Holger Levsen wrote on Fri, 21 Feb 2020 00:26 +00:00: > On Sun, Feb 16, 2020 at 07:35:39PM +0000, Daniel Shahaf wrote: > > > & patches welcome. > > Here you go, against current git: > > wheeehoo, very nice. >
Thanks for the review! > just two comments: > > > + write_file ($list_limited, <<__EOS__); > > +php5 See README.Debian.security for the PHP security policy > > +__EOS__ > > why is php5 mentioned here? I'm not sure I understand the question. I based the new test on the existing "simple ($awk)" test. That test uses php5 as the example, so that carried over to the new test. However, I suspect that doesn't answer your question. Could you clarify it, please? > > % ./check-support-status > > --except=binutils,binutils-common:amd64,binutils-x86-64-linux-gnu,libbinutils:amd64,libctf0:amd64,libctf-nobfd0:amd64 > > I'm not sure if the handling of the ":amd64" architecture suffixes is > > ideal. Thoughts? > > I'd rather not have this there: > - it makes things complicated if I need to know if a package is > arch:all or the host binary arch > - it's also redudant, if the system is amd64, all packages will be > amd64. (well, modulo multi-arch i guess) Okay, so what would you prefer? To have --except=foo match both foo and foo:bar for any value of bar? (and 'foo' documented as a bare package name without a ":arch" suffix) Cheers, Daniel
