Indeed...painful and slow :( What I need is R 3.1.1 or later...so if I can get a pre-compiled package in toolshed that works I’m also happy! I understand Dave manages this process? @Dave: can you help me?
Thanks, Pieter From: bjoern.gruen...@googlemail.com [mailto:bjoern.gruen...@gmail.com] Sent: woensdag 12 november 2014 17:18 To: Lukasse, Pieter Cc: galaxy-dev@lists.bx.psu.edu; Dave Bouvier Subject: Re: [galaxy-dev] Tool shed environment dependency - HOW TO do this? Yes, this is correct. Do you really need to compile it by yourself? This is really a pain, I tried it during the R2 lifetime and this wasn't fun. 2014-11-12 17:10 GMT+01:00 Lukasse, Pieter <pieter.luka...@wur.nl<mailto:pieter.luka...@wur.nl>>: Ok, I didn’t know that...but it does explain the behavior we are observing. So the <actions_group> tag is a kind of switch/try mechanism where the next applicable <actions> part is only executed if the previous one fails? Is that how I should read it? So my script is based on the last part, which actually tries to compile R with cairo....and unfortunately this is not working. But maybe this was not detected by others because as you wrote, the compilation part is only a fall back (and probably will not be executed in practice)...? Thanks, Pieter. From: bjoern.gruen...@googlemail.com<mailto:bjoern.gruen...@googlemail.com> [mailto:bjoern.gruen...@gmail.com<mailto:bjoern.gruen...@gmail.com>] Sent: woensdag 12 november 2014 17:01 To: Lukasse, Pieter Cc: galaxy-dev@lists.bx.psu.edu<mailto:galaxy-dev@lists.bx.psu.edu>; Dave Bouvier Subject: Re: [galaxy-dev] Tool shed environment dependency - HOW TO do this? Hi Pieter, the compilation and hence the dependencies on the libraries for compilation are a fallback mechanism. This recipe will try to fetch binaries, if this does not work, or if the admin reconfigured Galaxy ... we try to compile R by our own. Hope this helps, Bjoern 2014-11-12 16:57 GMT+01:00 Lukasse, Pieter <pieter.luka...@wur.nl<mailto:pieter.luka...@wur.nl>>: Hi Bjoern, Dave, I have been trying to understand how https://toolshed.g2.bx.psu.edu/repos/iuc/package_r_3_0_3 works . It contains a number of package dependencies as shown in the XML below. However, none show up in the toolshed UI (next screenshot). Is this because the package entries (libpng, cairo, etc) are inside a <actions> tag while they should be outside of this tag? Could it be these are not executed after all? Furthermore, the installation actually seems to be a simple download of a pre-compiled R version from depot.galaxy.org<http://depot.galaxy.org> . So why the ./configure and make commands further down the XML? Thanks for your help! Pieter -----Original Message----- From: Björn Grüning [mailto:bjoern.gruen...@gmail.com<mailto:bjoern.gruen...@gmail.com>] Sent: donderdag 30 oktober 2014 11:25 To: galaxy-dev@lists.bx.psu.edu<mailto:galaxy-dev@lists.bx.psu.edu>; Lukasse, Pieter; Dave Bouvier Subject: Re: [galaxy-dev] Tool shed environment dependency - HOW TO do this? Hi Pieter, Am 30.10.2014 um 11:15 schrieb Lukasse, Pieter: > Hi, > > I have a Bioconductor package which depends on R 3.1.1 , so I think I cannot > use the "setup_r_environment" trick. Why? Do you need a new release? The latest current release in the Tool Shed is R 3.1.0. If this is not enough we need to poke Dave to build a new package. But this way would be the preferred way. > I already got a tool_dependency.xml working that installs R 3.1.1 and the > necessary Bioconductor packages using bioclite method (see attachment). Now > my question is: > > > * I would like to split up this into two steps as I don't want to > trigger the compilation of new R environment every time I when I need to just > update the >Bioconductor package....the question is: how to do such things in general? How >can I access the INSTALL_DIR of another tool from within another >tool_dependency.xml? If I can do this, then my problem is solved. If you really want to build your R packages by your own. Have a look at https://github.com/bgruening/galaxytools/blob/master/packages/package_r_3_0_3/tool_dependencies.xml You will find an exhaustive installation instruction how to build R properly. Also note that we export a variable called R_ROOT_DIR that is set to $INSTALL_DIR. You can now access R_ROOT_DIR from any other tool_dependency file ... Hope this helps, Bjoern P.S. Do you mind asking this question on biostar again, I think this is a very nice question that can be of interest to a lot more people. > Thanks! > > Pieter Lukasse > Wageningen UR, Plant Research International Department of > Bioinformatics (Bioscience) Wageningen Campus, Building 107, > Droevendaalsesteeg 1, 6708 PB, Wageningen, the Netherlands > T: +31-317481122<tel:%2B31-317481122>; > M: +31-628189540<tel:%2B31-628189540>; > skype: pieter.lukasse.wur > http://www.pri.wur.nl<http://www.pri.wur.nl/<http://www.pri.wur.nl%3chttp:/www.pri.wur.nl/>> > > > > > ___________________________________________________________ > Please keep all replies on the list by using "reply all" > in your mail client. To manage your subscriptions to this and other > Galaxy lists, please use the interface at: > http://lists.bx.psu.edu/ > > To search Galaxy mailing lists use the unified search at: > http://galaxyproject.org/search/mailinglists/ >
___________________________________________________________ Please keep all replies on the list by using "reply all" in your mail client. To manage your subscriptions to this and other Galaxy lists, please use the interface at: http://lists.bx.psu.edu/ To search Galaxy mailing lists use the unified search at: http://galaxyproject.org/search/mailinglists/
