Programatic procedure to retrieve the output path of a test

[umarcor]

Ref #467.

@umarcor:
Is it possible to read test_name_to_path_mapping.txt directly as some Python type (dict or list)?

@kraigher:
The file was intended for human use. The format is folder_name SPACE test_name.
I do not recommend that you use it. It is better we add this to --export-json instead.

@kraigher:
Currently the commands are logged if you use --log-level=DEBUG for human use. If an external tool needs this information we need to add it to --export-json.

@umarcor:
IMHO, there should be a simple and realiable programatic procedure to retrieve the output/log of a test. It don't really mind if I need to set some stdout in the python script, or get it as a return field from vu.main(), or parse some output file. I am trying to process test_name_to_path_mapping.txt just because I didn't find a better way to 'see' the content in output.txt.

About adding it to --export-json, although useful, I'd rather avoid it for this specific use case. Using JSON is a better approach when it is to be used from Python, Go, etc. But, with Bash or C I find it easier not to require a JSON parsing library. The current format, as you described it, is straighforward, as one only needs to search for the first space.

[kraigher]

Are you aware that the output path is accessible as an attribute on the post_check and pre_config hooks as well? Depending on the use case this might be what you need rather than having it written in the export-json file.

[umarcor]

No, I was not. And I can very probably use it to retrieve the path in the run.py file, without actually running the simulation. I will give it a try. Thanks!

[kraigher]

It will only be accessible when running the pre config hook before simulation. So you need to run the simulation.

[umarcor]

It will only be accessible when running the pre config hook before simulation. So you need to run the simulation.

I think I should be able to 'run the simulation' as --elaborate with sim_option ghdl.elab_e, so the pre config hook is triggered and the binary is compiled (thus, the output path needs to be created), but no simulation is executed. What do you think?

[kraigher]

That would work.

[umarcor]

I implemented it in #568 as follows:

def preconf(output_path):
    global opath
    opath = output_path
    return True

for tb in lib.get_test_benches(pattern='*', allow_empty=False):
    tb.set_pre_config(preconf)

ui.set_sim_option("ghdl.elab_e", True)
ui._args.elaborate = True
try:
    ui.main()
except SystemExit as exc:
    if exc.code is not 0:
        exit(exc.code)

args = [line.rstrip('\n') for line in open(join(opath, 'ghdl', 'args.txt'))]