Skip to main content

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] [List Home]
Re: [cross-project-issues-dev] Tycho target environments

As I understand it, each of those combinations is relevant if Maven/Tycho needs to run on that environment, if you need to produce fragments/products for that environment, or if you need to compose such fragments into your update site, i.e., each is relevant if that combination is actually used/needed.

One way to look at it is to ask, what happens if the fragment/IU for any particular combination is missing?  E.g., locally on Windows your build will need the win32 fragments of SWT even just to compile anything that depends on SWT.  But on Windows, if the mac or linux fragments are missing, most likely nothing will care, unless of course you also are trying to include all these fragments in your update site or trying to build products for multiple platforms and want to include a product for that combination.

This is why the Platform can add things like linux riscv64 or ppc64le without anyone else necessarily taking action.  EPP for example now provides a linux package for riscv64 but not for ppc64le.

https://github.com/eclipse-packaging/packages/blob/a1bdda42da3bd7c4d781adf225cb8bc46f6b3c66/releng/org.eclipse.epp.config/parent/pom.xml#L130-L172

So it all boils down to which fragments you need to consume and/or produce as well as which fragments you need in order to run on the build on  host machine; ci instances are generally linux/gtk/x86_64 by default, so you can't leave that out.    There really aren't many fragment producers on the train.

Keep in mind too that EMF is consumed by the Platform and still the Platform is able to provide support for os/ws/arch combinations so no wrist slapping was or is required (and that pure Java can run anywhere that a JVM with the required EE and packages exists).

On 28.12.2024 07:36, Edward Willink via cross-project-issues-dev wrote:

Hi

While revising pom.xml's for Java 21, I noticed the following possibly stale target-platform-configuration section

<environments>

<!--environment>

<os>win32</os>

<ws>win32</ws>

<arch>x86</arch>

</environment-->

<environment>

<os>win32</os>

<ws>win32</ws>

<arch>x86_64</arch>

</environment>

<!--environment>

<os>linux</os>

<ws>gtk</ws>

<arch>x86</arch>

</environment-->

<environment>

<os>linux</os>

<ws>gtk</ws>

<arch>x86_64</arch>

</environment>

<environment>

<os>macosx</os>

<ws>cocoa</ws>

<arch>x86_64</arch>

</environment>

</environments>

The commenting was necessitated by the removal of 32 bit support.

However IIRC we now have ARM support and https://download.eclipse.org/eclipse/downloads/drops4/R-4.34-202411201800/ identifies macosx.cocoa.aarch64 but no linux equivalent.

The table in https://eclipse.dev/eclipse/development/plans/eclipse_project_plan_4_34.xml#target_environments is not clear to me, but seems to suggest aarch64 for Ubuntu Long Term Support. (It would help if there was a mapping to the Tycho target-platform-configuration spellings.)

Looking for an exemplary project to copy, I find that EMF has just the same three x86_64 environments.

Are these environment settings irrelevant, or are EMF and my projects and no doubt many others needlessly excluding their usage with some supported Eclipse environments?

If so, perhaps the SimRel aggregator needs to start slapping wrists.

Regards

Ed Willink





_______________________________________________
cross-project-issues-dev mailing list
cross-project-issues-dev@xxxxxxxxxxx
To unsubscribe from this list, visit https://www.eclipse.org/mailman/listinfo/cross-project-issues-dev

Back to the top