Open KingDuckZ opened 3 years ago
Is it possible to get some feedback on this please?
About the failures it has to be expected, there are some combination of header options that not checked frequently. So in particular if you are testing the garbage torture test, you'll see potentially a good number of them and times out. Some are unchecked returns values, race condition against ourself...
Thanks for the reply. Failures are the ones I listed already.
So wouldn't the project benefit from making tests easier to run? If you try my meson.build
you'll be able to run all tests with ninja test
.
I have no control on the build system, so wait for an official answer of @ruby0x1
Hey wren people, I'm maintaining a small meson script as part of my c++ wren wrapper project. I updated it to work with wren 0.3.0 just now, and I can make whatever change to make sure it works with the latest commit if you're interested in adding that directly into wren? As visible from the
meson_options.txt
file it supports optional unit test compilation and optional meta and random modules. Previous version also supported optional cli, but I see wren-cli's been split into a separate repository now, so I removed that option.As an aside, when I run
ninja test
I get some failures in the unit tests (assuming I'm running the unit test program correctly), is that something you're aware of? Namely it'scall_calls_foreign
,get_variable
,call_wren_call_root
unexpected pass,resolution
,new_vm
,handle
,error
,lists
,user_data
,slots
,foreign_class
,api_call
,api_foreign_method
,module
,get_module_variables
.In order to detect which tests are expected to fail, I wrote a python script that searches for an "expected fail" regex in each .wren file. This is brittle and very slow, would it be possible to have that information in the filename itself as a convention? For example
my_good_test.wren
andmy_bad_test.fails.wren
.