New ebuild,Build and tested on ~amd64. Split off from https://bugs.gentoo.org/show_bug.cgi?id=579580 as net-misc/gnurl can be released before gnunet and gnunet-gtk As put into the metadata.xml, I will be the maintainer via proxy-maint. Reproducible: Always
Created attachment 434774 [details] =net-misc/gnurl-7.45.0
Created attachment 434776 [details] =net-misc/gnurl-9999
Created attachment 434778 [details] metadata.xml
One possible fix remains: HOMEPAGE currently has 2 values, the gnunet.org/gnurl and the git page displaying the description. Should it be reduced to gnunet.org/gnurl? I initially went with 2 for the verbosity of information.
Additional comment: both versions additionally compiled and running successfully (so far) on gentoo-hardened base system (gcc).
g-p-m has been added in CC however, has proxy-maint@gentoo.org yet declared he wishes to be set as the proxied maintainer?
(In reply to Ian Delaney from comment #6) > g-p-m has been added in CC however, has proxy-maint@gentoo.org yet declared > he wishes to be set as the proxied maintainer? I hope I read this right and you are asking me if I have declared to become proxied maintainer. This is an ebuild directly depending on bug 579580, where I declared to become proxied maintainer of all packages included in it, which also included gnurl before I split it off. Long sentence, explicit answer: Yes, I want to become the proxied maintainer. In case I did not understand it and answered wrong, please explain the parts I did not understand.
ok thanks for the contribution. This ebuild is close but will need 'shoring up'. 1. in metadata.xml <description> tag has been decided within g-p-m as a waste of space. I can just take it out otherwise fine. 2. ebuild. a) LICENSE="MIT" can you double check this? Recent submissions have been found to have either or an actual license of similar name. b) DEPEND="${RDEPEND}" is almost a clichee. Are all deps of runtime really required for the build? Yes there are only two. b) The running of eautoreconf normally takes place after application of patches and sed statements, not before. c) There are 4 distinct sed statements. This many edits questions the state of the build system. Are all of these gentoo specific? Any that are generic require submission to upstream for consideration to incorporate into the source of the package. It also beckons prompt for use of a classical patch; short sharp and clean. d) # FIX: Skip examples and man3 Why? I generally include the IUSE flag examples and install them upon use of the flag. If a man page is made available, why hold it back. I'd be looking at its installation as default. e) The P.M.S. has a function to install docs einstalldocs. An option for the rather extended DOCS var is to perform this installation, then any missed top up with " dodoc the rest"
(In reply to Ian Delaney from comment #8) > ok thanks for the contribution. This ebuild is close but will need 'shoring > up'. > > > 1. in metadata.xml <description> tag has been decided within g-p-m as a > waste of space. I can just take it out otherwise fine. > > 2. ebuild. > > a) LICENSE="MIT" can you double check this? Recent submissions have been > found to have either or an actual license of similar name. > b) DEPEND="${RDEPEND}" is almost a clichee. Are all deps of runtime really > required for the build? Yes there are only two. > b) The running of eautoreconf normally takes place after application of > patches and sed statements, not before. > c) There are 4 distinct sed statements. This many edits questions the state > of the build system. Are all of these gentoo specific? Any that are generic > require submission to upstream for consideration to incorporate into the > source of the package. It also beckons prompt for use of a classical patch; > short sharp and clean. > d) # FIX: Skip examples and man3 > Why? I generally include the IUSE flag examples and install them upon use of > the flag. If a man page is made available, why hold it back. I'd be looking > at its installation as default. > e) The P.M.S. has a function to install docs einstalldocs. An option for the > rather extended DOCS var is to perform this installation, then any missed > top up with " dodoc the rest" Thanks for looking into it. To a): it is the same license as net-misc/curl, which is "MIT" in portage. http://www.git.taler.net/?p=gnurl.git;a=blob_plain;f=COPYING;h=a98663e944f2081605d3a0d5445d08eb5c439e88;hb=HEAD vs https://curl.haxx.se/docs/copyright.html I will adjust the rest.
(In reply to Ian Delaney from comment #8) .... > c) There are 4 distinct sed statements. This many edits questions the state > of the build system. Are all of these gentoo specific? Any that are generic > require submission to upstream for consideration to incorporate into the > source of the package. It also beckons prompt for use of a classical patch; > short sharp and clean. > d) # FIX: Skip examples and man3 > Why? I generally include the IUSE flag examples and install them upon use of > the flag. If a man page is made available, why hold it back. I'd be looking > at its installation as default. I am in discussion with the developers to figure out the best way to deal with this, see what's generic and what's distro specific. At its core, gnurl is a reduced curl for gnunet, build against libgcrypt and leaving out many config options (dnssec breaks for example). That's reason enough to have 2 packages, however gnurl is never intended to be run by users from the commandline, so some parts are really not needed. The discussion has been moved to the mailinglist, what happened before can be read here: https://gnunet.org/bot/log/gnunet/2016-05-26 I will adjust the gnurl ebuild, and wait for how the discussion develops with upstream.
(In reply to Ian Delaney from comment #8) > b) DEPEND="${RDEPEND}" is almost a clichee. Are all deps of runtime really > required for the build? Yes there are only two. Extract from https://www.git.taler.net/?p=gnurl.git;a=blob;f=README;h=28e8cd89956d9b8f6fa0736b407fa52e42e02f27;hb=HEAD : Note that the compilation requirements were not hard-coded, but are rather socially enforced: if you compile libgnurl, please use the following options to configure: ./configure --enable-ipv6 --with-gnutls --without-libssh2 --without-libmetalink --without-winidn --without-librtmp --without-nghttp2 --without-nss --without-cyassl --without-polarssl --without-ssl --without-winssl --without-darwinssl --disable-sspi --disable-ntlm-wb --disable-ldap --disable-rtsp --disable-dict --disable-telnet --disable-tftp --disable-pop3 --disable-imap --disable-smtp --disable-gopher --disable-file --disable-ftp --disable-smb Naturally, you're free to specify additional options, such as "--prefix". The result should have support only for HTTP, HTTPS (via GnuTLS), IDN, zlib and TLS-SRP. /End Extract This leads me to think that zlib and gnutls are runtime and build dependencies, though if you know that zlib is not required at runtime but at buildtime only for curl, then I'll fix the ebuild DEPEND and RDEPEND.
I hereby pull back my request to become maintainer as I can't maintain works (and most importantly, double work sometimes) on two OS. Whoever likes to pick it up, feel free to do so. Adding to this package for the record: it will probably be maintained for a longer time at the overlay at http://youbroketheinternet.org/#overlay If other people in the overlay feel like fixing it, I have already fixed it in the other OS and package manager, volunteered double effort is energy draining for me.
Created attachment 437896 [details] =net-misc/gnurl-7.45.0 adding last fixes before i leave this to the overlay.
Created attachment 437898 [details] =net-misc/gnurl-9999
Created attachment 437900 [details] metadata.xml
See bug https://bugs.gentoo.org/show_bug.cgi?id=579580
Update: I continue working on this and apply as gentoo developer in 2017 when I think this is ready to move from gnunet.org repository to portage.