bagder

The idea is to make a two and a half hour (give or take) video course explaining and detailing curl, the command line tool. How it works, how to use it, from the basics to some more advanced uses. This will be done by Daniel Stenberg, founder and lead developer of the curl project.

The recording and live-stream is scheduled for August 31, 2023. In the US morning and Euro evening.

See blog post for details.

The project (10 min)

bagder

scoring h2
TLS Handshake
  localhost...ok.
  curl.se...ipv4...ipv6...ok.
  nghttp2.org...ipv4...ipv6...ok.
httpd downloads
  https://one.http.curl.se:47541/score1.data: single...serial...parallel...ok.
  https://one.http.curl.se:47541/score10.data: single...serial...parallel...ok.
  https://one.http.curl.se:47541/score100.data: single...serial...parallel...ok.
httpd requests

bagder

#!/usr/bin/perl

# [scheme][divider][userinfo][hostname][port number][path][query][fragment]

my $num = $ARGV[0];

sub get_part
{
    my ($part, $g) = @_;
    my @a;

bagder

/***************************************************************************
 *                                  _   _ ____  _
 *  Project                     ___| | | |  _ \| |
 *                             / __| | | | |_) | |
 *                            | (__| |_| |  _ <| |___
 *                             \___|\___/|_| \_\_____|
 *
 * Copyright (C) Daniel Stenberg, <daniel@haxx.se>, et al.
 *
 * This software is licensed as described in the file COPYING, which

bagder

/***************************************************************************
 *                                  _   _ ____  _
 *  Project                     ___| | | |  _ \| |
 *                             / __| | | | |_) | |
 *                            | (__| |_| |  _ <| |___
 *                             \___|\___/|_| \_\_____|
 *
 * Copyright (C) Daniel Stenberg, <daniel@haxx.se>, et al.
 *
 * This software is licensed as described in the file COPYING, which

bagder

HTTP/3 options

• "HTTP/3 please". Attempt to use HTTP/3 but fallback to h1/h2 if h3 does not work. This option makes curl try h3 first and if that has not succeeded after N milliseconds, a second attempt using TLS+TCP is initiated and then the first to connect wins and is used. This might then also make h3 not get used if it is too slow. In a future, if a significant portion of servers run HTTP/3, we could consider enabling this option by default.

• "HTTP/3 only". Only try HTTP/3 against the given server and fail if h3 was

bagder

diff --git a/lib/http2.c b/lib/http2.c
index 3c8674cbf..ead356307 100644
--- a/lib/http2.c
+++ b/lib/http2.c
@@ -391,12 +391,14 @@ static bool http2_connisdead(struct Curl_cfilter *cf, struct Curl_easy *data)
        not in use by any other transfer, there shouldn't be any data here,
        only "protocol frames" */
     CURLcode result;
     ssize_t nread = -1;
 

bagder

cd tests/httpd
$ pytest
===================================== test session starts =====================================
platform linux -- Python 3.10.9, pytest-7.2.0, pluggy-1.0.0+repack
httpd: [/usr/sbin/apache2]
rootdir: /home/daniel/src/curl/tests/httpd
collected 0 items / 1 error                                                                   

=========================================== ERRORS ============================================
______________________________ ERROR collecting test_01_basic.py ______________________________

bagder

$ cd tests/httpd/
$ pytest
===================================== test session starts =====================================
platform linux -- Python 3.10.9, pytest-7.2.0, pluggy-1.0.0+repack
httpd: [/usr/sbin/apache2]
rootdir: /home/daniel/src/curl/tests/httpd
collected 0 items / 1 error                                                                   

=========================================== ERRORS ============================================
______________________________ ERROR collecting test_01_basic.py ______________________________

bagder

HackerOne feedback

The site is slow

Going to the page and render it can take many seconds, even on a powerful machine with fast Internet.

Sometimes it seems that clicking on a notification does nothing and you have to reload the page.