I have renamed app.js to app.coffee (as well as translated the contents), but now when I compile I get this error:
Running "validate-imports:tests" (validate-imports) task
>> client/tests/helpers/start-app: Cannot find module "client/app"
This error goes away when I translate the file back to javascript.
I have added the grunt-contrib-coffee and confirmed it works correctly, the problem I believe is that the coffee-script compilation happens after the validate-imports task which looks for .js files in the app folder. Does this need to be tweaked to look in the tmp/javascripts folder where the coffee-script gets compiled to?
Here is the task in question:
// Scripts
grunt.registerTask('buildScripts', filterAvailable([
'jshint:app',
'jshint:tests',
'validate-imports:app',
'validate-imports:tests',
'coffee',
'emberscript',
'copy:javascriptToTmp',
'transpile',
'concat_sourcemap'
]));
Anyone know of this bug?
I found an answer here. Thus, when I tried it, I renamed start-app.js to start-app.coffee, converted the code into coffescript and now it works without that error.
Bryan
One approach to this would be to add this line to your testem.json file:
"before_tests": "coffee -c tests/**/*.coffee"
This should compile the .coffee files within your /tests directory before test execution, meaning they will compile down to their .js equivalent before they are run. While you could technically change the EAK boilerplate from .js to .coffee with a similar trick, it might be better just to write your tests in .coffee, while leaving the default .js testing harness to maintain compatibility with EAK.
You can also remove these files when the test run is over, as shown within the tapas-with-ember testem.json file.
Related
(Been looking for this for ages over internet...).
Been working with very old npm packages recently and seems that they cannot be loaded properly within Rails-6 (I know I could create a question by each cases, but I'd like to learn to handle this myself since every time it is for a different reason).
I'd like (as for debugs and in development mode only) to display the list of available all assets, including js, css, images, and anything else available at public level (the client could load). So it should be a set of compiled assets ?
Similarly to http://localhost:3000/rails/mailers like http://localhost:3000/rails/assets ?
Displaying the list of all available assets (prior compilation, also for debug intents) could also be great.
This might do what you want:
Rails.application.assets.each_file do | pathname |
# ...
end
This will enumerate every file of which Sprockets is aware. You can run that at rails c or build a simple controller to dump the list into a view if you prefer. For more information, see https://stackoverflow.com/a/11005361.
If you're not using Sprockets (e.g. a newer Webpacker-based application) then, obviously, you'll need a different solution.
You can run rails assets:reveal and see all the available assets:
$ rails assets:reveal
application.js
application.js.map
application.css
application.scss
...
To view the full path, use rails assets:reveal:full:
$ rails assets:reveal:full
/home/{REDACTED}/app/assets/builds/application.js
/home/{REDACTED}/app/assets/builds/application.js.map
/home/{REDACTED}/app/assets/builds/application.css
/home/{REDACTED}/app/assets/stylesheets/application.scss
...
I clearly must be Doing Something Wrong here. I'm wrestling with the asset pipeline (again). I have a custom font, and it seems to me to get everything to compile properly I need to use asset_path() in multiple places, but it's having an unexpected effect.
I realize there are several ways to do this, but here's what I have currently:
In application.css.scss.erb:
#import "<%= asset_path("my-font.css") %>";
my-font.css's source file is app/assets/stylesheets/my-font.css.erb (it needs to be an .erb because I am also using asset_path() there as well).
In application.rb I am adding my-font.css to the precompile list.
config.assets.precompile << 'my-font.css'
When I clean out public/assets and run rake assets:precompile Everything's getting compiled, with digests, but the digest applied to the actual file is not the same as the digest calculated and put in to application.css.
The resulting file is
public/assets/my-font-2f25682a1ea904a866ef9f44101a5a2e.css but in public/assets/application-bba2edaee36771f4bdb5c89b8ec90aaf.css the reference to it is:
#import url(/assets/my-font-ed843d3b174ca427edf963e473ad9e9b.css);
I realize I'm probably using asset_path() more than I should, and also importing files via url() instead of requiring them, but this has gotten me the closest to having things working.
I suspect one of the digests is being calculated on my-font.css before it goes through ERB, and the other after, but I don't understand why nor how to fix it.
Suggestions?
I would guess that you're cleaning your assets just by emptying public/assets. That's not enough, you'll also need to empty your tmp/cache/assets, or just run rake assets:clobber to do both.
