add qt cmake macros #4577
add qt cmake macros #4577
Conversation
|
Failure in build 1 (
Note: To save resources, CI tries to finish as soon as an error is found. For this reason you might find that not all the references have been launched or not all the configurations for a given reference. Also, take into account that we cannot guarantee the order of execution as it depends on CI workload and workers availability. |
|
Failure in build 3 (
Note: To save resources, CI tries to finish as soon as an error is found. For this reason you might find that not all the references have been launched or not all the configurations for a given reference. Also, take into account that we cannot guarantee the order of execution as it depends on CI workload and workers availability. |
and remove unused cmake files
|
Failure in build 4 (
Note: To save resources, CI tries to finish as soon as an error is found. For this reason you might find that not all the references have been launched or not all the configurations for a given reference. Also, take into account that we cannot guarantee the order of execution as it depends on CI workload and workers availability. |
|
I am clueless as well but recall seeing this error some time ago 🤔 |
|
Failure in build 5 (
Note: To save resources, CI tries to finish as soon as an error is found. For this reason you might find that not all the references have been launched or not all the configurations for a given reference. Also, take into account that we cannot guarantee the order of execution as it depends on CI workload and workers availability. |
|
@danimtb could it be linked to conan-io/conan#6379 or conan-io/conan#8172 ? |
|
Failure in build 6 (
Note: To save resources, CI tries to finish as soon as an error is found. For this reason you might find that not all the references have been launched or not all the configurations for a given reference. Also, take into account that we cannot guarantee the order of execution as it depends on CI workload and workers availability. |
blocked by conan-io#4577
|
Failure in build 7 (
Note: To save resources, CI tries to finish as soon as an error is found. For this reason you might find that not all the references have been launched or not all the configurations for a given reference. Also, take into account that we cannot guarantee the order of execution as it depends on CI workload and workers availability. |
|
@danimtb and the curse goes on |
|
took the erratic machine offline to avoid the error ATM. Let's see if that works |
|
All green in build 8 (
|
| "set(Qt5Core_QMAKE_EXECUTABLE ${CMAKE_CURRENT_LIST_DIR}/../../../bin/qmake)\n" | ||
| "set(Qt5Core_MOC_EXECUTABLE ${CMAKE_CURRENT_LIST_DIR}/../../../bin/moc)\n" | ||
| "set(Qt5Core_RCC_EXECUTABLE ${CMAKE_CURRENT_LIST_DIR}/../../../bin/rcc)\n" | ||
| "set(Qt5Core_UIC_EXECUTABLE ${CMAKE_CURRENT_LIST_DIR}/../../../bin/uic)") |
There was a problem hiding this comment.
TLDR: Would it be ok to provide cmake targets in Qt5:: namespace (using this build module) whereas conan generated target are in the qt:: namespace ?
Director's cut: With the change as it is now, the user has to manually call qt5_wrap_cpp on the sources in the CMakeLists.txt. Most projects do not use this, but instead rely on CMAKE_AUTOMOC, which automatically invokes MOC on sources of all targets. This currently does not work, because cmake tries to invoke the target Qt5::moc, which does not exist. Would it be fine if I added this add_executable(Qt5::moc IMPORTED); set_target_properties(Qt5::moc PROPERTIES IMPORTED_LOCATION ${Qt5Core_QMAKE_EXECUTABLE}) ? I'm not sure because all the targets generated by conan are in the qt namespace, not in the Qt5 namespace, and I don't want to make this shift without components support.
WDYT ?
There was a problem hiding this comment.
I think this is what is making my PR fail
#4562
There was a problem hiding this comment.
yes: take a look at ericriff#5
|
Where is my merge?! Where is my merge?! 😆 |
According to my bot... and that's a LOT ready. |
|
Nice to see there are so many packages ready 🎉 |
|
This build should take ~11hrs based on the last one [
{
"url": "https://api.github.com/repos/conan-io/conan-center-index/statuses/d5549e2d0ff2ca744828a00305eef392173ce9c5",
"avatar_url": "https://avatars.githubusercontent.com/u/54393557?v=4",
"id": 12192890712,
"node_id": "MDEzOlN0YXR1c0NvbnRleHQxMjE5Mjg5MDcxMg==",
"state": "success",
"description": "This commit looks good",
"target_url": "https://ci-conan-prod.jfrog.team/job/cci/job/PR-4577/8/display/redirect",
"context": "continuous-integration/jenkins/pr-merge",
"created_at": "2021-02-16T21:58:36Z",
"updated_at": "2021-02-16T21:58:36Z",
"creator": {
"login": "conan-center-bot",
}
},
{
"url": "https://api.github.com/repos/conan-io/conan-center-index/statuses/d5549e2d0ff2ca744828a00305eef392173ce9c5",
"avatar_url": "https://avatars.githubusercontent.com/u/54393557?v=4",
"id": 12185253569,
"node_id": "MDEzOlN0YXR1c0NvbnRleHQxMjE4NTI1MzU2OQ==",
"state": "pending",
"description": "This commit is being built",
"target_url": "https://ci-conan-prod.jfrog.team/job/cci/job/PR-4577/8/display/redirect",
"context": "continuous-integration/jenkins/pr-merge",
"created_at": "2021-02-16T10:59:24Z",
"updated_at": "2021-02-16T10:59:24Z",
"creator": {
"login": "conan-center-bot",
}
}
] |
|
We cannot afford building after merging to |
|
We would probably still need to rebuild anyway after merge, because the dependencies could have been updated in the mean-time too, right ? |
That's another big challenge and, IMO, it should be addressed the other way around: rebuild all the consumers each time we merge a package to Nevertheless, I'm not sure if it would require too much computational power. The C++ graph is not very big, but it is very connected (link), I'm not sure if we can afford to build 10-20 references each time we merge a change to some libraries... 😓 |
|
Random thought: I have no clue how much free capacity CCI has on average (or in the future if CCI growths further), but maybe some not-that-urgent rebuilds could be added to an "only build it, if there is free capacity" waiting list? |
|
TBH we are not at that point yet. I would say that we have different objectives with different priorities (not in order):
I'd like to think that we are improving in all those areas, some of them are more visible and others are only appreciated by the internal team. What I'm sure of is that we can work on those objectives only because other people are doing an amazing job here. If we were required to review every PR we wouldn't be able to work on the internals. |
Correct me if I am missing something but CCI builds are not locked to a recipe revision. By the time a consumer gets it it might be too late and changed again anyways? 💭 The package version can not change in the merge commit... I am not sure it adds much benefit, in fact it might introduce instability if the targets changed.
I really appreciate the work from the entire Conan, it's truly making a difference about how we look at OSS and the few teams using it have really be able to deliver massive new technologies with a low velocity. We might not "see" the changes but we certainly feel them. So thank you 🤗 |
* Add QuaZip 1.1 * Projects that link against Qt5 need c++ 11 * invoke moc manually blocked by #4577 * Fix licence name Co-authored-by: Uilian Ries <uilianries@gmail.com> Co-authored-by: Eric Lemanissier <eric.lemanissier@gmail.com> Co-authored-by: Uilian Ries <uilianries@gmail.com>
fixes #4574
Specify library name and version: qt/5.15.2
conan-center hook activated.