Getting fedora boost-mingw to include BoostConfig.cmake

[Albert Astals Cid]

How hard would be for mingw64-boost/mingw32-boost to include 
BoostConfig.cmake?

boost-devel does but it's on the wrong "sysroot".

Cheers,
  Albert

[Neal Gompa]

On Sun, Aug 11, 2024 at 7:04 PM Albert Astals Cid  wrote:
> How hard would be for mingw64-boost/mingw32-boost to include
> BoostConfig.cmake?
>
> boost-devel does but it's on the wrong "sysroot".

It shouldn't be hard, currently we're deleting the files instead of
keeping them. I'm not sure if the files are usable, but if you're
willing to test a build, I can attempt to restore them and provide a
test build for you to try?

[Neal Gompa]

On Sun, Aug 25, 2024 at 7:20 AM Neal Gompa  wrote:
> On Sun, Aug 11, 2024 at 7:04 PM Albert Astals Cid  wrote:
> > How hard would be for mingw64-boost/mingw32-boost to include
> > BoostConfig.cmake?
> >
> > boost-devel does but it's on the wrong "sysroot".
>
> It shouldn't be hard, currently we're deleting the files instead of
> keeping them. I'm not sure if the files are usable, but if you're
> willing to test a build, I can attempt to restore them and provide a
> test build for you to try?

For reference, it looks like it was disabled ten years ago with this message:

* Fri Feb 10 2012 Erik van Pienbroek  - 1.48.0-3
- Don't provide the cmake files any more as they are broken and cmake
itself already provides its own boost detection mechanism.
Should fix detection of boost by mingw32-qpid-cpp. RHBZ #597020, RHBZ #789399

Those bugs are:

* https://bugzilla.redhat.com/597020
* https://bugzilla.redhat.com/789399

If CMake has its own way to detect Boost that works for this, is there
a reason you're not using it?

[Albert Astals Cid]

El diumenge, 25 d’agost del 2024, a les 13:23:19 (CEST), Neal Gompa va 
escriure:
> On Sun, Aug 25, 2024 at 7:20 AM Neal Gompa  wrote:
> > On Sun, Aug 11, 2024 at 7:04 PM Albert Astals Cid  wrote:
> > > How hard would be for mingw64-boost/mingw32-boost to include
> > > BoostConfig.cmake?
> > >
> > > boost-devel does but it's on the wrong "sysroot".
> >
> > It shouldn't be hard, currently we're deleting the files instead of
> > keeping them. I'm not sure if the files are usable, but if you're
> > willing to test a build, I can attempt to restore them and provide a
> > test build for you to try?
>
> For reference, it looks like it was disabled ten years ago with this
> message:
> Fri Feb 10 2012 Erik van Pienbroek  -
>
> 1.48.0-3

- Don't provide the cmake files any more as they are broken and
> cmake itself already provides its own boost detection mechanism.
> Should fix detection of boost by mingw32-qpid-cpp. RHBZ #597020, RHBZ
> #789399
> Those bugs are:
> https://bugzilla.redhat.com/597020
> https://bugzilla.redhat.com/789399
>
> If CMake has its own way to detect Boost that works for this, is there
> a reason you're not using it?

Because cmake says you is deprecating its FindBoost.

"Port projects to upstream Boost's BoostConfig.cmake package configuration 
file"

https://cmake.org/cmake/help/latest/module/FindBoost.html

Chers,
  Albert
> -- 
> 真実はいつも一つ！/ Always, there's only one truth!

