[WIP] Enable option of localhost test server

[sten0]

Continuing from #64

What I had in mind was something like this PR. Unfortunately it doesn't yet work correctly, and seems to have exposed an unhandled failure case:

   passed  27/32  websocket-server-close
Invalid client headers found in: 


   passed  28/32  websocket-server-filter
   passed  29/32  websocket-to-bytes
Websocket client connection: Host header not found
Websocket client connection: Upgrade: websocket not found
Websocket client connect: No key sent
Websocket client connect: No key sent
Websocket client connection: HTTP/1.1 not found
   passed  30/32  websocket-verify-client-headers

Anyways, I thought I'd take a quick stab at the problem today, and this is what I came up with. Sorry if there are any dumb errors, I will confess it's rushed work.

[ahyatt]

Thanks for the change. I should have time this weekend to take a close look at it.

[sten0]

Thanks for the change. I should have time this weekend to take a close look at it. Thanks, much appreciated! For a preview, it looks like we'll need to add

a workaround for the TLS issue you fixed when using the echo.websocket.org (I forget the host name) server. If I find a space of free time long enough to copy your solution (crediting you of course) before the weekend, then I'll give potential fix a shot. ``` passed 2/32 websocket-check gnutls.el: (err=[-110] The TLS connection was non-properly terminated.) boot: (: priority NORMAL:%DUMBFW :hostname 127.0.0.1 :loglevel 0 :min-prime-bits 256 :tru stfiles (/etc/ssl/certs/ca-certificates.crt) :crlfiles nil :keylist nil :verify- flags nil :verify-error nil :callbacks nil) Buffer " *stream buffer*" has a running process; kill it? (yes or no) Test webso cket-client-with-remote-server backtrace: yes-or-no-p("Buffer \" *stream buffer*\" has a running process; kill process-kill-buffer-query-function() kill-buffer(#<buffer *stream buffer*>) #f(compiled-function () #<bytecode 0x4973c9>)() open-network-stream("websocket to wss://127.0.0.1:9999" nil "127.0.0 (condition-case nil (open-network-stream name nil host port :type ty (if (eq type (quote plain)) (make-network-process :name name :buffer (let* ((type (if (equal (progn (or (and ... t) (signal ... ...)) (ar (if (member (progn (or (and (memq (type-of url-struct) cl-struct-url (let* ((name (format "websocket to %s" url)) (url-struct (url-generi (progn (let ((--cl-keys-- --cl-rest--)) (while --cl-keys-- (cond ((m (let* ((protocols (car (cdr (plist-member --cl-rest-- (quote :protoc websocket-open("wss://127.0.0.1:9999" :on-message (lambda (_websocke (let* ((tls-checktrust nil) (wstest-closed nil) (wstest-msg) (wstest (progn (let* ((tls-checktrust nil) (wstest-closed nil) (wstest-msg) (if (>= (string-to-number (substring emacs-version 0 2)) 24) (progn (lambda nil (if (>= (string-to-number (substring emacs-version 0 2)) ert--run-test-internal(#s(ert--test-execution-info :test #s(ert-test ert-run-test(#s(ert-test :name websocket-client-with-remote-server : ert-run-or-rerun-test(#s(ert--stats :selector t :tests [#s(ert-test ert-run-tests(t #f(compiled-function (event-type &rest event-args) # ert-run-tests-batch(nil) ert-run-tests-batch-and-exit() eval((ert-run-tests-batch-and-exit)) command-line-1(("-l" "package" "--eval" "(add-to-list 'package-direc command-line() normal-top-level() Test websocket-client-with-remote-server condition: (error "Error reading from stdin") FAILED 3/32 websocket-client-with-remote-server ```

[ahyatt]

BTW, do you have FSF papers signed? I will need them in order to merge any significant change.

[sten0]

Hi Andrew, Andrew Hyatt <notifications@github.com> writes:

BTW, do you have FSF papers signed? I will need them in order to merge any significant change.

I have not, and won't be able to complete that process in a reasonable amount of time. On the upside, you'll see that the bulk of the diff is the resurrection of the Tornado server test.py file. I don't consider any of my changes significant (by the FSF definition), but if you think they are then please consider them to be © me under the CC0 license. This is a GPL-compatible license that is pretty much public domain that is protected in countries for which public domain doesn't exist. eg: no attribution required, plagiarise away! :-p As I mentioned, I don't think they're significant changes... As far as I can tell this CC0 method works thus: 1. If the couple of lines are significant for ©, then consider them CCO 2. Given this, you can claim them as your own and do anything you want with them 3. Including asserting you hold copyright to them 4. Thus, the effect is similar to as if I had formally assigned copyright to you. 5. You then have the ability to exercise your right as copyright holder to assign your rights to the FSF. I'm not a lawyer, but this seems like a logical, and simple workaround to FSF bureaucracy. Cheers, Nicholas P.S. if you want to make the big changes on your branch I'd be happy to rebase to make it 100% clear that my contributions are ≤10 lines (oh, that's new, it used to be 25 for GNU projects!)

[ahyatt]

Generally, this looks good - I agree, it doesn't look like more than 10 lines of actual changed code. I'll work on completing this shortly.

[ahyatt]

I've put this into a branch on my side, and added commit 5aaf9d1, which makes everything work in a more or less seamless fashion. Let me know if you have any thoughts, if not, I'll merge this into the main branch.

[sten0]

Hi Andrew, Andrew Hyatt <notifications@github.com> writes:

I've put this into a branch on my side, and added commit 5aaf9d1, which makes everything work in a more or less seamless fashion. Let me know if you have any thoughts, if not, I'll merge this into the main branch.

I've tested your additional work and it looks good to me :-) Just to be safe, I'd like to confirm that you'll be doing a simple merge, and not a rebase, because I'd like to push an updated Debian package asap, and would like to keep our branches in sync. Why the rush? Autoremoval in one week ;-) https://tracker.debian.org/pkg/emacs-websocket We met the deadline with time to spare :-) Thanks again, it was a truly a pleasure collaborating on this issue! Nicholas

[ahyatt]

I've merged in your changes, plus my commit. Thank you for your contribution!