The author mentions PEP 8. In the Python community, there is a style checker called pep8, which implements the PEP 8 guidelines. Style only. There's also a tool called pyflakes, which does static code analysis. Obviously, it can't catch much, but it does alright. Then, there are lints: pylint and pychecker. These last two are horribly dogmatic and cannot stand Python which isn't written in their style. We usually don't recommend pylint and pychecker because of their overbearing nature and insistence on Java-in-Python code.
I recently built a couple websites, and we used Sass for the CSS, so we ended up with very nicely formatted SCSS maintained by hand, and minified CSS for deployment. I tried CSS Lint and was horrified at what it did to our CSS, so it's no longer part of our cycle.
If people wanna use it, fine, but it's not helpful to me.
What exactly do you mean by Java-in-Python code? Do you have examples? I find that code written using Pythonic idioms gets a high Pylint score, without having the slightest whiff of Java about it. That's not to say that its defaults aren't excessively fussy in some cases, but it's easy to tweak so that it doesn't whine about too many public methods or using magic &c.
Bingo. It's really annoying that I can't use tuple/dict (un)packing without it complaining, let alone that if I have a class which is too big, too small, or does too much/too little, it throws a fit. Worse, it doesn't understand zope.interface and gets quite irate when confronted with interface-heavy code.
I'm usually working in Twisted or Django, and it's so very unfun to try to get pychecker/pylint to be okay with those libraries. I'd rather have pyflakes.
This is extremely week argument for CSS Lint.
CSS is not the starting point (unless you are talking about CSS sprites) for the performance optimization, and, frankly, it may never be a point.
On a normal page it would be hard even measure difference between performance of class vs. descendent selector and littering your markup with classes may offset any benefits by simply getting bigger document which takes longer to parse.
And tell me, how making id selector (the fastest) persona non grata helps performance?
"Rebutting" a single point (which was not even a point, just a note, based on how do you "fee" about the author, does not rebut anything.
I am a programmer and I tell you: CSS Lint won't make your pages faster. It may make the slower, but that does not matter, cause nobody will notice it either way.
6 comments
[ 2.9 ms ] story [ 33.0 ms ] threadI recently built a couple websites, and we used Sass for the CSS, so we ended up with very nicely formatted SCSS maintained by hand, and minified CSS for deployment. I tried CSS Lint and was horrified at what it did to our CSS, so it's no longer part of our cycle.
If people wanna use it, fine, but it's not helpful to me.
I'm usually working in Twisted or Django, and it's so very unfun to try to get pychecker/pylint to be okay with those libraries. I'd rather have pyflakes.
And tell me, how making id selector (the fastest) persona non grata helps performance?
"Rebutting" a single point (which was not even a point, just a note, based on how do you "fee" about the author, does not rebut anything.
I am a programmer and I tell you: CSS Lint won't make your pages faster. It may make the slower, but that does not matter, cause nobody will notice it either way.