Subject: Re: changing anchor filenames to lowercase
Tue May 26 08:51:43 1998



>> It would still have to be done by hand -- you could do a global change on
>> .HTML to .html ... but for the file names themselves - you'd have to
>> highlight, then select "change case" -- maybe you could write a macro to do
>> that.
>> 
>> >>Its the references inside each HTML document that are the problem.
>> >>The only solution I have to date is the manual one, but as there
>> >>are about 300 pages I am not too excited at the prospect. Does anyone
>> >>have a better idea?
>
>Perl, perl, perl. A simple script that finds all SRC or A HREF or A
>NAME tags, takes their contents, lowercases them and puts 'm back. My
>perl's a bit rusty so I'm not going to actually post the code, but it
>should be, like, 5 lines of code.....
>
>Should even be possible with some 'perl -pi -e .... ' one-line
>command.


    perl -pi.bak -e 's/(href=".*?")/\L$1/ig;' *.html



which says:


    perl -pi.bak -e 's/(href=".*?")/\L$1/ig;' *.html
                                              ^^^^^^
for every file in this directory whose extension is '.html', 


    perl -pi.bak -e 's/(href=".*?")/\L$1/ig;' *.html
           ^^^^^
create a backup copy (with the extension '.bak'), then write the
output of this command back into the original file.   the
official term for this is "in-place editing", hence the 'i'.


    perl -pi.bak -e 's/(href=".*?")/\L$1/ig;' *.html
         ^^
loop through the file one line at a time, printing the results
of the command (technically, the contents of the "$_" internal
variable) before going on to the next line.


    perl -pi.bak -e 's/(href=".*?")/\L$1/ig;' *.html
                 ^^
decide what to print (and thus write back into the file) by
executing the string which follows.   and that string says:


    perl -pi.bak -e 's/(href=".*?")/\L$1/ig;' *.html
                     ^
perform a substitution


    perl -pi.bak -e 's/(href=".*?")/\L$1/ig;' *.html
                                          ^
for every occurrence of a matching string (not just the first,
as is the default), which makes this a 'global' search & replace
function.


    perl -pi.bak -e 's/(href=".*?")/\L$1/ig;' *.html
                                         ^
ignore case when looking for items that match the target pattern.


    perl -pi.bak -e 's/(href=".*?")/\L$1/ig;' *.html
                        ^^^^^^
the pattern we're looking for starts with the string 'href="' 


    perl -pi.bak -e 's/(href=".*?")/\L$1/ig;' *.html
                              ^^
followed by any number of characters, of any kind,


    perl -pi.bak -e 's/(href=".*?")/\L$1/ig;' *.html
                                ^^
up to the next '"' mark.


    perl -pi.bak -e 's/(href=".*?")/\L$1/ig;' *.html
                       ^          ^
when you find a string of that type, remember it (store it in a
temporary variable named "$1")


    perl -pi.bak -e 's/(href=".*?")/\L$1/ig;' *.html
                                    ^^^^
and replace it with a copy of itself, converted entirely to
lowercase.




after running that, all your URLs should be converted to
lowercase in the original files, but the backups contain
original versions of each file.   if something goes wrong, you
can back up and start again.   if the command did what you
wanted, you can delete all the backup files and be done with it.