[PATCH] D22172: Added REQUIRES to TestingGuide documentation

[Mehdi AMINI via llvm-commits]

mehdi_amini added inline comments.

================
Comment at: llvm/trunk/docs/TestingGuide.rst:406-434
@@ +405,31 @@
+List of features that can be used in ``REQUIRES`` and ``REQUIRES-ANY``:
+- asan
+- not_asan
+- asserts
+- can-execute
+- debug_frame
+- default_triple
+- dfsan
+- fma3
+- global-isel
+- ld64_plugin
+- ld_emu_elf32ppc
+- ld_plugin
+- linux
+- loadable_module
+- long_tests
+- lsan
+- msan
+- not_msan
+- native
+- object-emission
+- python-psutil
+- shell
+- system-windows
+- ubsan
+- not_ubsan
+- x86_64-linux
+- xar
+- zlib
+- nozlib
+
----------------
Prazek wrote:
> mehdi_amini wrote:
> > MatzeB wrote:
> > > Should we really list all of them here? I see this list getting out of date quickly as it is not obvious to people changing lit.cfg that this list exists here. I'd recommend to only refer to lit.cfg and maybe mention 1 or 2 examples...
> > Agree.
> I think having list somewhere is usefull. It wasn't obvious for me where I can find the code that defines the features - and all features are splitted between different files. The argument that it get outdated fast ins invalid - just keep track of it in reviews (I am willing to be added as a subscriber to every patch that touches any lit.cfg to check it) 
> or find a way to generate it from code.
> 
> The list is useful. e.g. I can see that the list doesn't mention anything similar to LLVM_ENABLE_STATS that I was looking for, so I know that I should probably go different way.
> 
> The argument that it get outdated fast ins invalid - just keep track of it in reviews (I am willing to be added as a subscriber to every patch that touches any lit.cfg to check it) 

Sorry but you can't just say that the argument "is invalid". Experience shows it is.

> or find a way to generate it from code.

Sure, in the meantime I think we should remove it.


Repository:
  rL LLVM

http://reviews.llvm.org/D22172