I've resolve this kind of dependency by injecting the raw contents of the asset you want to bundle into a composite asset like application.scss.erb. You need to explicitly declare the dependency (so that changes in my-fonts.css cause a regenerations of application.css) and then inject the contents:
application.scss.erb:
// Should be at Top of File
//= depend_on_asset my-fonts.css
//... wherever in the file you want the contents injected:
<%= environment['my-fonts.css'] %>
How does this work? During asset pipeline compilation, the environment here has a hash of all pre-compiled asset contents. This allows you to inject the compiled contents from my-fonts.css into your composite asset. Because we're manually injecting the value, we need to explicitly create a dependency to track this relationship via depend_on_asset.
One thing to keep in mind is that multiple asset pre-processors (SCSS, ERB, etc) are processed from the "outside in", so the contents of the my-fonts.css asset will be compiled/injected during the ERB processing as CSS output. They will be included within the asset before SCSS processing. This shouldn't pose a problem, because if this is an SCSS asset, any SCSS references will be compiled before injection into the parent asset.
I really don't know how to debug this, maybe somebody has an idea.
I have many LESS files which I import in application.css.less. We use Bootstrap, and we want to use the variables that are defined in it within our own styles, so we can't require the LESS files in the manifest (as required files don't seem to make their variables public to other required files).
Everything works nicely, except a few of the LESS files - when edited - don't trigger a recompile of the CSS! It seems to be quite random which do and which don't, and it's only 6 of them which don't (compared to about 25 in sum). When I require one of them in the manifest, it successfully leads to a recompile - if I #import it, it doesn't.
Any idea on how to debug this? If I rename one of them (e.g. from time_records.less to time_records2.less), it successfully trigger recompile after changes... So it has do do something with the names of these 6 specific files:
calendars.less
contacts.less
folders.less
handout.css.less
print.css.less
time_records.css.less
Any help is greatly appreciated. Thank you.
The latest edition of less-rails implements Import Hooks that should solve your problem. I realize this is a late answer and it may have not existed at the time.
Import Hooks
Any #import to a .less file will automatically declare that file as a sprockets dependency to the file importing it. This means that you can edit imported framework files and see changes reflected in the parent during development. So this:
#import "frameworks/bootstrap/mixins";
#leftnav { .border-radius(5px); }
Will end up acting as if you had done this below:
/*
*= depend_on "frameworks/bootstrap/mixins.less"
*/
#import "frameworks/bootstrap/mixins";
#leftnav { .border-radius(5px); }
The depend_on Directive
depend_on path declares a dependency on the given path without including it in the bundle. This is useful when you need to expire an asset's cache in response to a change in another file.
I want to make use of the new json fixtures ability of jasmine-jquery, but can't get it to work. I have a Rails 3.2 app with evergreen for JS testing. As the readme of https://github.com/velesin/jasmine-jquery said, putting the jasmine-jquery.js inside the spec/javascripts/helpers dir would load jasmine-jquery automatically.
I put jasmine-jquery (and other libraries like jasmine-ajax) into
spec/javascripts/helpers dir (so they are automatically loaded) and
fixtures into spec/javascripts/fixtures dir.
But it doesn't. I tried to load the file inside the spec_helper.js, but the require only points to the public dir.
I also found this answer: https://groups.google.com/forum/?fromgroups=#!topic/ruby-evergreen/0Tma5RKqZB8
But the solutions aren't working in my case.
So, where can I tell evergreen / jasmine to load the jasmine-jquery.js file, so that I can use the loadFixtures methods...?
When I use the "Show me the page" step, with Capybara / Cucumber, can I control where the files get output? I can't seem to find the config for this, and my google fu is failing me.
Right now it appears that by default they go to the root of my rails folder and clutter up things there.
There is indeed a config option that allows you to specify where to output the files:
Capybara.save_and_open_page_path
I believe it was added in the latest version (0.3.9)
In your env.rb file you can do something like:
Capybara.save_and_open_page_path = '/Users/jsboulanger/my-rails-project/tmp'
In Capybara 2.10, Capybara::save_and_open_page= is deprecated. Instead, call Capybara::save_path=
Nice. Thanks for this.
To be really neat about it I added the config line to config/environments/test.rb, since you generally only use capybara in test, and that works fine.
Since there's a bunch of subfolders in tmp/ I used:
Capybara.save_and_open_page_path = 'tmp/capybara'
and created that folder.