[Neal Gompa]

On Sun, Aug 25, 2024 at 3:17 PM Albert Astals Cid  wrote:
> El diumenge, 25 d’agost del 2024, a les 13:23:19 (CEST), Neal Gompa va
> escriure:
> > On Sun, Aug 25, 2024 at 7:20 AM Neal Gompa  wrote:
> > > On Sun, Aug 11, 2024 at 7:04 PM Albert Astals Cid  wrote:
> > > > How hard would be for mingw64-boost/mingw32-boost to include
> > > > BoostConfig.cmake?
> > > >
> > > > boost-devel does but it's on the wrong "sysroot".
> > >
> > > It shouldn't be hard, currently we're deleting the files instead of
> > > keeping them. I'm not sure if the files are usable, but if you're
> > > willing to test a build, I can attempt to restore them and provide a
> > > test build for you to try?
> >
> > For reference, it looks like it was disabled ten years ago with this
> > message:
> > Fri Feb 10 2012 Erik van Pienbroek  -
> >
> > 1.48.0-3
> Don't provide the cmake files any more as they are broken and
> > cmake itself already provides its own boost detection mechanism.
> > Should fix detection of boost by mingw32-qpid-cpp. RHBZ #597020, RHBZ
> > #789399
> > Those bugs are:
> > https://bugzilla.redhat.com/597020
> > https://bugzilla.redhat.com/789399
> >
> > If CMake has its own way to detect Boost that works for this, is there
> > a reason you're not using it?
>
> Because cmake says you is deprecating its FindBoost.
>
> "Port projects to upstream Boost's BoostConfig.cmake package configuration
> file"
>
> https://cmake.org/cmake/help/latest/module/FindBoost.html

I've kicked off a build in COPR for this:
https://copr.fedorainfracloud.org/coprs/ngompa/mingw-boost-withcmake/build/7...

Once it's built, you can enable it and try it by doing the following:

$ sudo dnf copr enable ngompa/mingw-boost-withcmake
$ sudo dnf upgrade 'mingw*-boost*'

You can revert by doing the following:

$ sudo dnf copr disable ngompa/mingw-boost-withcmake
$ sudo dnf distro-sync 'mingw*-boost*'

--
真実はいつも一つ！/ Always, there's only one truth!

[Albert Astals Cid]

El dilluns, 26 d’agost del 2024, a les 3:22:49 (CEST), Neal Gompa va escriure:
> On Sun, Aug 25, 2024 at 3:17 PM Albert Astals Cid  wrote:
> > El diumenge, 25 d’agost del 2024, a les 13:23:19 (CEST), Neal Gompa va
> >
> > escriure:
> > > On Sun, Aug 25, 2024 at 7:20 AM Neal Gompa  wrote:
> > > > On Sun, Aug 11, 2024 at 7:04 PM Albert Astals Cid

wrote:
> > > > > How hard would be for mingw64-boost/mingw32-boost to include
> > > > > BoostConfig.cmake?
> > > > >
> > > > > boost-devel does but it's on the wrong "sysroot".
> > > >
> > > > It shouldn't be hard, currently we're deleting the files instead of
> > > > keeping them. I'm not sure if the files are usable, but if you're
> > > > willing to test a build, I can attempt to restore them and provide a
> > > > test build for you to try?
> > >
> > > For reference, it looks like it was disabled ten years ago with this
> > > message:
> > > Fri Feb 10 2012 Erik van Pienbroek  -
> > >
> > > 1.48.0-3
> > Don't provide the cmake files any more as they are broken and
> > > cmake itself already provides its own boost detection mechanism.
> > > Should fix detection of boost by mingw32-qpid-cpp. RHBZ #597020, RHBZ
> > > #789399
> > >
> > > Those bugs are:
> > > https://bugzilla.redhat.com/597020
> > > https://bugzilla.redhat.com/789399
> > >
> > > If CMake has its own way to detect Boost that works for this, is there
> > > a reason you're not using it?
> >
> > Because cmake says you is deprecating its FindBoost.
> >
> > "Port projects to upstream Boost's BoostConfig.cmake package configuration
> > file"
> >
> > https://cmake.org/cmake/help/latest/module/FindBoost.html
>
> I've kicked off a build in COPR for this:
> https://copr.fedorainfracloud.org/coprs/ngompa/mingw-boost-withcmake/build/7
> 942069/
>
> Once it's built, you can enable it and try it by doing the following:
>
> $ sudo dnf copr enable ngompa/mingw-boost-withcmake
> $ sudo dnf upgrade 'mingw*-boost*'
>
> You can revert by doing the following:
>
> $ sudo dnf copr disable ngompa/mingw-boost-withcmake
> $ sudo dnf distro-sync 'mingw*-boost*'

Works perfectly (for my very limited ase of building poppler).

