I want to xhr some opal script, and use ruby code defined there. I tried to get it with $.getScript. But no success for me.
$.ajax({
url: 'assets/foo.js.rb',
success: function(data){
#{ClassInFooJs.new}
},
dataType: "script"
});
Moreover the script is evaled (in strange manner), I mean I can call Opal.modules["file_i_required"] from console, and it will return basically the compiled code.
BUT inside of it nothing is evaled, (no console.logs, window["foos"] = #{something}), and I can't reference anything from that file.
Any help?
You should use either require "foo" from Opal or Opal.require("foo") from JavaScript. If the load "foo" behavior is required should be noted that Opal.load("foo") is also present.
The alternative is to compile files on the server with the :requirable flag set to false but that's kinda difficult unless you compile manually. See the API docs for more info.
Calling directly Opal.modules["foo"](Opal) should be avoided as it doesn't properly update $LOADED_FEATURES and in general can change in the future.
Related
I have installed angular universal on my app.
Running npm run build:ssr - DONE. WORKS.
Running npm run server:ssr - DONE.WORKS.
After accessing the server URL (localhost:4000), the page is not fully loaded and the following error is raised on the Terminal:
I also faced the same problem, so I would just like to share my findings for the same.
For me, there were two plausible causes/solutions for it:
First, in my project's I18N default JSON file that was en.json, I was having a problem In the structure of the JSON file.
For example, I had the below mistake. I missed the comma after the second label 'FINISH' :
{
"COMMON": {
"EDIT": "Edit",
"FINISH": "Finish"
"QUIT": "Quit",
}
}
So after correcting the structure, the application ran fine without an error.
Secondly, another cause of the issue could be, at runtime transloco was not able to find the correct label in the selected language, so it looked for a fallback language and it could not even find that in the transloco-root.module.ts so after adding my fallback language, it tried to find the same in the fallback language as specified in the transloco-root.module.ts.
So it found out that label and the issue got resolved.
BUT in the second solution provided, you need to have that incorrect label in at least that fallback language's json file in correct format.
I added the fallback language like below:
useValue: translocoConfig({
availableLangs: ['fr', 'en'],
defaultLang: 'en',
reRenderOnLangChange: true,
fallbackLang: 'fr',
prodMode: environment.production,
missingHandler: {
logMissingKey: true
}
})
i18n Transloco wasn't fully configured on the module file.
In my Puppet module, I have something like this:
file {'myfile':
ensure => 'file',
path => '/whatever/myfile',
content => inline_template(file(
"puppet:///modules/my_module/templates/${domain}/${::hostname}_myfile.erb",
"puppet:///modules/my_module/templates/${domain}/myfile.erb"
))
}
And my spec looks like:
require 'spec_helper'
describe 'my_module' do
context 'with defaults for all parameters' do
it { should compile }
end
end
If try to run it, I get:
Failure/Error: it { should compile }
error during compilation: Evaluation Error: Error while evaluating a Function Call, Could not find any files from puppet:///modules/my_module/templates/dev.contaazul.local/myfile.erb, puppet:///modules/my_module/templates/dev.contaazul.local/myfile.erb at /home/carlos/Code/Puppet/modules/my_module/spec/fixtures/modules/my_module/manifests/init.pp:48:33 on node becker.local
Obviously, it cannot find any of the ERB templates. If I replace the puppet:// part for /home/carlos/Code/Puppet/ (where the code actually lives), it passes, but in production it is /etc/puppet/, so, it will not work.
How can I make this work?
RI Pienaar has released a code snippet for a function with this behavior. You will need to copy this code into a file in one of your modules at the path lib/puppet/parser/functions/multi_source_template.rb.
Once you do that, you should be able to use it like so:
file {'myfile':
ensure => 'file',
path => '/whatever/myfile',
content => multi_source_template(
"my_module/${domain}/${::hostname}_myfile.erb",
"my_module/${domain}/myfile.erb"
)
}
As to why the original approach doesn't work: URLs are usually used with the source property only and transferred to the agent as is. The agent consumes the URL and makes an according request to a Puppet fileserver (which is just another master).
The file function on the other hand (as used here with the content property in conjunction with inline_template) will not process URLs and expects local paths instead.
All that being said, the issue could possibly have been side-stepped by specifying both the paths for the test box and the production system, with the latter just acknowledging that /home/carlos is missing. But that would have been far from a clean solution.
I need to be able to report qUnit tests to a file so my build server can parse them.
I'm using qUnit (grunt-contrib-qunit) through Grunt along with the jUnit reporter found here.
I can get the report to write to the log just as it states but I'm having trouble getting it into a file. I've tried qunit callbacks in my gruntfile but none of them seem to get the xml info. I also tried to simply redirect stdout but it (of course) printed all of the non-xml command-line stuff along with the xml.
In short, I've got the XML echoing properly in the console.log statement. I just need to get this to a file somehow. Either through Grunt, phantomjs, or any other means.
Well, if you're running QUnit tests from Grunt, then you have the full power of Node at your disposal. I've never used that JUnit plugin, but if it just gives you callback in your QUnit HTML file, then you would need a browser solution (even if that is phantomjs).
