You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Any changes we make are really hard to test the impact without having an example to test with.
I suggest we create a sub-folder called test containing its own streamlined package.json, gulpfile.js, and a sass folder containing some simple sass rules.
As for how to properly write the tests themselves, I'm not sure. Maybe we should copy the generated contents of the css folder to other folder names like css-gulp-sasscss-gulp-sass-no-sourcemaps etc and write javascript that recursively compares the generated output in css from running gulp sass and gulp sass --no-sourcemaps with these other folders.
Unfortunately though this wouldn't necessarily allow us to test things like purposeful linting failures.
It also wouldn't test to confirm BrowserSync commands work, so we'd need a separate test for that. Not sure how we'd write that either, to confirm that the watch command doesn't fail, and confirm that it starts browsersync (not sure how to detect the port).
@balintk any suggestions on how we might go about testing that the tasks and command line options?
The text was updated successfully, but these errors were encountered:
Any changes we make are really hard to test the impact without having an example to test with.
I suggest we create a sub-folder called
test
containing its own streamlinedpackage.json
,gulpfile.js
, and asass
folder containing some simple sass rules.As for how to properly write the tests themselves, I'm not sure. Maybe we should copy the generated contents of the
css
folder to other folder names likecss-gulp-sass
css-gulp-sass-no-sourcemaps
etc and write javascript that recursively compares the generated output incss
from runninggulp sass
andgulp sass --no-sourcemaps
with these other folders.Unfortunately though this wouldn't necessarily allow us to test things like purposeful linting failures.
It also wouldn't test to confirm BrowserSync commands work, so we'd need a separate test for that. Not sure how we'd write that either, to confirm that the
watch
command doesn't fail, and confirm that it starts browsersync (not sure how to detect the port).@balintk any suggestions on how we might go about testing that the tasks and command line options?
The text was updated successfully, but these errors were encountered: