5.0-rc-2, groovy.util.Cl
In 2.5.0-rc-2, groovy.util.CliBuilder delegates to
groovy.cli.picocli.CliBuilder. The error is that the `parser` property of
this class is no longer writable.
You can resolve this with 2.5.0-rc-2 by either not setting the `parser`
property in the CliBuilder constructor or using the
groovy.cli.comm
FYI 2.5.0-rc-2 breaks Gant. Specifically, it's caused by changing
groovy.util.CliBuilder to use Picocli
java.lang.reflect.InvocationTargetException
at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
at
sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImp
14 matches
Mail list logo
