Is there an easier way to "fix" a control instead of building everything (i.e., java, native, etc.)?

[Martin Fox]

The most likely reason for a Mac Glass compilation error is using Xcode 16 to build an older version of JavaFX (anything before PR #1812 was integrated in May of this year). Xcode 16 enabled a new check which caught an old type mismatch error in GlassTouches.m.

If that’s what you’re seeing you can either grab the latest master or just cut out enough code in GlassTouches.m to get it to compile. The event tap that drove GlassTouches was disabled years ago so it’s all dead code. The May PR removed it entirely.

Martin

> On Jul 23, 2025, at 6:34 AM, Kevin Rushforth <kevin.rushforth at oracle.com> wrote:
> 
> [redirecting to openjfx-dev]
> 
> You will need to build all of JavaFX using gradle.
> 
> You might be running into an issue with the version of Xcode you are using. You will need macOS 14 or 15 with a recent version of Xcode. We use Xcode 15.4 for our production builds.
> 
> Can you list the version of macOS and Xcode you are using? And copy/paste the specific error message(s) you see?
> 
> -- Kevin
> 
> 
> On 7/22/2025 8:21 PM, Paul Furbacher wrote:
>> I want to investigate fixing what I consider a bug or deficiency in a particular control in the easiest manner possible.
>> 
>> Initially, I thought to just build everything, but that hasn’t worked because I run into an issue with native compilation, notably in “:graphics:ccMacGlass”.
>> 
>> Instead of battling gradle to build the whole shebang, it seems to me it would be much quicker to simply patch an existing distribution. Browsing “build.gradle”, nothing jumps out at me.
>> 
>> Is it possible to just patch a published distribution with a “fixed” javafx.controls jar?
>> 
>> (Leave jmods out of the picture for the moment because I’ll initially use javafx on the classpath. If my idea of a fix pans out, then I’d proceed with filing a bug report with suggested changes, and maybe doing a pull request — that’s for later.)
>> 
>> If that’s not possible, I might be back here asking for help diagnosing what’s causing clang to finish with a non-zero exit value
>> 
>> 
>> Regards,
>> 
>> Paul Furbacher
>> 
>