Bug 2381102
| Summary: | portmidi: FTBFS with change proposal CMake: Use ninja generator by default | ||
|---|---|---|---|
| Product: | [Fedora] Fedora | Reporter: | Cristian Le <fedora> |
| Component: | portmidi | Assignee: | Michael J Gruber <mjg> |
| Status: | CLOSED ERRATA | QA Contact: | Fedora Extras Quality Assurance <extras-qa> |
| Severity: | unspecified | Docs Contact: | |
| Priority: | unspecified | ||
| Version: | rawhide | CC: | mjg, oget.fedora, xavier |
| Target Milestone: | --- | ||
| Target Release: | --- | ||
| Hardware: | Unspecified | ||
| OS: | Unspecified | ||
| Whiteboard: | |||
| Fixed In Version: | portmidi-217-75.fc43 | Doc Type: | --- |
| Doc Text: | Story Points: | --- | |
| Clone Of: | Environment: | ||
| Last Closed: | 2025-07-17 16:00:24 UTC | Type: | --- |
| Regression: | --- | Mount Type: | --- |
| Documentation: | --- | CRM: | |
| Verified Versions: | Category: | --- | |
| oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |
| Cloudforms Team: | --- | Target Upstream Version: | |
| Embargoed: | |||
| Bug Depends On: | |||
| Bug Blocks: | 2376112 | ||
|
Description
Cristian Le
2025-07-16 15:21:39 UTC
I have a fix here: https://src.fedoraproject.org/fork/mjg/rpms/portmidi/c/66c2c18cbf122575df5d6cc3109cd2a99d8276e3?branch=rawhide It builds both on rawhide (with make) and lecris/cmake-ninja (with ninja). BUT: Once I up the cmake minimum version for the CMake 4.0 proposal as I do here: https://src.fedoraproject.org/fork/mjg/rpms/portmidi/c/584568a8ffcd2dcee2ed5087d47b48ec5046bd04?branch=rawhide Then the build works on rawhide and lecris/cmake-4.0, but not any more on lecris/cmake-ninja because ninja suddenly fails to find a file. So, the change proposals appear to be interdependent, and a fix to one may break an apparant fix to another one. And there's a third one, of course, for non-standard variables. That is very odd. First thing that I notice, why did you add ``` %global __cmake_in_source_build 1 ``` That would likely break a lot of stuff. `CMAKE_CACHEFILE_DIR` is also a very weird one that needs checking. The documentation mentions it's not supposed to be writeable. It should expand to `%_vpath_builddir` if you need it anywhere else. Do you have a log of how it breaks? Also very much appreciate the notes on the policy incompatibility. FEDORA-2025-ab78c5fd26 (portmidi-217-75.fc43) has been submitted as an update to Fedora 43. https://bodhi.fedoraproject.org/updates/FEDORA-2025-ab78c5fd26 FEDORA-2025-ab78c5fd26 (portmidi-217-75.fc43) has been pushed to the Fedora 43 stable repository. If problem still persists, please make note of it in this bug report. Portmidi (1) is basically a stale ancient codebase, so I'm going for minimal fixes rather than trying to clean everything up. For example, the cmake setup for the java part uses custom commands which - as such - work for an in-source build only. That's why I enforce that now in the proper way (it used to be "enforced" with "-B." and a direct make call). Adjusting the version range makes cmake behave differently, so that should probably be the first change which one works out. Just as a hint :) Ah, I see that upstream has changed to portmidi2. Good luck with that transition :) I guess you've found a workaround, whatever works I guess :) |