After the --dump-header writes a file, how to do read those headers back into the next request? I would like to read them from a file because there are a number of them. <
Easy:
$ curl -H @header_file $URL
... where the header file is a plain text file with a HTTP header on each line. Like this:
Color: red
Shoesize: 11
Secret: yes
User-Agent: foobar/3000
Name: "Joe Smith"
curl had no way to "bulk change" headers like that, not even from a file.
Your best approach with an old curl version is probably to instead write a shell script that gathers all the headers from the file and use them, like:
#!/bin/sh
while read line; do
args="$args -H '$line'";
done
curl $args $URL
Invoke the script like this:
$ sh script.sh < header_file
As answered by @dmitry-sutyagin, if your curl is at least version 7.55.0 you can use the @
notation to read headers from a file:
curl -H @headerfile.txt https://www.google.com/ # requires curl 7.55.0
If your curl is NOT 7.55.0 or newer, there's a useful hack:
-K/--config <config file>
, and put several -H/--header <header>
lines in the text file.For instance:
curl --dump-header foo.txt https://www.google.com/
dos2unix foo.txt
Convert the file to -H 'header'
lines, manually or with a script:
cat foo.txt |
awk '$1 == "Set-Cookie:"' |
perl -ne "chomp; next if /^\\s*\$/; if (/'/) { warn; next } print \"-H '\$_'\\n\";" |
tee headerfile.txt
This might output something like:
-H 'Set-Cookie: 1P_JAR=2018-02-13-08; [...]'
-H 'Set-Cookie: NID=123=n7vY1W8IDElvf [...]'
curl --config headerfile.txt https://www.google.com/
Starting with curl 7.55.0 it can now read headers from a file:
curl -H @filename
It's that easy now.
how about this:
curl -v -H "$(cat headers.txt)" yourhost.com
where headers.txt
looks like
Header1: bla
Header2: blupp
works in BASH.
curl $(xargs -a headers.txt printf "-H '%s'") example.org