Style rules for the seqbind transform

[gnuvince]

The seqbind parser transform allows Erlang code to reuse the "same" variable name multiple times:

X@ = 10,
X@ = X@ + 1.

Though useful sometimes, its usage is typically a code smell. I have created two new style rules (gnuvince/elvis_core#1):

• The no_seqbind rule warns when a module declares its intention to use seqbind (-compile({parse_transform, seqbind}).);
• The no_useless_seqbind rule warns when a module has activates the seqbind transform, yet never makes any use of it (i.e., no Var@ names anywhere in the module).

[elbrujohalcon]

@gnuvince just curious: where do you get that parse_transform from?
I haven't seen anything like that in my ~10 years as an Erlanger. 😱

[gnuvince]

@elbrujohalcon It's an option for the compiler: http://erlang.org/doc/man/compile.html (scroll down a few screenfuls).

[elbrujohalcon]

I know what parse_transform is ;). I was asking where did you get this one (seqbind) in particular.

[gnuvince]

Oops, misread, sorry! Here's their Github repo: https://github.com/spawngrid/seqbind

[elbrujohalcon]

Thanks. Please send a PR over with your rules, then :)

[gnuvince]

PR submitted: inaka/elvis_core#67

The gadget/compiler check failed, but I'm not sure what the problem is.

[DBoroujerdi]

@gnuvince I commented on issue #418 about failed tests but I hadn't seen this issue. It's related to this seqbind merge and could be the cause of the gadget/compiler check you mention.