R devel - Apr 2009 - Linux Huge R/lib64/R/doc/html/search/index.txt (PR#13683)

Full_Name: Bruce Foster
Version: 2.9.0
OS: Red Hat Linux 5.3
Submission from: (NULL) (129.105.110.38)


The R/lib64/R/doc/html/search/index.txt file for R2.9.0 is much too large and
contains multiple copies of its entries. In comparison with R 2.8.0:

[root at seldon search]# pwd
/sscc/opt/R-2.8.0/lib64/R/doc/html/search

[root at seldon search]# ll index.txt 
-rw-r--r-- 1 root root 4496494 Nov 11 16:39 index.txt

[root at seldon search]# grep "^Entry: abind" index.txt | wc -l
1

[root at mule2 search]# pwd
/sscc/opt/R-2.9.0/lib64/R/doc/html/search

[root at mule2 search]# ll index.txt
-rw-r--r-- 1 root root 1858821960 Apr 28 21:31 index.txt

[root at mule2 search]# grep "^Entry: abind" index.txt | wc -l
732

There are 732 copies of the abind entry in 2.9.0! 

The two installations are comparable, in that they are standard builds of R with
the addition of 18 overlapping task views, which were individually installed
with the command update.views("viewname").

1. The index is too large to work with the web search engine. All searches
fail.
2. Can I rebuild this index file somehow to eliminate multiple entries?

Thanks!
Bruce