Re: How can I just get the output name?

On 3/26/24 09:10, Dan Fandrich via curl-users wrote:
> On Tue, Mar 26, 2024 at 07:18:21AM -0700, ToddAndMargo via curl-users wrote:
>> On 3/24/24 05:41, ToddAndMargo via curl-users wrote:
>>> Firefox CliGet gives me this when I click on a particular
>>> download link:
>>>
>>> curl --header 'Host: dl.duosecurity.com'
>>> ...
>>> 'https://dl.duosecurity.com/duo-win-login-latest.exe'
>>> --output 'duo-win-login-4.3.0.exe'
>>>
>>> All I want is the "duo-win-login-4.3.0.exe" text.
>>> Using cURL, is there a way to get?
> 
> This will do it:
> 
>    curl -sI -o /dev/null --write-out '%header{content-disposition}\n' \
>      https://dl.duosecurity.com/duo-win-login-latest.exe | \
>    sed -e 's/^.*="//' -e 's/".*$//'
> 
> This requests only the headers from the site and writes out just the
> Content-Disposition: header, then strips off everything that's outside
> quotation marks. You'd want a more robust header parser if you want to work it
> to work reliably with any file name on any site, but it's likely good enough
> for this one URL. You'll need a curl version at least 7.83.0 to get the %header
> feature.
> 
> Dan

Hi Dan,

Awesome!

$ curl -sI -o /dev/null --write-out '%header{content-disposition}\n' 
https://dl.duosecurity.com/duo-win-login-latest.exe
attachment; filename="duo-win-login-4.3.0.exe"

$ curl -sI -o /dev/null --write-out '%header{content-disposition}\n' 
https://dl.duosecurity.com/duo-win-login-latest.exe | sed -e 's/^.*="//' 
-e 's/".*$//'
duo-win-login-4.3.0.exe

Thank you!

-T
-- 
Unsubscribe: https://lists.haxx.se/mailman/listinfo/curl-users
Etiquette:   https://curl.se/mail/etiquette.html