Skip to main content

Settling the bracket discussion once and for all

There are quite a few ways you can format a multi-line function call/list/dictionary in a programming language. With Python (at least with PEP8) it's just two ways: hanging indentation and vertical indentation. The latter is quite straight-forward and I mention it in the notes below. The former one is what this post is about. Well, about a minor but important question of where to place the closing bracket when using it.

Yes, style is much more than formatting the commas and brackets so let's settle this and move on to the more important questions.

When using hanging indentation, the only true way is this one(*):
some_tuple = (  # Nothing here, only opening bracket.
    item1,
    item2,
    item3, item4,  # Several items on one line is acceptable though not recommended.
    item5,  # Last comma is important. And never, never place the closing bracket here.
)  # Again, nothing here but the bracket and it's on the same level as the opening line.
Empty line after the opening comma is mandated by PEP8 even though there are a few cases where doing otherwise would've been a possible alternative, like:
class DoNotDoThis(object):

    def just_do_not(self,
        long_argument1,
        long_argument2,
        long_argument3=long_default_value,
    )
and there is a reason for that — anything after the bracket makes the opening line less visible. That's part of the reason for the two spaces before inline comment rule.

The last comma(†) and separate line for the closing bracket is important because you don't really want to see another commit in git blame just because another item was added or deleted and you had to add a comma or move the bracket. You you're using Vim, it also makes adding/deleting easier because you can operate with the whole lines (it helps with just any editor but with Vim it's more so).

I'll repeat: closing bracket should be on separate line and on the same indentation level with the beginning of the opening line (if some_tuple is indented by four spaces, closing bracket should be indented by four spaces).

It might sound like a minor issue but 1) it's the only meaningful way of doing brackets once you think about it, and 2) they should be formatted in uniform way throughout the code so why not pick this one from the three or four PEP8-compatible alternatives(‡). It's one of the things that are easily controllable with pep8 linter (or flake8 which I actually prefer, the same thing but better) — you might wanna configure some auto-check in your editor/IDE, most of them allow that easily (plus, autoindent helps tremendously).

(*) The only case when you place the closing bracket on the line with the last item is this (so-called vertical indentation):
call_some_function(argument1, argument2,
                   argument3, argument4)
Although I'm not fan of this style, it's been around since Kernighan and Ritchie... But still, don't do that, especially with long arguments. Also, while it's still acceptable to put a closing comma on a separate line when using vertical indentation, it looks kinda ugly for me.

(†) There is one case where there should be no last comma. If it's a function call/def and you're using *args or **kwargs. It's almost not worth mentioning because you don't add stuff after **kwargs and it would just result in syntax error if you placed a comma there (not sure why it was implemented that way, but that's the way things are). But still a separate line for the bracket is preferred.

(‡) The text of PEP8 is little ambiguous on this point and some variations acceptable by the letter of it are not allowed by the recent versions of the linter, but it's essentially three options for hanging indentation: a) closing bracket on the separate line on the same level of indentation as the items (doesn't work well with if/while/def blocks), b) closing bracket on the same line as the last item (makes editing harder and messes with git blame when you add/remove items), c) what I have presented above. (There is also d) bracket on any other other level of indentation, like four spaces more than for the items, but that just makes zero sense.)

Popular posts from this blog

「この番組は、ご覧のスポンサーの提供でお送りします」Inference Labs

It's official, I have corporate sponsors now (which will hopefully be closer to strategic partnership soon, at least I'd like that). Thanks for the steaks and wine and equipment and all other resources for my distributed hacker team (I'm hoping for a cool demo for ETHOnline hackathon), and local Web3 meatup group I facilitate in Tbilisi (steaks and gas money for everyone participating in the ecosystem), and my Web3 community support army in training. A large bag of kudos for our general sponsor: Inference Labs! If anybody needs help onboarding into the decentralised world, reach me privately and we'll help you navigating the options, we have the expenses covered and good people for this (better points of contact are being established, things are being worked on, stay tuned). I should probably tell a few words about what we actually do at Inference Labs but things change so often I hope we're still bringing decentralized AI to the Web3 world when you read this (maybe...

New horizons

I originally started this blog with ideas of reviewing devices and services and hoping that eventually if it gets popular enough somebody starts sending me stuff to review. A lot has changed since, I stopped obsessing that much about new gadgets and got into vintage electronics, many of things that were new and interesting a few years ago are a commodity now. I thought about reviewing the phone I finally got to refreshing last year (S23 Ultra is ok upgrade for Note 8, I'm glad that new ultras will finally have flat screen again, I might upgrade next year or so just for that) but I don't really feel like it or think it would mean much for the readers. Most of my vintage electronics is at home where I haven't been for a couple years and it's not something I can currently do something about, I touched a soldering iron like once or twice this year. I might post something work-related once I get the hang of what I'm actually doing there and have some rough ideas wen dece...

Using virtualenv for more than Python projects

Sorry, it's not a complete instruction, just a thought. It occurred to me (some time ago) that Python's virtualenv is, essentially, a simplified version of system "prefix", it has bin, lib, include, and can have more stuff when needed. If you're willing to experiment (you'll probably have to set a few additional environment variables and/or build flags but that's no big deal), you can install various other tools there up until you have a complete system with its own compiler and complete set of libraries although it's much simpler to keep using system compiler and libraries only complimenting them when needed. Granted, prefixes are nothing new, people were using /opt (and their home directory) this way since the beginning of time. But with little help of virtualenv-wrapper or pyenv you can easily switch between them and isolate environments better. Binaries and stuff installed in virtualenv would override system defaults but only when venv is activat...