http://mgltools.scripps.edu Should be available for protein structure science. Perhaps the science overlay. Reproducible: Always
Created attachment 142835 [details] sci-libs/msms-2.6.1.ebuild sci-libs/msms-2.6.1.ebuild
Created attachment 143150 [details] update critical thing fix (bins are installed in /bin not /usr/bin amd64 added, but at this time upstream bug in the compression (tar.gz is just tar) some syntax optimisations
Created attachment 143153 [details] msms-2.6.1.ebuild typo, sorry about that!
Comment on attachment 143153 [details] msms-2.6.1.ebuild ><HTML><HEAD/><BODY><PRE># Copyright 1999-2008 Gentoo Foundation ># Distributed under the terms of the GNU General Public License v2 ># $Header: $ > >inherit eutils > >DESCRIPTION="This allows you to run msms as a standalone application. A wrapped version of the library comes with MGLTools" >SRC_URI="x86? ( http://mgltools.scripps.edu/downloads/tars/releases/MSMSRELEASE/REL${PV}/${PN}_i86Linux2_${PV}.tar.gz ) > amd64? ( http://mgltools.scripps.edu/downloads/tars/releases/MSMSRELEASE/REL${PV}/${PN}_i86_64Linux2_${PV}.tar.gz )" >HOMEPAGE="http://mgltools.scripps.edu" >RESTRICT="mirror" >LICENSE="academical free" >SLOT="" >KEYWORDS="~x86 ~amd64" >IUSE="static" > >RDEPEND="" >DEPEND="${RDEPEND}" > >src_install() { > insinto /usr/lib/msms/ > doins atmtypenumbers > exeinto /usr/lib/msms/ > doexe pdb_to_xyzr pdb_to_yzrn > x86? && doexe msms.i86Linux2.2.6.1 > amd64? && !static? && doexe msms.x86_64Linux2.2.6.1 > amd64? && static? && doexe msms.x86_64Linux2.2.6.1.staticgcc > x86 && dosym /usr/lib/msms/msms.i86Linux2.2.6.1 /usr/bin/msms > amd64? && static? && dosym /usr/lib/msms/msms.x86_64Linux2.2.6.1.staticgcc /usr/bin/msms > amd64? && !static? && dosym /usr/lib/msms/msms.x86_64Linux2.2.6.1 /usr/bin/msms > dosym /usr/lib/msms/pdb_to_xyzr /usr/bin/pdb_to_xyzr > dosym /usr/lib/msms/pdb_to_xyzrn /usr/bin/pdb_to_xyzrn > dodoc README ReleaseNotes > doman msms.1 >} > ></PRE></BODY></HTML>
Comment on attachment 143153 [details] msms-2.6.1.ebuild wrong use of arch specific code
Created attachment 143286 [details] msms-2.6.1.ebuild
Hi Justin, Thanks much for your efforts! Is there any chance the we can have upstream release the sources for msms? Only in very special circumstances should binary packages ever end up in portage. Thanks, Markus
I do not know. They provide on there webpage only the bins. I wrote this ebuild originaly, because I needed msms as a depency. Think there is no really requirement for many people. Just wanted to provide it. The effort I made were simply for training my skills. The original ebuild was just for x86, but I tried to enlarge it at least to amd64. As I am an uprising ebuild writer I made some big mistakes, thats why there were so many version. As an summary: Think no sources available, if there is a need here is the ebuild, else forget about it. Thanks justin P.s. I just found out that there sources but inside a larger program package (MGLTools). If I got time I will look on that. http://mgltools.scripps.edu/downloads
If it does end up being a binary package, it should get the -bin suffix in its name and install to /opt rather than /usr. Although since you noticed that its source is released as part of another package, it really shouldn't be a major issue for upstream to also release it separately.
Created attachment 143412 [details] msms-bin-2.6.1.ebuild You are right. Corrected now to /opt and -bin. This ebuild is a dependency of bug 209970.
msms is a pretty nice tool. I've used it a lot as part of VMD to render surfaces. Let's look into the possibility of getting at the sources, either by pulling them out of a larger source tarball or asking upstream if they'd be willing to post ones for msms only. cheers, Markus
I contact the author with the request for publication of the sources. Let's wait for that.
Hi Justin, That sounds great! Thanks, Markus
Created attachment 143504 [details] msms-bin-2.6.1.ebuild We use gawk and not nawk, don't we?
Comments: - If a package is named msms-bin, it should install to /opt/msms-bin - Try a case statement for the ARCH section instead. It's a bit cleaner. - Relative symlinks - I'm a little surprised RDEPEND is empty. Is that right? It could be. - What are you doing with the eclasses this inherits? - Should the keywords have -* ?
(In reply to comment #15) > Comments: > > - If a package is named msms-bin, it should install to /opt/msms-bin > - Try a case statement for the ARCH section instead. It's a bit cleaner. I did. > - Relative symlinks I reformed them. Hope that is what you mean, although the most ebuild use absolute statements. > - I'm a little surprised RDEPEND is empty. Is that right? It could be. just glibc > - What are you doing with the eclasses this inherits? They are looking very important!!! :) > - Should the keywords have -* ? Right. >
Created attachment 146033 [details] msms-bin-2.6.1.ebuild
There are some unclarity with the license plus I found some sources. Thus I clear things with upstream and reopen this Bug or setup a new one.
Okay, I learned something in the last time. Here is the msms lib source ebuild. http://git.overlays.gentoo.org/gitweb/?p=proj/sci.git;a=tree;f=sci-libs/msms;h=2b1ece8e400aa1dd3a08dcab8c7ab726739da9e2;hb=HEAD
I added it to tree with its own MSMS licence