This is what commit ca722eac6e should
have been.
Sorry for breaking a non-deprecated MSVC build once again. The reason it
"appeared to work" for me was that on deprecated builds, Manager.h
includes Manager.hpp for backwards compatibility. On non-deprecated it
doesn't and MSVC ends up dying on a linker error. But the problem was
not related to extern templates, the problem was that there was a
compiler-specific ifdef in the cpp file which exported the class for all
compilers but MSVC, somehow. And so when I removed extern template from
the header, it still worked everywhere, except MSVC and except a
non-deprecated build.
Similar to the change done in Corrade, see the commit for details:
878624ac36
Wow, this is probably the most backwards-compatibility code I've ever
written. Can't wait until I can drop all that.
The boring dry usage info grew significantly, and listing it as the
first thing on the page would scare people off. Put the examples first
instead, and for the imageconverter and sceneconverter add --info
listing examples, as that's what is eye-catchy.
Originally I copied this over from SceneGraph template classes, where it
was used to prevent the compiler from needlessly instantiating a
template that was already available elsewhere. But this is a different
case, the extern template is not preventing any instatiation of
anything, no code is inline, so it apparently should not have been there
at all, instead of being disabled for ḾinGW GCC and clang-cl, and then
subsequently discovering it also breaks MinGW Clang.
Since I'm not testing with MinGW Clang on the CI (only with MinGW GCC),
this went unnoticed for a while -- sorry.
Hey, do you also remember the times where people were excited to upgrade
from Clang 3.6 to 3.7, GCC 4.7 to 4.8 or Firefox 1.1 to Firefox 1.5? Not
Clang 12 to 14, GCC 10 to 13 or Chrome 102 to 126.
Until now, these were only transitively tested in concrete plugin
implementations, meaning that if the base implementation would have an
error (such as accessing a null optional), it would only get discovered
when building a plugin, worst case a plugin in a completely different
repo.
Here the benefit is especially clear -- as Containers::Pair is trivially
copyable with trivial types, all growable arrays can make use of
std::realloc() while with the STL variant a silly constructor, copy
constructor, destructor had to be used.
Additionally, we no longer need to take explicit care of libc++ and MSVC
STL where returning a std::pair<bool, Containers::String> as
return {{}, Containers::String{..., <deleter>}};
would caused an unnecessary copy instead of a move, losing the custom
deleter in the process. Yay!
There's a <Corrade/Containers/PairStl.h> include for backwards
compatibility purposes, but obviously it would only work for the return
type of validate*() and cases where an initializer list was passed to a
list-of-pairs-taking functions, and not a concretely typed ArrayView.
Those functions were though mostly the linker API which isn't
implemented by any plugin yet, so it shouldn't be *that* breaking to
users. Neverteless, I'm trying to do this breaking change rather sooner
than later to prevent pain further down the road when the Vulkan APIs
and SPIR-V pipeline gets widely used.
Consistently with changes done to Utility::Path, this enforces proper
error handling on user side. Originally I didn't want to do this and
instead wanted to have a special Array instance devoted for an error
state, but that still would allow the error state be errorneously
treated as a successful but empty array.
It's more useful if the Error class is directly referenced than saying
just "error output" -- so people can grab it, redirect it, etc. Also
drop the useless "does what it is expected to do on success" sentences
that add no value whatsoever.
It limits the support for CMake 3.12+, but it's much less verbose and I
don't expect people to use ancient CMake versions with IDEs like Xcode
or VS anyway, so this should be fine.
I wanted to avoid including extra stuff with the Manager.hpp split, but
this would make it even worse than having Array and String included
unconditionally. Fortunately it's enough to simply not even have the
declaration.
Probably a leftover from when these dependencies were handled in a
much shittier way? For as long as I remember, enabling WITH_GL_INFO
always enabled WITH_GL and WITH_WINDOWLESSWHATEVERAPPLICATION
implicitly.
Minor but very important convenience feature, especially useful when
dealing with command-line apps. This now works:
magnum-imageconverter a.png a.jpg -c jpegQuality=0.75
The AnyImageConverter gets the jpegQuality option and then
automatically propagates it to the concrete plugin (which is either
JpegImageConverter or StbImageConverter), possibly warning in case the
target plugin doesn't recognize given option (i.e., doesn't list it in
its default configuration). Previously the user had to always specify a
concrete converter implementation using -C, which was rather annoying
and nonintuitive.
There will be Flag::FlipY for images at some point, enabled by default
for compatibility with existing GL code, and so it makes sense to start
discouraging setFlags() as early as possible to avoid people resetting
the default by accident.
Also update the imageconverter, sceneconverter and shaderconverter utils
to use these instead of setFlags().
Currently those will be needed mainly by the Vk library to patch around
a SwiftShader bug. I'm not sure yet how the public API should look so
it's all hidden in the Implementation namespace for now.
If it's ignored, a warning is printed to catch accidents, but not an
error since it should be possible to just append --info to existing
command line to see what the input is about.
Vladimír Vondruš