On Fri, Dec 20, 2019 at 09:22:07PM +0000, Stuart Henderson wrote:
> When updating ports I'd find it helpful to have an easier way to check
> signatures from upstream. What does anyone think about the general idea
> or the specifics of this? List the signature file in SIGFILES, set
> CHECK_PGPSIG=Yes in mk.conf, and with the sample I've included for
> mutt in the patch below (if you already have the signer's public key
> in your keyring) it looks like this:
>
> $ make checksum
> ===> Checking files for mutt-1.13.2v3
> `/y/Download/ftp/pub/OpenBSD/distfiles/mutt-1.13.2.tar.gz' is up to date.
> `/y/Download/ftp/pub/OpenBSD/distfiles/mutt-1.13.2.tar.gz.asc' is up to date.
> >> (SHA256) mutt-1.13.2.tar.gz: OK
> >> (SHA256) mutt-1.13.2.tar.gz.asc: OK
> gpg: assuming signed data in `mutt-1.13.2.tar.gz'
> gpg: Signature made Wed Dec 18 22:23:48 2019 GMT using RSA key ID 80316BDA
> gpg: Good signature from "Kevin J. McCarthy <[email protected]>"
> gpg: WARNING: This key is not certified with a trusted signature!
> gpg: There is no indication that the signature belongs to the owner.
> Primary key fingerprint: 8975 A9B3 3AA3 7910 385C 5308 ADEF 7684 8031 6BDA
>
> I'm thinking of this as a ports maintainer tool only - I don't think
> it's useful for bulk builds as we already check hashes against the ports
> tree. Intentionally not switching uid because that will make it tricky
> to manage public keys.
>
> Thoughts?
>
Looks good in general. Comments inline.
> Index: infrastructure/mk/bsd.port.mk
> ===================================================================
> RCS file: /cvs/ports/infrastructure/mk/bsd.port.mk,v
> retrieving revision 1.1514
> diff -u -p -r1.1514 bsd.port.mk
> --- infrastructure/mk/bsd.port.mk 16 Dec 2019 09:18:27 -0000 1.1514
> +++ infrastructure/mk/bsd.port.mk 20 Dec 2019 21:07:55 -0000
> @@ -117,7 +117,7 @@ _ALL_VARIABLES_PER_ARCH =
> # consumers of (dump-vars) include sqlports generation and dpb
> # dpb doesn't need everything, those are speed optimizations
> .if ${DPB:L:Mfetch} || ${DPB:L:Mall}
> -_ALL_VARIABLES += DISTFILES PATCHFILES SUPDISTFILES DIST_SUBDIR MASTER_SITES
> \
> +_ALL_VARIABLES += DISTFILES PATCHFILES SUPDISTFILES SIGFILES DIST_SUBDIR
> MASTER_SITES \
Maybe PGPSIGFILES ? a bit more typing, but SIGFILES is not very explicit.
> MASTER_SITES0 MASTER_SITES1 MASTER_SITES2 MASTER_SITES3 MASTER_SITES4 \
> MASTER_SITES5 MASTER_SITES6 MASTER_SITES7 MASTER_SITES8 MASTER_SITES9 \
> CHECKSUM_FILE FETCH_MANUALLY MISSING_FILES PERMIT_DISTFILES
> @@ -1289,6 +1289,11 @@ DISTFILES ?= ${DISTNAME}${EXTRACT_SUFX}
>
> PATCHFILES ?=
> SUPDISTFILES ?=
> +SIGFILES ?=
> +CHECK_PGPSIG ?= No
> +.if ${CHECK_PGPSIG:L} == yes
> +BUILD_DEPENDS += security/gnupg
> +.endif
>
> # the following loop "parses" DISTFILES-style files
> # _PATH_x contains filenames with SUBDIR prepended when necessary
> @@ -1300,10 +1305,10 @@ SUPDISTFILES ?=
> #
> # _FILES is used to de-duplicates names
> # the order matters: DISTFILES PATCHFILES SUPDISTFILES
> -# - we never have the same names in DISTFILES and PATCHFILES
> +# - we never have the same names in DISTFILES and PATCHFILES or SIGFILES
> # - SUPDISTFILES has to happen later
> _FILES=
> -.for v in DISTFILES PATCHFILES SUPDISTFILES
> +.for v in DISTFILES PATCHFILES SIGFILES SUPDISTFILES
> . if !empty($v)
> . for e in ${$v}
> . for f m u in ${e:C/:[0-9]$//:C/(.*)\{.*\}(.*)/\1\2/}
> MASTER_SITES${e:M*\:[0-9]:C/.*:([0-9])/\1/}
> ${e:C/:[0-9]$//:C/.*\{(.*)\}(.*)/\1\2/}
> @@ -1327,7 +1332,7 @@ _LIST_$v =
> .endfor
> _FULL_FETCHLIST ?=
>
> -CHECKSUMFILES = ${_PATH_DISTFILES} ${_PATH_PATCHFILES}
> +CHECKSUMFILES = ${_PATH_DISTFILES} ${_PATH_PATCHFILES} ${_PATH_SIGFILES}
Now this assumes you always check SIGFILES even if you don't intend to check
them later.
