New output for fpm test and/or fpm run breaks my CI

[everythingfunctional]

Description

fpm run or fpm test now includes Project up to date as part of the output produced if the project has already been built as part of PR #622. The test suite for my iso_varying_string library is now failing because one of the tests is for a procedure that does output, and this extra output is unexpected.

This is probably something I need to devise a better solution for in my test suite, but was an unexpected change. I think it deserves a bit of discussion whether we want this behavior though.

Expected Behaviour

fpm test or fpm run shouldn't output anything extra if the project is already up to date.

Version of fpm

current

Platform and Architecture

Linux

Additional Information

No response

[milancurcic]

Good question and I agree. I think it's also idiomatic for a UNIX-style CLI application to emit no output if there is no error and there's nothing to report.

But I think this issue title can be more clear if re-written as "Should fpm build emit output if the project is up-to-date?", as fpm test and fpm run both implicitly invoke fpm_build, if I understand it correctly. I think the answer is no, it should not.

[everythingfunctional]

I'm not opposed to fpm build reporting up-to-date, as make would do, but it probably shouldn't be said for fpm run or fpm test. That complicates things a bit, but maybe not too much.

[urbanjost]

All messages that are informative should not appear unless the --verbose|-V switch is present; other possibilities are all messages go to stderr instead of stdout; and/or the user can set a verbosity level including no messages. Having messages like that appear when using --list and --runner broke a lot of my scripts where I just wanted a pathname. Somewhat ironic it appears to have showed up along with the quiet mode for building :>

[awvwgk]

It is not uncommon for build systems to report that there is no work to be done

❯ make
make: Nothing to be done for 'all'.
❯ ninja -C _build_gcc11/
ninja: Entering directory `_build_gcc11/'
ninja: no work to do.

[urbanjost]

And all that I know of write to stderr and/or have a quiet mode; as is the convention on GNU/Linux and Unix. Even grep(1) has a quiet mode!

[Carltoffel]

I ran into this problem while following Milan's 'Modern Fortran' Book and doing it with fpm:
fpm run > tsunami_output.txt leads to fpm's (re-)building output to appear in the output file, which then isn't parsable by the plotting routine.
IMHO all of fpm's output should go to stderr.
Do you think an toml option for program output redirection could be useful?