Downloading a Text file from a Unix server (Apace) using CGI

stumbler has asked for the wisdom of the Perl Monks concerning the following question:

Dear Monks,

I am having a hard time downloading a text file in the correct format from a Apache server (running in Unix).

All I am trying to do is to have a button to download a text file (let us say, 123.inp residing in htdocs folder) from the server. When I click on 'Download Text File' button, it should provide the option to either "Open" or "Save" the file.

Everything works fine except that if I click on 'Open', the file opens as one single line in "NotePad". I know it is because the file is downloaded in "binary" format but I don't know how to fix it. Any suggestion would be of great help

use CGI qw(:standard);
use CGI::Carp 'fatalsToBrowser';

my $query = new CGI;

print $query->header();
print $query->start_html();
print $query->button( 
        -name => "Download", 
    -value=>"Download Text File",
    -onclick=> "window.location.href='/123.inp';"
    );
print $query->end_html();
[download]

Comment on Downloading a Text file from a Unix server (Apace) using CGI Download Code

Replies are listed 'Best First'.
Re: Downloading a Text file from a Unix server (Apace) using CGI by swares (Monk) on May 03, 2007 at 20:28 UTC
Try wordpad. Sounds like the file may be in Unix format and Notepad does not recognize the end of line character. Windows uses a different format... or you can use the unix2dos command to convert it.	[reply]
Re^2: Downloading a Text file from a Unix server (Apace) using CGI by Anonymous Monk on May 04, 2007 at 00:51 UTC
Thanks for the response. I used the ux2dos unix utility to convert the file to windows readable format and it works fine now.	[reply]
Re: Downloading a Text file from a Unix server (Apace) using CGI by almut (Canon) on May 03, 2007 at 21:18 UTC
You could deliver the file's content "dynamically", and do explicit line endings conversions yourself as required. For example, you could have a generic CGI script, which takes the file name as parameter (e.g. `deliver_plain_text.pl?file=123.inp`). The CGI would read the file, convert the line endings (`s/\n/\r\n/`), and write the result to stdout (which apache sends to the browser, as usual). If you don't want to have `deliver_plain_text.pl` appear in the URL, you could use `mod_rewrite` to do some rewriting magic for you... Also, you might want to apply some heuristics to detect whether the client actually is a Windows machine (e.g. by inspecting the User-Agent header), and only do line ending translations in that case. Alternatively, you could try to add an Apache output filter for the appropriate type of files, using `mod_mime`'s `AddOutputFilter`. I'm not aware of a ready-to-use apache filter for just that purpose, but - if performance is not a primary issue - you could use `mod_ext_filter` to pipe the content through a simple external program, which does `s/\n/\r\n/` ...	[reply] [d/l] [select]