* Add install-qt-commercial feature and tests
* Make the auto-answers parameters, fix linter issues
* Fork and execv instead of using subprocess
* Return to simpler process execution method version
* Fix test
* Move commercial installer into its own file
* Fix shadowing of symbol platform causing errors
* Adapt test_cli for argparse format changes on py 3.13+
* Fix some errors, monkeypatch install test
* Add --override super command
* Properly handle --override and grab all the remaining commands when no quotes are given
* Fix tests
* Add base for modules, some niche features are not yet entirely implemented, and there are no updates to the testsuite
* Fix some mistakes
* Fix errors made with the monkeypatch, update Settings to make sure its init
* Tests commercial (#20)
* Full support of installation of all modules and addons
* Add auto setup of cache folder for each OS, add unattended parameter
* Fix settings folders
* Add graceful error message for overwrite case
* Fix windows issue
* Hidden summon works
* Remove both subprocess direct calls
* Dipose of temp folder
* Fix path issue
* Add list-qt-commercial command
* Fix help info
* Make no params valid for list-qt-commercial
* Fix lint errors, and param overflow when no args are passed to list
* Fix search
* Add tests for coverage, fix lint
* Test for overwriting, and for cache usage coverage
* Return to clean exec, ignoring CI fail to preserve code clarity
* Fix parsing of subprocess.run output for some python versions
* Make output more readable in console for list-qt-commercial
* Forward email and password to list request for users without a qtaccount.ini
* Change default settings
* Fix lint errors
* Fix check error
* Rewrite WASM support
* Add WASM tests to CI, update CI to test more the latest versions, add auto EMSDK version detection function
* Fix some mistakes, typos, moved emsdk version function into BuildJob
* Fix issue related to extensions interfering with wasm on 6.8+
* Fix tests
* Remove dep on Version in CI
* Remove safety before patch
* handle cases where extensions don't exist.
for example with windows desktop 6.8.1 win64_msvc2022_arm64_cross_compiled
both qtwebengine and qtpdf don't exist.
Signed-off-by: Alexandre 'Kidev' Poumaroux <1204936+Kidev@users.noreply.github.com>
* for --long-modules assume extension doesn't exist on download error.
Signed-off-by: Alexandre 'Kidev' Poumaroux <1204936+Kidev@users.noreply.github.com>
* for --modules assume extension doesn't exist for download failures.
Signed-off-by: Alexandre 'Kidev' Poumaroux <1204936+Kidev@users.noreply.github.com>
* reformat with black
Signed-off-by: Alexandre 'Kidev' Poumaroux <1204936+Kidev@users.noreply.github.com>
* fix flake8 regression that doesn't occur locally.
Signed-off-by: Alexandre 'Kidev' Poumaroux <1204936+Kidev@users.noreply.github.com>
* Fix autodesktop by also updating the OS when searching for a valid desktop version to download
* Fix extension issue, reduce the possible retry for getting extensions to prevent server spam
* Fix CI asking for msvc2019 on 6.8+ but its no longer supported
* Make CI use C++20 and MSVC2022
* Fix linux build
* Update runners to windows-2022
Signed-off-by: Alexandre 'Kidev' Poumaroux <1204936+Kidev@users.noreply.github.com>
* Fix patching
Signed-off-by: Alexandre 'Kidev' Poumaroux <1204936+Kidev@users.noreply.github.com>
* Add back the semantic version changes to prevent crashes, add tests for it
* Update checks
* Cast 'https://mirrors.ustc.edu.cn' to the shadow realm
* Again
* Update settings.ini
* Update settings.ini
* Update settings.ini
* Remove one_rep on silent
* Update settings.ini
* Restore master settings, remove hash check
* ci: Use specific mirror
Attempt to work around download errors in Azure due to Qt's official download site often redirecting to mirrors to which the network connection is unstable
Signed-off-by: Alexandre 'Kidev' Poumaroux <1204936+Kidev@users.noreply.github.com>
* Re enable hash checking
* Treat read timeout error during download as connection error
Signed-off-by: Alexandre 'Kidev' Poumaroux <1204936+Kidev@users.noreply.github.com>
* Add test for modules in WASM with autodesktop
* Fix format
* Fix test
* Make '--autodesktop' trigger its own install process, and test it
* Fix older autodesktop tests
* Add mock update files for 680 wasm, add test for wasm 680 autodesktop
* Passes the additional tests
* Fix format
* Improve coverage, fix format
* Fix tests and improve logging or install
* Fix format
* Fix regression in other tests
* Use flavor
* Fix line len
* Fix codeql
* Fix list-qt for WASM arch on 6.5.x and 6.6.x, restore to original download URL
* Fix test error
* Revert ci settings URL as it is never used by clients, only in CI
* Add comment for clarity in ci/settings.ini
---------
Signed-off-by: Alexandre 'Kidev' Poumaroux <1204936+Kidev@users.noreply.github.com>
Co-authored-by: tsteven4 <13596209+tsteven4@users.noreply.github.com>
Co-authored-by: J.D. Purcell <jdpurcell@gmail.com>
This improves the warnings about unknown versions, modules, and
architectures of Qt, so that it is more explicit what the message
actually means.
This changes the purpose of `Cli._check_modules_arg` from checking if
modules exist in combinations.json, to returning a list of modules
that do not exist in that file. The function has been renamed as well.
This change was necessary to make the warning message more informative.
This will allow testing multiple qt installations, as required by the
feature that installs the default desktop qt where required
Add autodesktop test for ios
`MetadataFactory.fetch_http` must often download HTML pages, not
Updates.xml files. download.qt.io does not store checksums for these
files, so this particular function must be allowed to download these
pages without using a checksum.
This modifies `aqt tool` to allow `-` characters in tool versions, so
that `aqt tool` can pass the same unit tests that `install-tool` can.
This makes it less like the `aqt tool` from v1.2.5, but it should work
better with more tools with different version formats.
In an earlier PR, I added calls to `disable_socket()` from
`pytest_socket` where I thought they were needed to prevent some tests
from accessing the network, in case they weren't monkeypatched properly.
Today, I discovered that `disable_socket()` disables sockets globally
for all tests, which means that the tests that use remote data cannot
run if they are executed after another test calls `disable_socket()`.
This change calls `disable_socket()` once from `conftest.py`, so that
no tests are allowed to use network data unless they are marked as ok
to use the network, with `@pytest.mark.enable_socket`. See example of
usage in `tests/test_connection.py`.
Changed return code for unexpected exceptions:
This allows us to write tests that can discover whether an unexpected
exception occurred just by checking the return code, rather than reading
stderr. This will allow us to write less friable tests that don't break
every time some insignificant output details change.
This change catches exceptions derived from Exception and
KeyboardInterrupt raised by `installer`, while run by multiple
processes, and propagates them back to earlier stack entries. This will
prevent any OSError and BrokenPipe exceptions that would otherwise be
raised when one process has an exception while the other processes are
still running.
This also handles the MemoryError exception we saw in #416, and offers
some suggestions for solving the issue.
This recreates the legacy behavior, where `aqt tool` would fail to
install a tool variant when the version in the xml file does not match
the version requested in cli arguments.
This also recreates a bug where `aqt tool` cannot install tools that
do not have a valid semantic version.
`test_cli_legacy_commands_with_wrong_syntax` relies on `argparse` to
filter out bad argument lists; the new test uses a mocked getUrl that
raises ArchiveDownloadError.
