New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Implement Travis CI #80
Comments
Congratulations! That's great work! |
Thanks Brett, this was indeed quite a big job, and we're not there yet as there are many fails now, but it's an essential first step in raising - and keeping quality at a higher level! PS This will eventually also make it very easy to judge if merging PR requests would cause any regression as travis integrates well with github: but obviously we first need to fix the ~80 fails before it becomes truly useful |
It looks like a lot of the array tests may be failing because the example has a (required) dependency on the large array() function (the tests are for supporting the ordered "associative arrays" with a INI change). |
And the "shuffle" array function is failing because the example is demonstrating random behavior. |
"lcg_value" is also a random issue.... Though I've fixed with the header I found in cli.js, as for shuffle, but not for the arrays. |
Sorry, I see the problem as far as arrays may be my own problem with the more recent INI-array functionality. |
|
Finally, do you know what is causing a problem with strptime? |
Nevermind about |
Sometimes the examples rely on default settings, but the cli is apparently remembering the old ini values from previous examples. Is there some way the environment could be re-set up after each function's examples are run (if not each example), or would that add too much of a load? |
How about we reset the phpjs object before each function is evaluated here https://github.com/kvz/phpjs/blob/master/tests/phpjsutil.js#L260? By the way awesome work on the fixed Brett! 💖 |
Thanks :) And sure, that sounds like it ought to work. Btw, you saw my other bazillion comments above too? |
Hey Brett, Sorry I didn't get to your messages earlier! Here goes:
Fixed in 0738285
Passes now, thanks
👍
Not sure about this one
Not in favor of an otherwise very useful function like
I thinks as a rule of thumb, for type fetching we should stay as close to php as possible. If people want good/strict JS type fetching they should use things like underscore/lowdash. That's not something we can ever / should win nor aim for. So in short, if it's an 'associative array' we return array, cause that's what would make code work in php. It's not pretty, but it's as consistent as this project can try to be (I'm thinking of people trying to run php code in V8).
I already hacked Let me know what you think! PS It makes more sense as php is a serverside thing and node is too. It would greatly simplify our functions ( |
Replies below, but first, I hope you can forgive my long reply below, On 1/18/2014 7:04 AM, Kevin van Zonneveld wrote:
I think our approaches may diverge here due to your strong focus on the With better standardization in the browser, there is less of a need for While I can see your argument about PHP and Node both being server-side, As far as the likes of chmod not having a place in the browser, believe We can all no doubt agree that whenever developers access potentially But I presume you wouldn't be happy with someone working on the Node Or, to take a client-side example where one is accustomed to privileges, What is so inherently different between OS binary executables being There is no difference whatsoever that I can see between binary So, I disagree on chmod not having a place in the browser; it could be In my opinion, I even think that if the browser did more to enhance That all being said, I have to admit that for now, there is |
Apparently since a round of js-beautify changes, it appears Travis building is broken: "Error: Cannot find module 'underscore'". Any ideas what is needed to put this back on track? |
Where are you seeing this? Latest build seemed to work "fine" https://travis-ci.org/kvz/phpjs/builds/18664970 ? |
Strange...seems ok now, thanks! |
Can you add |
Or if it makes a difference in your testing environment maybe it needs to be |
Alright, added |
Hmm... I think the challenge is that it needs to be a circular reference...Is there a way to ensure window.window refers back to window (and that this will refer to window so that this.window refers to window.window and thus to window)? |
Ah I think get it now, sorry. How about this? 5a50fe1 Also: thanks a lot for the fixes! |
Thanks for the commit, but even the latest one doesn't fix it. For testing, I think the easiest one to follow is |
Ah you require everything to actually live inside window. In the testsuite, This works now though: function is_finite(val) {
window['isFinite'] = 'Brett';
return window['isFinite'] + ' - ' + window.window['isFinite']; $ node bin/phpjs.js --action test --name is_finite
returned:
"Brett - Brett" But I doubt that's a big help to you then. (~2am here, need to 💤 now) |
Profit because:
This will be a big boost for php.js quality in the long-run.
Anybody who feels up to this task and make name in php.js history?
The text was updated successfully, but these errors were encountered: