вс, 23 мая 2021 г. в 17:43, Paco Esteban <p...@e1e0.net>: > > Hi ports@, > > This makes portcheck(1) not complain about modules.inc being an extra > file. This file is now being included on some Go ports, and I guess it > can be used also on Rust ports. > > I was wondering if we should perform any other tests on this files, as > we do for Makefiles with the check_makefile function. For now I only > included the basic test that modules.inc is actually a file. > > suggestions ? ok to commit ?
You definitely want to have '((++non_portmk))' line here as well. Maybe it'll be good to check if '.include "./modules.inc"' line was not forgotten or mangled in port Makefile. Or maybe it isn't worth it. I can't say since I do not touch Go/Rust ports. Also, we could count the number of MODGO_MODULES and MODCARGO_CRATES a bit later, during 'make show=...' analysis, and suggest creation of modules.inc when those numbers get too high. But what threshold should be? 10? 16? 42? > diff efc26b0628a018a694a05768ccbc773a0ed2e5ad /usr/ports > blob - bb8568b21b82fed260785c869a3cecfc4d4816cb > file + infrastructure/bin/portcheck > --- infrastructure/bin/portcheck > +++ infrastructure/bin/portcheck > @@ -612,6 +612,10 @@ check_port_dir() { > portmk_exists=true > ;; > > + modules.inc) > + test -f "$F" || err "$F is not a file" > + ;; > + > files|patches) > if [[ -d $F ]]; then > check_${F##*/}_dir "$F" > > -- > Paco Esteban. > 0x5818130B8A6DBC03 > -- WBR, Vadim Zhukov