Cheers,
  Albert
> --
> 真実はいつも一つ！/ Always, there's only one truth!

[Neal Gompa]

On Mon, Aug 26, 2024 at 11:17 AM Albert Astals Cid  wrote:
> El dilluns, 26 d’agost del 2024, a les 3:22:49 (CEST), Neal Gompa va escriure:
> > On Sun, Aug 25, 2024 at 3:17 PM Albert Astals Cid  wrote:
> > > El diumenge, 25 d’agost del 2024, a les 13:23:19 (CEST), Neal Gompa va
> > >
> > > escriure:
> > > > On Sun, Aug 25, 2024 at 7:20 AM Neal Gompa  wrote:
> > > > > On Sun, Aug 11, 2024 at 7:04 PM Albert Astals Cid
>
> wrote:
> > > > > > How hard would be for mingw64-boost/mingw32-boost to include
> > > > > > BoostConfig.cmake?
> > > > > >
> > > > > > boost-devel does but it's on the wrong "sysroot".
> > > > >
> > > > > It shouldn't be hard, currently we're deleting the files instead of
> > > > > keeping them. I'm not sure if the files are usable, but if you're
> > > > > willing to test a build, I can attempt to restore them and provide a
> > > > > test build for you to try?
> > > >
> > > > For reference, it looks like it was disabled ten years ago with this
> > > > message:
> > > > Fri Feb 10 2012 Erik van Pienbroek  -
> > > >
> > > > 1.48.0-3
> > > Don't provide the cmake files any more as they are broken and
> > > > cmake itself already provides its own boost detection mechanism.
> > > > Should fix detection of boost by mingw32-qpid-cpp. RHBZ #597020, RHBZ
> > > > #789399
> > > >
> > > > Those bugs are:
> > > > https://bugzilla.redhat.com/597020
> > > > https://bugzilla.redhat.com/789399
> > > >
> > > > If CMake has its own way to detect Boost that works for this, is there
> > > > a reason you're not using it?
> > >
> > > Because cmake says you is deprecating its FindBoost.
> > >
> > > "Port projects to upstream Boost's BoostConfig.cmake package configuration
> > > file"
> > >
> > > https://cmake.org/cmake/help/latest/module/FindBoost.html
> >
> > I've kicked off a build in COPR for this:
> > https://copr.fedorainfracloud.org/coprs/ngompa/mingw-boost-withcmake/build/7
> > 942069/
> >
> > Once it's built, you can enable it and try it by doing the following:
> >
> > $ sudo dnf copr enable ngompa/mingw-boost-withcmake
> > $ sudo dnf upgrade 'mingw*-boost*'
> >
> > You can revert by doing the following:
> >
> > $ sudo dnf copr disable ngompa/mingw-boost-withcmake
> > $ sudo dnf distro-sync 'mingw*-boost*'
>
> Works perfectly (for my very limited ase of building poppler).

Updates proposed:

* F41: https://bodhi.fedoraproject.org/updates/FEDORA-2024-02963dd4a5
* F40: https://bodhi.fedoraproject.org/updates/FEDORA-2024-7869158891

Testing and karma welcome! :)

--
真実はいつも一つ！/ Always, there's only one truth!

[Albert Astals Cid]

El diumenge, 25 d’agost del 2024, a les 13:20:36 (CEST), Neal Gompa va 
escriure:
> On Sun, Aug 11, 2024 at 7:04 PM Albert Astals Cid  wrote:
> > How hard would be for mingw64-boost/mingw32-boost to include
> > BoostConfig.cmake?
> >
> > boost-devel does but it's on the wrong "sysroot".
>
> It shouldn't be hard, currently we're deleting the files instead of
> keeping them. I'm not sure if the files are usable, but if you're
> willing to test a build, I can attempt to restore them and provide a
> test build for you to try?

I could do that, yes.

Cheers,
  Albert
> -- 
> 真実はいつも一つ！/ Always, there's only one truth!