Inesita's livereload works! https://www.youtube.com/watch?v=7h6wn8tF9Rw
Needs lot of cleanups but it works :tada:
@fazibear sweet!
@fazibear retweeted!
whats inestia?
is there an opal library for encrypting data before transport via ajax?
@/alll need opinions on opal/opal#1181
that also matches equivalent ruby gem library?
Thermatix, not that I know of
sigh
can you point me in the right direction for encryption/decryption?
what kind of encryption do you need?
I want to encrypt form data before sending to the back end
I also want to know of a good js password hasing library
@Thermatix what's wrong with using https? Why encrypt in JS? This is generally considered a Bad Idea AFAIK
does sinatra/padrino suppourt https out of the box?
@Thermatix yes and no. HTTPS is usually handled by the web server (apache, nginx) that proxies to your application server (webrick, thin, puma). If you expose your application server directly to the interwebs (without a front proxy), you can make webrick/thin/puma handle https for you. This is not how it's usually done for production apps, but each of these rack servers supports https.
I c so beyond changing the protocol to https, I don't have to do anything else?
how do I ensure that the protocol is set to http when doing an ajax request in opal?
I'm using the jquery library for ajax requests
@Thermatix you still need the certificate, I suggest finding a tutorial of some kind depending on what you need (e.g. https://duckduckgo.com/?q=https+ruby+server+setup&t=osx&ia=qa)
@Thermatix correct, once https is set up on the server side, all you have to do it use https:// urls instead of http://. Opal does not care about that, and neither does jquery. If you load the initial html page over https, you can make ajax calls to relative urls (no need to specify https etc), and these requests will be going through https. Hope this makes sense. In addition to what @elia said, you can try to use a self-signed SSL certificate to test with. The browser will complain that the connection is not secure (since your self-signed certificate does not come from a "trusted" source), but that's ok for testing purposes.
can I use opal to then ensure that the protocol is set to https in the first place?
just do not expose a clear-text port
I think you can check this client-side, but a better approach is what @meh suggested: either do not expose a clear-text port (http port 80) at all, or check on the server-side if a request is using SSL and if not, redirect to the https URL.
@Thermatix I think the general theme here is: do not make encryption a client-side concern in any way :smile:
ok
@elia re the
opal -R server my_code.rb trick that you mentioned, I am getting Empty reply from server for http://localhost:3000 What am I missing?
@vais not sure, I used it with a pipe:
pbpaste | opal -R server, does it work that way?
let me try...
@elia my bad, the issue seems to be with my vagrant forwarded port (I decided to use Vagrant running Ubuntu even on my Mac, so I have the same environment everywhere. The jury is still out on whether this was a good idea :wink2:)
As they say, you can always add another layer of indirection.
@vais go docker!
and eventually dind!
@elia Is it possible to run generated tmp/mspec_node.rb this way? As I know, it uses some node-specific stuff...
@elia :weary:
@iliabylich maybe try mspec_phantom instead :wink:
@iliabylich rake should output the actual commands it runs and it should be fairly easy to reuse them adding that option (and probably remove the node/phantom specific
-r requires)
@elia wow. I'm so tired of debugging through
node debug .... Thanks you so much :)
@elia hmm, so it's not my vagrant port forwarding, because other web things work (even just
ruby -run -e httpd . -p 3000) It's gotta be something in that server cli runner, but wth? I'm not having a good day today, one mystery after another :smile: Thought maybe it's thin, but the same result with webrick. Better luck tomorrow perhaps.
@iliabylich the only regret is that I didn't publicize it more :)
@vais does it work from inside the machine? (pseudocode:
vagrant ssh curl localhost:3000)
@elia yep, it does with curl inside the ssh session, but not from the outside (outside Chrome shows me the empty reply error).
I have used this vagrant setup with rails and sinatra before, using thin and webrick, so I have no idea what gives in this case.
how do I do a setTimeout in opal without using native javascript?
@Thermatix you can use native (which is already required if you're using opal-jquery) in this way:
$$.setTimeout(-> { callback code }, 500)
thanks that worked
@Thermatix As I remember, there is a method
delay in opal-browser that takes a timeout and a block
thanks but I don't really want to install an entire gem just for a single function
although...
it might be worth using over the jquery library
do I need jquery included for opal-browser?
I don't think so, there is no
opal-browser dependency in opal-jquery's gemspec
I mean do I need the jquery library to use opal-browser
No,
opal-browser is a wrapper for a native browser functionality
AH ok, thats what I wanted to know, thanks.
in that case I think I'm going to ditch the jquery library and switch to opal-browser
Thermatix, opal-browser is cherry-pickable
if you just need delay, just require 'browser/delay'
and it will just be one class
Thermatix, if you find it's not cherry-pickable enough and it's bringing in stuff you don't want, open an issue and I can make it more granular
... it's saying that Browser::HTTP doesn't exist...
because it's cherry-pickable
require 'browser/http'
i'm requiring 'browser' for the moment
yes
require 'browser'; require 'browser/http'
I c, so I can't just require 'browser'
'browser' requires only the basic browser stuff
events, window, dom and cssom
if you need additional stuff you also have to require it
I seee
and browser/delay is not a basic thing?
nope
you have to require it too
gotcha
hmmm I don't think it mentions that in the readme for opal-browser
doubt it does
shouldn't it?
probably
:D
why doesn't
object.inspect return the object and it's attributes? only the object itself?
what object?
lets say I'm doing:
Request.get("/api/forms/tree/#{id}") do |result|
result.on :success do |response|
puts response.inspect
end
endit only displays #<Request:0x2d6> and nothing else I have to do puts response.instance_variables to see anything and I'd still have to iterate over each one to find the value of each attribute (if they are a public attribute at all).
