On Thu, Nov 29, 2018 at 6:31 PM Rob Landley <[email protected]> wrote: > > On 11/29/18 5:37 PM, enh wrote: > >> Anyway, I'm happy to give a clearer error. _How_? (What does "uname -o" > >> say on a > >> mac?) > > > > "illegal option" :-) > > > > just plain `uname` gets me "Darwin". > > > > (given that `toybox uname -o` behaves differently than coreutils and > > busybox ["Linux" versus "GNU/Linux"], and isn't even documented in > > toybox, you might want to steer clear of that anyway.) > > Sigh. Serves me right for looking at "man uname" and thinking the result meant > anything... > > But [ "$(uname)" == "Darwin" ] I can work with. > > >> At the very leas I can make $SED default to "gsed" when [[ "$(uname -o") > >> =~ Mac ]] or some such, then at least the error is gsed missing. But that > >> assumes I'm detecting MacOS... > > > > the reason why i haven't sent you a patch is that there's a bunch of > > other scripts that need similar treatment... should i factor it out > > into scripts/findsed.sh and source that from the affected scripts? > > What other scripts need this treatment?
i assume you mean which scripts *absolutely* need it (or you'd have just run `grep -rw sed`)... iirc, it was test.sh and single.sh. but presumably if we had a findsed.sh we'd just call it everywhere and switch everything over to $SED? > Rob _______________________________________________ Toybox mailing list [email protected] http://lists.landley.net/listinfo.cgi/toybox-landley.net
