Update to fmt 8.1.1 #10367
Update to fmt 8.1.1 #10367
Conversation
Pokechu22
force-pushed
the
fmt-8.1.1
branch
2 times, most recently
from
878b405
to
c76b741
Compare
|
I'm not sure what's going on with the debian builders failing to find |
Yep. If you look at line 40 (when expanding the log for shell_1) you can see this in the cmake output: |
This was a longer discussion in various PRs (with the most documenting one being #8856 I guess) with the conclusion that we want at least one compiler to tell us about potential standard issues in advance; and MSVC just happened to fit the bill (IIRC because the earliest ones where we wanted to use C++14 or C++17 didn't have a formal switch for it, and we just said "use latest" instead) |
We will be using fmt::runtime, which was added in fmt 8.0.0.
This was deprecated in fmt 8.0.0.
This format string is by definition dynamic and can't be checked at compile time. There are other similar strings in the log handler and in asserts, but they use vformat and thus don't need fmt::runtime. We might be able to do a similar thing where the untranslated string is compile-time checked, but FmtFormatT is used in so few places that I don't want to handle that in this PR.
fmt 8.0.0 requires this.
At least in MSVC (which is not restricted from targetting C++20), these can be resolved to either std::format_to or fmt::format_to (though I'm not sure why the std one is available). We want the latter.
Using unsigned char* or signed char* results in a deprecation warning, which is treated as an error. It needs to be casted to regular char* for it to work.
|
FifoCI detected that this change impacts graphical rendering. Here are the behavior differences detected by the system:
automated-fifoci-reporter |
|
I've updated the minimum fmt version in CMakeLists.txt to 8.0, which has resulted in the debian builder using the fmt in Externals. |
Fixes issue 12798. Fmt 8.1.1 uses compile-time format string checking by default if
constevalis supported (i.e. C++20 is in use), which is the case for MSVC (since we don't restrict the C++ version on MSVC for some reason, though we do on other compilers). There are a large number of changes needed to fix compile errors for that. Otherwise, the main source of errors isenum classes no longer being implicitly converted toints; in most cases I've added a cast (but I created anEnumFormatterfor WD's command enums because it seemed useful there).Note that the last time I updated a library (#10188), the windows builds broke due to a bad build cache. I don't think that applies here (as fmt doesn't actually have any cpp files; we built it as a header-only library), but it's something to look out for.