Em Thu, 05 Sep 2019 02:28:05 -0700
Joe Perches <j...@perches.com> escreveu:

> On Thu, 2019-09-05 at 05:56 -0300, Mauro Carvalho Chehab wrote:
> > I'll accept
> > this patch and keep such rule in mind for next SPDX changes.
> > 
> > Anyway, with regards to script, we need to change the rules in order to
> > allow adding SPDX for python scripts, as otherwise the addition of SPDX
> > headers may cause regressions.  
> 
> Not really.
> 
> $ git grep -n '^# SPDX-License-Identifier' -- '*.py' | wc -l
> 62
> $ git grep -n '^# SPDX-License-Identifier' -- '*.py' | \
>   cut -f2 -d: | sort | uniq -c
>      32 1
>      27 2
>       3 3
> $ git grep -n '^# SPDX-License-Identifier' -- '*.py' | \
>   cut -f1 -d: | xargs ./scripts/spdxcheck.py --verbose
> 
> License files:               14
> Exception files:              2
> License IDs                  19
> Exception IDs                 2
> 
> Files checked:               62
> Lines checked:               95
> Files with SPDX:             62
> Files with errors:            0

Not quite sure what you're meant with the above, but:

1) this won't cover all python scripts, as there are some without
.py extension:

        $ git grep -lE "\#\!.*python"|grep -v \.py$
        drivers/staging/greybus/tools/lbtest
        scripts/bloat-o-meter
        scripts/diffconfig
        scripts/show_delta
        tools/hv/lsvmbus
        tools/kvm/kvm_stat/kvm_stat

2) we have already some files that are violating PEP-263 due to
SPDX headers. On a quick look at the results of:

        $ git grep -lE "^\#.*coding" $(git grep -lE "\#\!.*python")
        Documentation/sphinx/kernel_include.py
        Documentation/sphinx/rstFlatTable.py
        tools/perf/python/tracepoint.py
        tools/perf/python/twatch.py
        tools/power/x86/intel_pstate_tracer/intel_pstate_tracer.py

Those three files already have such problem:

        tools/perf/python/tracepoint.py
        tools/perf/python/twatch.py
        tools/power/x86/intel_pstate_tracer/intel_pstate_tracer.py

On those scripts, the coding line is not at the right place, and will be
silently ignored by python, causing them to crash with UTF-8 encoding
chars, if any.

And the other two doesn't have a SPDX header file. Adding a SPDX header
there would violate the "up to the second line" rule.

Thanks,
Mauro

Reply via email to