Add support for -S, -emit-llvm and output to stdout (-o -) #9014
Conversation
jgravelle-google
left a comment
jgravelle-google
left a comment
There was a problem hiding this comment.
Woot, mostly lgtm
| }''') | ||
| run_process([PYTHON, EMCC, 'src.c', '-s', 'NO_FILESYSTEM=1']) | ||
|
|
||
| def test_dashS(self): |
There was a problem hiding this comment.
Nit: could be dash_S to match has_dash_S? test_dashS_stdout looks like it's mixing camel and snake case to me, the underscore makes that less odd I feel.
There was a problem hiding this comment.
I'm just mimicking the test_dashE_xx and test_dash_M_xxx tests below.
quantum5
left a comment
quantum5
left a comment
There was a problem hiding this comment.
Looks fine if we don't care about .bc file generation.
emcc.py
Outdated
| has_dash_S = '-S' in newargs | ||
| if has_dash_c: | ||
| assert has_source_inputs or has_header_inputs, 'Must have source code or header inputs to use -c' | ||
| target = target_basename + '.o' |
There was a problem hiding this comment.
Do we want a if '-emit-llvm' in newargs here as well to generate a .bc file?
emcc.py
Outdated
| target = target_basename + '.o' | ||
| final_suffix = '.o' | ||
| elif has_dash_S: | ||
| if '-emit-llvm' in newargs: |
There was a problem hiding this comment.
I'm not sure it's worth checking for -emit-llvm - if we do that, we may also want to check -flto and WASM_OBJECT_FILES=0 as those also affect whether we emit bitcode. But I think it might be overkill to do all that.
There was a problem hiding this comment.
This is all for the default output filename if otherwise unspecified by the user, right? So letting this be best-effort is probably good enough.
There was a problem hiding this comment.
Regardless of WASM_OBJECT_FILES or flto settings -emit-llvm will always generate bitcode output from clang, and this is already supported in way since we pass this through from the command line.
Or are you saying the -emit-llvm might not have come from the user? .. I'll check that.
No description provided.