Thermatix, because inspect isn't implemented for it, pull request welcome
isn't .inspect a basic ruby object method that standardly returns an object and it's instance variables and there values?
@Thermatix Not for every single object. For example, for Object - yes, for String - no :)
This message was deleted
This message was deleted
lets try that again
I'm having some trouble sending JSON data to my server.
here's my request class that makes the request:
require 'browser/http'
class Request
class << self
%w{get post delete put patch head}.each do |http_verb|
define_method http_verb do |*args,&block|
self.make(http_verb,*args,&block)
end
end
def make(type,url,payload=nil, &block)
data_hash = {dataType: 'json'}
data_hash.merge!({payload: payload.to_json})if payload
Browser::HTTP.send type, url, *data_hash do |response|
puts response.inspect
# results = [response.status_code,response.json]
if block_given?
block.call(*response)
else
response
end
end
end
end
end
and I'm using it like so:
def handle_validation event
value = event.target.value
::Request.post("/api/forms/validate/#{Form.tree[:meta][:id]}",format_validation(event)) do |result|
result.on :success do |response|
self.tree = response.json
set_pair_validation_field_value true, value
set_validation_messages 'correct', false
end
result.on :failure do |response|
set_pair_validation_field_value false
set_validation_messages response.json[:data][:message], true
end
end
end
the problem is that the server is receiving the data like this and erroring:
unexpected token at 'dataType=json&payload=%7B%22meta%22%3A%7B%22context%22%3A%22form.validation.check%22%2C%20%22target%22%3A%22form.account%22%7D%2C%20%22data%22%3A%7B%22field%22%3A%22user%20name%22%2C%20%22value%22%3A%22%22%7D%7D':
all
format_validation does is just make sure it's in the correct hash structure, nothing more
Check out this code: https://github.com/opal/opal-browser/blob/master/opal/browser/http/request.rb#L297 (parameters is your
data_hash)
Thermatix, the problem is HTTP::Request is not a normal object
most of the stuff that you care about isn't in instance variables, but on the underlying native object
When you pass a
Hash to the send method, Browser::HTTP::Request encodes it as "key1=value1&key2=value2" (while you need json). The solution is: encode your data with to_json manually before passing it to Browser::HTTP.send
I see
does it pick up on JSON automaticly so you don't need to tell it that your sending JSON?
automatically*
@Thermatix The only http-related example in readme is a simple http get (without any parameters), so start reading the source :)
Thermatix, nope
so how do you tell it the content type? because
hash[:payload] is a JSON string
well, first of all you have to use post
wait, I think I get it
so just HTTP.post(url, hash)
Thermatix, there is documentation
almost everything has docs
how do I set the content_type header on the request object when
It seems strange that
Browser::HTTP.post returns a promise and not the request object?It seems strange that
Browser::HTTP.post only takes a url and the payload but not the payload type, I don't see how to set the content_type when I don't get access to the request object directly nor can you pass it into the request object.
Thermatix, read the docs
you just pass a block to post
it yields the Request object
It doesn't, it returns a promise
yields, not returns
HTTP.post(url, data) { content_type 'application/json' }
that will also return a promise
when I do:
Request.get("/api/forms/tree/#{id}") do |result|
puts result
endI get a promise not a request object is what I meant
Thermatix, `HTTP.get("/api/forms/tree/#{id}").then { |result| }
for the post
HTTP.post(url, hash) { content_type 'application/json' }.then { |result }
I realised what I was doing wrong, but It's now saying
NoMethodError: undefined method 'content_type=' for #<Request:0x2d6>
it's
content_type not content_type=
it's made so you don't have to provide an argument
it's either
HTTP.post(..) { |r| r.content_type 'application/json' } or `HTTP.post(..) { content_type 'application/json' }
got it, thanks for the help.. I don't think I'm firing on all cylinders atm
Thermatix, no worries, it ain't your fault, I suck at explaining things :P
@meh I think you're doing great :+1:
@elia, @vais, I still think we're fucked one way or the other, I didn't consider the fact
$$s is currently used for the inheret self of a block
this means we have to figure something else out
actually, I think it's easy
basically
$$s remains, but instead of the common check being self = $$s || this, it becomes self = $$m.self || $$s || this
that way we still DO NOT have to ensure the caller try/catches it to reset it properly, and we get all the benfits
that should work
@meh sorry, what is
$$m.self?
vais, the new metadata stuff
it will probably be
$m or $meta or $metadata in the prelude
does opal support
prepend ?
nope
not yet
damn
would have been awesome, oh well
would easily allow me to add method extentions
umm question are cookies sent with every ajax request?
yeah
by the browser
kk
@/all 0.9.0.beta2 is out! here's the new stuff from beta1: https://gist.github.com/elia/72872e47b23423228f09
(would post the rubygems tweet but apparently I have some connection issues ๐ )
@elia my connection seems fine, so here's the tweet: https://twitter.com/rubygems/status/662404555448164352