[pkg-bioc] NETTAB meeting in Pisa, Installation of local packages dated newer than repository

David Vernazobres dv at uni-muenster.de
Mon Jun 18 23:11:24 UTC 2007


On Mon, Jun 18, 2007 at 08:30:44PM +0200, Steffen Moeller wrote :
> Hi, 
> 
> I returned from Pisa last Saturday and think our presentation was 
> well-perceived. There may be an adoption of our R packages or of our 
> principles for a dynamic installation of packages (possibly also comprising 
> our R packages) by the BioinfoGrid EU project people (that run Scientific 
> Linux (still)) but this is too early to tell, really.
Good to know it ! That's a good point.

> I'll run into them in 
> September again when we are planning for a workshop+conference
> of our NorduGrid in Copenhagen 
> (http://indico.hep.lu.se/conferenceDisplay.py?confId=318, some fancier web 
> site is in preparation). Things will operationalise then. You are all invited 
> to join in for that meeting, btw. It is a grid school plus workshop plus 
> conference, no fees, there may be support for travel for students.
> 
> A major issue that the Grid people also run into is the synchronised update of 
> bioinformatics databases, alike the issues recently discussed on Debian-Med 
> and Debian-Devel.  I hope that we can think of some szenarios that allow for 
> an active involvement of the community in this issue.
Well, If I have followed the threads correctly, the point was
  The size is a bit too big.
  We(deb) can only provide some script to download the database (script which
  are mainly existing outside of Debian today).

Or are you referring to something else ?


> There are differences 
> between the packages we provided for R to those that represent databases only 
> (or AnnotationData as BioConductor puts it of which we ignore the bigger full 
> genome ones). I just need to make up my mind about what these differences 
> are.
Some of the small data package of Bioc are need for compiling some
other Bioc packages (If I remember correctly).
Mainly during the autotest of the package, load a known dataset, and
play with-it.


> Back to the bits with quick tangible successes:
> 
> I just ran against the following bug:
> 
> Error: package 'Biobase' 1.12.2 was found, but >= 1.13.41 is required 
> by 'affy'
> 
> for the compilation of affy a readily compiled version 1.4 of biobase was 
> available but not installed since an earlier version was already installed.
> 
> I might fix this any time unless anybody of yours comes first. I just refrain 
> from the building of packages just happening in the background.

I started squashing this one. It's related to a previous discussion. 
Some code are ready but not yet finalise. (grep experimental cran2deb)
I will try to commit a partial patch which is not breaking everything.
(split between debian and alioth, and making all comparaision using the
deb perl api)

> Btw, how can we automatically sign the packages we compile without entering 
> our keys all the time? Some gpg-agent magics?
Yes !
But this does not solved the Release.gpg problem!



david



More information about the pkg-bioc-devel mailing list