diff options
author | Rob Austein <sra@hactrn.net> | 2009-04-29 18:55:48 +0000 |
---|---|---|
committer | Rob Austein <sra@hactrn.net> | 2009-04-29 18:55:48 +0000 |
commit | 309dad3781b2259d9dccf59a87c9896c91f56ebd (patch) | |
tree | 8131a4ba693aac3dbe64b5b1bc1393f26ae15985 /scripts/async-http.py | |
parent | 2c7bea62c4e8a8a25c2e0e23033d37820a9165a3 (diff) |
Notes on test environments
svn path=/scripts/async-http.py; revision=2375
Diffstat (limited to 'scripts/async-http.py')
-rw-r--r-- | scripts/async-http.py | 17 |
1 files changed, 17 insertions, 0 deletions
diff --git a/scripts/async-http.py b/scripts/async-http.py index 66c502de..fe7847b7 100644 --- a/scripts/async-http.py +++ b/scripts/async-http.py @@ -21,6 +21,23 @@ PERFORMANCE OF THIS SOFTWARE. # Command that may be useful for testing server side of this: # # lynx -post_data -mime_header -source http://127.0.0.1:8000/ +# +# Testing the client side of this is more entertaining, both because +# we need to be liberal in what we accept and also because servers do +# wildly different things depending both on HTTP version number and +# options signaled by the client and also on internal details in the +# server context (such as whether the content is static or generated +# by CGI). It's useful to test against static content, but also to +# test against CGI-generated output, eg, the following trivial script: +# +# print "Content-Type: text/plain; charset=US-ASCII\r" +# print "\r" +# for i in xrange(100): +# print "%08d" % i, "." * 120, "\r" +# +# At least with Apache 2.0, result of running this differs +# significantly depending on whether client signals HTTP 1.0 or 1.1; +# the latter produces chunked output. import sys, os, time, socket, asyncore, asynchat, traceback, urlparse import rpki.async |