Thanks Jonny, I was hoping you would have feedback😃.  I appreciate the
response and hopefully this is something that can be checked in / provided
in the future.

Regards,
James

On Mon, Apr 13, 2026 at 6:12 PM Jonny <[email protected]> wrote:

> Pardon the multiple messages, but I went ahead and just opened a PR.
> https://github.com/apache/groovy-geb/pull/323
>
> Comments welcome!
>
> On Mon, Apr 13, 2026 at 4:53 PM Jonny <[email protected]> wrote:
>
>> Also, to James's idea of publishing the skill on skills.sh, it looks like
>> the Apache Beam folks put their skills in a .agent/skills directory within
>> their repository:
>> https://github.com/apache/beam/tree/master/.agent/skills. Perhaps we
>> could do the same in Groovy and Geb?
>>
>> I know the Spock maintainers have also claimed space on Context7 (
>> https://github.com/spockframework/spock/commit/6eafb3de4c4042c789d2f508c42808d0b15e38ef
>> ).
>>
>> On Mon, Apr 13, 2026 at 4:48 PM Jonny <[email protected]> wrote:
>>
>>> I added a few more tips that I've picked up over the years.
>>> https://gist.github.com/jonnybot0/dcb7fb817ae6c1860eaec164391b49b7
>>>
>>> Most are additions (ways to make your test not flaky, link to the Geb
>>> docs), but there is one place that I pushed back hard against James's
>>> clanker:
>>>
>>> > *Overusing required: false* on optional content - the only time you
>>> really want to mark a page element as required: false is when your spec
>>> *needs* to try to interact with it when it's absent (for example, to
>>> assert that it isn't present, !page.buttons.sometimesThereButton). If
>>> the button may or may not be there, but you never test the case where it
>>> isn't there, you should just leave it as required. Remember, throwing an
>>> exception when something *exceptional* happens is okay, especially in
>>> tests!
>>>
>>> This was something that was drilled into me by Marcin, the former
>>> maintainer of Geb *and* hard experience. I saw more than one case where
>>> someone added `required: false` as a way to address flakiness in a test
>>> that only *hid* the flakiness and moved it downstream. And that someone
>>> was, often enough, me. AIs are even more prone to this kind of quick-fix,
>>> unhelpfully-defensive thinking, so it's probably best if we ward them off
>>> it out the gate.
>>>
>>> Best,
>>>
>>> Jonny
>>>
>>> On Mon, Apr 13, 2026 at 11:08 AM James Daugherty via dev <
>>> [email protected]> wrote:
>>>
>>>> This was generated by an AI, but it is probably a good starting point:
>>>> https://gist.github.com/jdaugherty/f63781ff72c826b14f20fc3a2a41020e
>>>> If anyone has feedback, it would be most welcome.
>>>>
>>>> I know that some people are using services like https://skills.sh/ to
>>>> index skills.  Creating a dedicated repo for skills may be useful for
>>>> Groovy / Geb.
>>>>
>>>> -James
>>>>
>>>> On Mon, Apr 13, 2026 at 8:00 AM Jochen Theodorou <[email protected]>
>>>> wrote:
>>>> >
>>>> > On 4/12/26 19:44, James Daugherty via dev wrote:
>>>> > > Hi Everyone,
>>>> > >
>>>> > > The Grails project has been gradually expanding its Geb test
>>>> coverage
>>>> > > and as we've reviewed the old tests / improved on them, it's become
>>>> > > clear there are some best practices with Geb that we didn't always
>>>> > > follow.  This got me thinking: in the age of AI, the Grails team has
>>>> > > discussed including AI skills (https://agentskills.io/home) as
>>>> part of
>>>> > > our development process to better help adoption.  Has such a topic
>>>> > > been discussed for Geb?  Does anyone have a good starting skill for
>>>> > > Geb?
>>>> > Not me, but I would say to just write something and then let other
>>>> > people look over it to improve it.
>>>> >
>>>> > bye Jochen
>>>>
>>>

Reply via email to