> MAKESUMFILES = ${CHECKSUMFILES} ${_PATH_SUPDISTFILES}
I think I would rather add them unconditionally to MAKESUMFILES,
and conditionnaly to CHECKSUM_FILES, depending on CHECK_PGPSIG...
(I haven't tried it, but from what I remember from the logic this should
work)
>
> # This is what is actually going to be extracted, and is overridable
> @@ -2491,6 +2496,16 @@ _internal-checksum: _internal-fetch
> fi; \
> fi
> . endif
> +. endif
> +. if !empty(SIGFILES) && ${CHECK_PGPSIG:L} == yes
> + @cd ${DISTDIR}; OK=true; \
> + for file in ${SIGFILES}; do \
> + gpg --verify $$file || OK=false; \
> + done; \
> + if ! $$OK; then \
> + echo "GPG signature invalid."; \
> + exit 1; \
> + fi
I assume you want to run thru *all* SIGFILES in every case ? because
it's very convoluted, as opposed to just running gpg and letting it fail.
Also, it's going to be a bit verbose, between gpg --verify output and the
echo.
If you want to keep it that way, the echo should probably be on stderr, as
it's an error message ?
> . endif
>
>
> Index: mail/mutt/Makefile
> ===================================================================
> RCS file: /cvs/ports/mail/mutt/Makefile,v
> retrieving revision 1.126
> diff -u -p -r1.126 Makefile
> --- mail/mutt/Makefile 19 Dec 2019 13:49:44 -0000 1.126
> +++ mail/mutt/Makefile 20 Dec 2019 21:07:55 -0000
> @@ -3,6 +3,7 @@
> COMMENT= tty-based e-mail client
>
> DISTNAME= mutt-1.13.2
> +SIGFILES= mutt-1.13.2${EXTRACT_SUFX}.asc
> EPOCH= 3
>
> CATEGORIES= mail
> Index: mail/mutt/distinfo
> ===================================================================
> RCS file: /cvs/ports/mail/mutt/distinfo,v
> retrieving revision 1.37
> diff -u -p -r1.37 distinfo
> --- mail/mutt/distinfo 19 Dec 2019 13:49:44 -0000 1.37
> +++ mail/mutt/distinfo 20 Dec 2019 21:07:55 -0000
> @@ -1,2 +1,4 @@
> SHA256 (mutt-1.13.2.tar.gz) = L3OTG/9AjvZLgp9ZWw2vg6yBjRVEWJUgQLcEguZ2nnQ=
> +SHA256 (mutt-1.13.2.tar.gz.asc) =
> 1qBz5va7E+IPbGTj4alZGpSQxXIv/JZZimgP1Rqs3Do=
> SIZE (mutt-1.13.2.tar.gz) = 4895334
> +SIZE (mutt-1.13.2.tar.gz.asc) = 833
>
>
This obviously also requires pieces in sqlports, but that can come in later.