- Trying to get away from untyped strings with special meaning.
- The "split identifier" branch tried the same thing, but
was duplicating the existing InstanceKey.h work.
- If a module exists, and has unmet dependencies, then
that is only a problem if the module itself is *used*.
Merely existing is ok.
This triggers on FreeBSD, where partition isn't built, but
bootloader depends on partition -- so you can never start
Calamares on FreeBSD, because bootloader depends on something
non-existent.
Relax the check: just warn, and only fail if a non-existent
module is used (all those with unmet dependencies are considered
non-existent).
- Calamares scans **all** subdirs of the module-directory
for a module.desc and complains about those that don't have
a module.desc.
- For ./calamares -d runs from the build-directory, this
leads to a few complaints when some plugins have been
ignored (and so no module.desc is generated for them).
- Things in libcalamares/ subdirectories are namespaced
according to that subdirectory (sometimes in namespace
Calamares, sometimes CalamaresUtils). Do that in modulesystem/ too.
- this is not entirely straightfoward, since we need
different constructor arguments for the objects
Calamares creates (no QVariantList& args, in particular).
Implement our own registerPlugin() and createInstance()
for that.
- work around a bug in K_PLUGIN_FACTORY_DECLARATION_WITH_BASEFACTORY
- Give each check a name (based on the module it runs for, so
there might be overlaps when there are multiple module instances).
- Log the remaining checks each time the timeout fires, to help
figure out which one is hanging.
- Avoid races which might double-delete the timer, or enter
the if twice (which would lead to duplicate emissions of
requirementsComplete and the associated UI glitches).
If the requirements checking is **really fast**, e.g. you don't
have a check for internet connectivity, then the checks
might be done as fast as the 0-timeout single-shot timer,
which means that finished() is called once by the QFutureWatcher,
and then after that by the QTimer .. leading to two messages
"All requirements have been checked", but also twice
requirementsComplete being emitted, so you end up with two
results lists being added by the CheckerContainer.
Stop that by using the results-progress timer as an additional
flag: the first time everything is complete, delete that timer
and set the pointer back to nullptr.
- This small header file contained a few unrelated typedefs.
Move those typedefs to the classes they relate to. This
**does** mean that some consumers need to #include something
else instead.
- Use type names more consistently.
Editorial: why are **pages** responsible for creating the jobs?
- Currently just moves a single enum, but this is prep-work for
moving the non-GUI parts of the module system into libcalamares,
to better support GUI-less operation.
- Use only utils/YamlUtils.h to pull in yaml-cpp and supporting code.
- When compiling with clang, turn off warnings that the system header
for yaml-cpp would generate.
- Avoid crash due to invalid iterator, when modules
are removed due to missing requirements.
- Simplify code, factor out the determination of
which required modules are missing.
- Move the actual checking into a separate object with some lifecycle-
management signals.
- Right now this is still single-threaded and blocking, so no net gain.
- Use DATA for the qml and branding directories (looks for qml/
and branding/<name>/ in those directories).
- Use CONFIG for the global settings and module settings (looks
for settings.conf and module/<name>/ in those directories).
FIXES#941
- Be more verbose when things go wrong while loading module-
configuration files.
- Allow more forms of paths to be specified. In debug mode, this
is mostyle helpful for the module test-loader.
- document new variable from the CMake module
- use it in libcalamaresui to simplify #include'ing the
header for the "all" extension.
Suggested by Denis Proskurin.
- Similar to the refactorings in Calamares proper, just split out
collecting the search paths into a static function. This makes
it a little easier to find places that will need expansion for
more-than-one-config-directory.
- There is more to failing out of loadModules() than just
emitting modulesFailed, so instead share the failure
code with the code after loading modules -- but don't load any.
- Module dependency-checking is done in two phases:
first, catch any unknown modules that are listed
in *requiredModules* and bail out before loading
anything. Second, check that the modules required
by X occur before X in the sequence.
- The value set in module.desc was never stored for use,
but isn't an attribute of the instance, either. It belongs
with the descriptor, in ModuleManager.
Introduce the notion of emergency modules and emergency jobs.
Initial use will probably center around the preservefiles module,
and possibly umount.
FIXES#928
A potentially emergency module is one that has EMERGENCY
(in CMake) or emergency: true (in module.desc) set.
Any such module must also set emergency: true in the
configuration of the module. This is to allow for
instances of a module that **don't** run as emergency
modules, alongside actual emergency ones.
- Add a TODO for allowing modules to come from somewhere other
than the module loader (this would allow "internal" modules
that are always present)
- Warnings are warnings
- Collect the failed modules, instead of bailing out on the first one
(this also prevents crashes caused by quit() called from a timer).
- Introduce a slot to report on failed module loading (no UI yet).
- Put the (constant) 'Calamares will now quit' on its own debug line.
- Tell the user what the search paths are if a module is not found
(prompted by a mis-configuration in a Neon live image).
- Move type and rename it; put in Calamares namespace
- Emit signals from the viewmanager as results come in
- Remove state changing from welcome view step based on its internal
requirements checking (for now this breaks progressing past the
welcome page)
- Log checking of the requirements
Introduce a method checkRequirements() into the module system so that
individual modules can do their own checking (as opposed to stuffing
it all into the welcome module).
- Was marked incomplete and unused; none of the existing modules
set any requirements, and the descriptors are not set up to
hold the requirements information anyway.
- Dependencies are generally through globalStorage values, or if
there are dependent jobs they should be created in-order by one
source (e.g. though a View or a subclass of CppJob which overrides
jobs() ).
- It is the responsibility of deployers to formulate a settings.conf
that includes all the required modules.
- A 'real' dependency system is going to lead to the introduction
of interface-definitions and a great deal of complexity, for a
use case that can be handled with careful deployment instead.
