Baslinux mailing list

[Lee Forrest <lforrestster AT gmail.com>]

On Fri, Feb 23, 2007 at 01:03:21PM -0600, James Miller wrote:
> I've run across a problem that a bash script might resolve. But my bash 
> skills are 
> rudimentary at best. I'd therefore like to ask for help with this project. 
> I've 
> downloaded a grouping of html files I want to have resident on my home 
> computer. The 
> files are linked together in series and link one to another from within 
> each html file 
> as well as from some centralized index pages. The problem is that this 
> grouping of 
> files was created by some (not too bright) M$ user who paid little or no 
> attention to 
> case sensitivity. Sometimes file names are in lower case, sometimes in 
> upper case 
> (mostly upper case). The problem, as should be clear, is that the links 
> don't work on 
> a Linux system if the case is not properly specified in the hyperlink. What 
> I 
> therefore need to do is convert the relevant file names, whether under 
> their resident 
> directory, or within the html files themselves, to either upper or lower 
> case. It 
> could be alot more complex to do this if the file names varied more than 
> they do, but 
> as things stand, there are only 5 words that need their case changed.
> 
> For demonstration purposes, let's say that the relevant words are index, 
> htm, word1, 
> word2, and word3. What would be the automated way to change all instances 
> of index, 
> htm, word1, word2, and word3 to INDEX, HTM, WORD1, WORD2, and WORD3?

> I suppose the routine for doing this would be different for
> renaming the files than it would be for changing the case of
> hyperlinks inside an html file. 

Only the sed part of the script would differ.

> Any help on this would be much
> appreciated.

This script should work as is (when you fill it in). But be sure
to test it first in a directory you create for the purpose with
copies of some or all of the files in question in it.

#!/bin/sh

dir=[where the files are - include trailing forward slash]

for _file_ in  

#list files, space seperated; same single line; include prefix
#$dir

#${dir}$_file -- if all files in dir, the do use ${dir}*  
#etc ; wildcards - sed uses regular expressions

do cat $_file |  \
sed -e 's/index/INDEX/g' \
-e 's/htm/HTM/g' \

#and so on

> tmp/xa2923
mv /tmp/xa2923 ${dir}$_file

done


For bash newbies, put that in a file and make it executable with:

--------------
chmod +rx file   #name it anything
--------------

Then run it  by putting the full path on the commandline:

-------------
/home/you/file     # for example
-------------


Of course, if there are other instances of those strings in the files, 
they'll be
changed too. I just don't think it matters. 

A more complex sed script would fix that.

Lee


-- 
BasicLinux: Small is Beautiful
http://www.basiclinux.com.ru