Re: cleaning up messy capitalization to transfer a web page from wintoes

[Levi Waldron <levi.waldron@utoronto.ca>]

On July 8, 2003 11:24 pm, K S Sreeram wrote:
> Check out this page:
>
> http://linuxvm.org/Info/HOWTOs/win2lin.html
>
> Looks like it has what you need

This script would be great, except that it doesn't actually work.  I first got 
an error because the end of one line was missing a \.  Then it ran, but 
didn't actually correct the capital letters in a way to fix broken 
hyperlinks.  It seemed to lower-case the first letter in the <a href=... but 
do nothing to the filenames?  Oh well.

If anyone's curious to look at the short bit of code, here it is.  Maybe the 
problem would be obvious to someone who really knows their sed and shell 
programming.  (I've fixed the first bug, the missing \ at the end of the line 
-e 's/|/ /' \)

#!/bin/bash
D='/tmp/htmlrename/firelab/'
grep -ni "<a .*href=" `find ${D} -name \*.html` \
   | sed -e 's/:.*href="/|/i' -e 's/".*//' \
         -e 's/<a//' \
         -e 's/#.*//' \
         -e 's/|/ /' \
   | grep '\..*\.' \
   | grep -v http: \
   | sort -u \
   | while  read source link
      do
         fd=`dirname $source`
         EF=$fd/$link
         [ -n "${link}" ] && \
            [ ! -f ${EF} ] && \
            {
               echo EF=${EF}  s=$source d=$fd l=,$link,
               AF=`find ${D} -iname ${link}`
               echo AF=${AF}
               cmd="grep -ni $link $source /dev/null"
#              echo ${cmd}
               ${cmd}
#              grep -i "$link" $source /dev/null
               echo
            }
      done