[Robast-commits] r870 - branches/robast-1.0/pkg

[noreply at r-forge.r-project.org]

Author: ruckdeschel
Date: 2016-04-23 23:37:22 +0200 (Sat, 23 Apr 2016)
New Revision: 870

Added:
   branches/robast-1.0/pkg/20160423NotToCRAN.txt
Log:
Note for CRAN as to RobAStRDA 

Added: branches/robast-1.0/pkg/20160423NotToCRAN.txt
===================================================================
--- branches/robast-1.0/pkg/20160423NotToCRAN.txt	                        (rev 0)
+++ branches/robast-1.0/pkg/20160423NotToCRAN.txt	2016-04-23 21:37:22 UTC (rev 870)
@@ -0,0 +1,30 @@
+                RobAStRDA once again ... 
+
+sorry to bother you with this again, we have been here before,
+and there is an extensive mail exchange on this on CRAN archives
+(last time 19.09.2013 13:43 "Re: CRAN submission RobAStRDA 0.9")
+and some offline exchange with Uwe on this afterwards.
+
+To cut it short: Our package fails by size. We do not see how
+to fix this; it is a database with interpolators; i.e., grids
+for speed up of computations, but in addition also many interpolating
+R-functions (with their environment) -- the results of approxfun (to 
+produce the interpolating function only once and not each time on the 
+fly on call). One clue might be that we simultaneously store pre R-3.0
+and post R-3.0 environments (and at run-time decide which branch to
+take according to version information), and that the compression does
+not work well in post R-3.0 for pre R-3.0-environments ...
+
+Anyway: Our sysdata.rda data set is xz compressed already but during 
+check gets extracted, throwing the note that our package is too large; 
+lazyload and putting the data somewhere else does not help either (and 
+even gets worse as to size).
+
+Uwe was so nice to offer some of his time to look at it, so maybe you
+find some trick? 
+
+Another way out would be to host the package somewhere else (in some
+drat type repo with infrastructure of Dirk Eddelbuettel's) if we may
+link to non CRAN-repos in a suggests/depends field of another package.
+
+Any help appreciated, best, Peter