Phantom uses QtWebKit, which has implemented the File API so you could implement a solution using that from JUnit's callback, but, of course, that would fail if you run the tests in certain other browsers (namely IE9 or under). Here's how that might look (no guarantees on this being exact, I have not run it):
QUnit.jUnitReport = function(report) {
function onInitFs(fs) {
fs.root.getFile('qunit_report.xml', {create: true}, function(fileEntry) {
fileEntry.createWriter(function(fileWriter) {
fileWriter.onwriteend = function(e) { /* if you need it */ };
fileWriter.onerror = function(e) { /* if you need it */ };
var blob = new Blob([report.xml], {type: 'application/xml'});
fileWriter.write(blob);
}, someErrorHandlerFunction);
}, someErrorHandlerFunction);
}
window.requestFileSystem(window.TEMPORARY, 1024*1024, onInitFs, someErrorHandlerFunction);
}
And again, if you need to do something to write the file in IE9 or under (or some mobile browsers) you'll need another solution, like kicking off an ajax request to upload the data to a server that stores the file. You could even run that "server" from within Grunt and have Node write the file.
I'm building an MMO using Node.js, and I'd like some scripters from my team to create scripts (duh) for bosses and other scripted objects. My first thought was to have a folder on the server where people can upload javascript files and let node.js automatically read and parse every script in the directory, but I don't want them to be able to execute process.exit() for example, or other hazardous stuff.
Is there a solution that lets me control what functions the scripters are able to call?
You can control what functions are they unable to call with vm module.
For example,
vm.runInNewContext(userCode, {
require: null,
process: null,
someFunc: function (x) { return x+1 },
someData: { abc: 'def' }
});
Will javascript work as scripting language. I guess it should.
so, I think vm.runInNewContext is might be all you need.
Have a look at http://nodejs.org/docs/latest/api/vm.html
On POSTing data to my expressjs app, this is what I am getting:
node(58287,0x7fff771ad960) malloc: *** error for object 0x7ff8a8600c58: incorrect
checksum for freed object - object was probably modified after being freed.
*** set a breakpoint in malloc_error_break to debug
Abort trap: 6
Any idea why?
Update:
Here is some code:
Client side:
$.ajax({
url: 'user/' + id,
type: 'POST',
dataType:'JSON',
data: JSON.stringify(data),
success: function(response){
console.log(response);
}
});
Server side:
app.post('/user/:id', function(req,res){
var id = req.params.id;
console.log(data);
});
When I use JSON.stringify on the client side, I am hitting this weird error:
node(58461,0x7fff771ad960) malloc: * error for object
0x7fa861d00e28: incorrect checksum for freed object - object was
probably modified after being freed.
* set a breakpoint in malloc_error_break to debug Abort trap: 6
When I don't use JSON.stringify on the client side, I get 'null' strings on the server side.
Any ideas on what I am doing wrong?
I experienced this same error recently. Here's the fix:
Node.js has had some bugs that cause it to install improperly from source under OS X (see e.g. issue 2061). The good news is that the packaged installer installs it correctly. So, simply uninstall Node, then head to http://nodejs.org/#download and run the Macintosh Installer.
I've reported this bug on the Node issue tracker here.
This is liable to be a bug in nodejs's internals. (Or, if expressjs has any native-code bindings, perhaps expressjs.)
There's no easy way for you to write this kind of bug yourself in JavaScript. If you can reproduce this at will, they'd probably like a bug report. Try to figure out the least amount of code that can reproduce the problem.
I just got this error today, and updating Node.js through MacPorts from 0.8.9_0 to 0.8.10_1 fixed the issue.
It doesn't seem this was specifically adressed though, as bug reports and the changelog don't indicate that (http://blog.nodejs.org/2012/09/25/node-v0-8-10-stable/).
I haven't looked into the cause of these issues (plural - there are clearly at least two bugs somewhere along the line - jquery, node, express), although a summary and workaround are as follows:
When your client side code looks like in your updated question the server obviously shouldn't crash with a malloc error (bug #1), although it is understandable that the query is mistreated, since you're telling jquery to send json and then you send a string.
bug #2 is simply null --> "null" along the pipe. This at least doesn't cause a server crash, because data types match headers (i.e. everyone think we're using json), however someone is converting nulls to strings. My baseless suspicion is that it's express/connect.
Finally, the workaround is simply to wrap your data in a way that manages to be transported, and then unwrap on the server side:
Client side:
$.post({
url: 'user/' + id,
data: {workaround: JSON.stringify(data)}, // no null strings this way
success: function(response){
console.log(response);
}
});
Server side:
app.post('/user/:id', function(req,res){
var id = req.params.id;
var data = JSON.parse(req.body.workaround); // unwrap
console.log(data);
});
If I find time, I'll investigate and try and post a bug report somewhere (it's a problem when you don't know whose fault this is...), please try do this as well.
You are are calling console.log server-side on 'data' which is not defined in the scope of your example.