fritz posted:

I found a line in some academic C code just now that has 5 "="s, two arithmetic operators, and a [].

fritz posted:

I found a line in some academic C code just now that has 5 "="s, two arithmetic operators, and a [].

eithedog posted:

a += b[0] == c + d == f

What do I win?

eithedog posted:

a += b[0] == c + d == f

What do I win?

Edison was a dick posted:

i = i++ is undefined behaviour in C, so Java could have inherited that feature.

EAT THE EGGS RICOLA posted:

Someone just handed me a ~7000 line javascript thing to look at and make major changes to by tomorrow at the latest. Every variable and method has a two or three character random name. For example something I might call getElementId is called sn.

I'm legitimately thinking of just pretending this conversation never happened and/or just killing myself so I don't have to deal with this.

dealmaster posted:

Why do we get added to reviews like this at the very last second when nothing can realistically be done???

Volmarias posted:

Jabor posted:

So that the person who's actually responsible for that situation has someone to pass the buck to when it all comes crashing down.

If you're not able to actually veto the change, it's not a real code review, and you should probably make sure it's not your responsibility by stating that you have reservations and <person who punted the review to you> can sign off on it if they think it's good.

eithedog posted:

a += b[0] == c + d == f

What do I win?

Snapchat A Titty posted:

There's no benefit to writing it like that. Use intermediate variables with clear names and let the compiler figure out how to optimize stuff.

Voted Worst Mom posted:

How about,

a ^= (b[0] == c) ^ (d == f);

which doesn't seem unreasonable at all?

Subjunctive posted:

But also doesn't do the same thing?

fritz posted:

You asked for it.

code:

umax=(vmin=input[kminus=k=k0])+(umin=lambda)-vmax;

fritz posted:

You asked for it.

code:

umax=(vmin=input[kminus=k=k0])+(umin=lambda)-vmax;

JawnV6 posted:

And everyone's cheating the "5 =" by assuming a single assignment and two comparisons.

a = (b=4)+(c=6) +(d=e=0);

Flagrama posted:

There wasn't a code sample posted. They are both guessing what it could have been based off the description.

Coylter posted:

Well, let's just say this made my day. This gentlemens, is part of the server code of something important.

Coylter posted:

code:

If pyramid poo poo code

Well, let's just say this made my day. This gentlemens, is part of the server code of something important.

dealmaster posted:

Has the writer of this never heard of a regular expression? Jesus Christ.

xtal posted:

In fairness, regular expressions are pretty horrific themselves.

xtal posted:

In fairness, regular expressions are pretty horrific themselves.

Subjunctive posted:

Yeah, "now you have two problems."

That code's not horrible to read once you squint past the boilerplate. If it were chained with && it would be pretty decent. Maybe better with the equivalent of a string stream or sscanf?

^^^

Regular expressions are a classic example of write-only code, especially without Perl's /x readability help. This case is simple enough that it could be OK, but it's a dangerous path. I've worked on regex engines (JS's replace-with-function-call is my fault), and I have to think pretty carefully to make sure I'm thinking about greed and backtracking correctly with all but the simplest stuff.

(Most regex engines are not actually regular expressions, and they all vary a bit in syntax or semantics, which also hurts understanding after the fact.)

Shinku ABOOKEN posted:

I agree that composability would be a nice feature to have.

Suspicious Dish posted:

code:

digit_group(n) = \d{n};

delimiter = '-' | ' ' | '';

area_code = (
    # (123)
    ('(' + digit_group(3) + ')') |
    # 123
    digit_group(3)
);

global_extension = (
    '+' + number{0:4}
);

phone_number = (
    # +1 (123)-456-7890
    global_extension? + area_code + delimiter + digit_group(3) + delimiter + digit_group(4)
);

Yes, this is more like EBNF notation, I know, but it can still be represented with the same exact state machine as a regex. It is, for all intents and purposes, a regular expression.

Subjunctive posted:

Yeah, "now you have two problems."

Suspicious Dish posted:

I'm saying that our current regular expression representations could go a long way towards encouraging more readability and composability, and right now they fail in those aspects, because every character matters.

Googling for "phone number regex", the first result I found is: ^(??:\+?1\s*(?:[.-]\s*)?)?(?:\(\s*([2-9]1[02-9]|[2-9][02-8]1|[2-9][02-8][02-9])\s*\)|([2-9]1[02-9]|[2-9][02-8]1|[2-9][02-8][02-9]))\s*(?:[.-]\s*)?)?([2-9]1[02-9]|[2-9][02-9]1|[2-9][02-9]{2})\s*(?:[.-]\s*)?([0-9]{4})(?:\s*(?:#|x\.?|ext\.?|extension)\s*(\d+))?$

Given enough time, you can break apart what it's doing. It's almost like compiler output, and that it really has to be reverse-engineered. I won't clean up that, since this is a dumb forum post and I have work tomorrow in the morning. But imagine an alternate representation:

code:

digit_group(n) = \d{n};

delimiter = '-' | ' ' | '';

area_code = (
    # (123)
    ('(' + digit_group(3) + ')') |
    # 123
    digit_group(3)
);

global_extension = (
    '+' + number{0:4}
);

phone_number = (
    # +1 (123)-456-7890
    global_extension? + area_code + delimiter + digit_group(3) + delimiter + digit_group(4)
);

Yes, this is more like EBNF notation, I know, but it can still be represented with the same exact state machine as a regex. It is, for all intents and purposes, a regular expression.

Suspicious Dish posted:

The second problem is that traditional regular expression languages don't have any inherent composability. There's no way to say delimiter = '-' | ' ' | ''; and then use (\d{3})? $delimiter (\d{3}) $delimiter (\d{4}). Because of this, you also see a lot of clever trickery in common regex syntax to group some parts together, and a lot of "code golf" that makes it dense and unreadable. Trying to understand a regular expression is like reverse engineering assembly code, almost.

Suspicious Dish posted:

Googling for "phone number regex", the first result I found is: ^(??:\+?1\s*(?:[.-]\s*)?)?(?:\(\s*([2-9]1[02-9]|[2-9][02-8]1|[2-9][02-8][02-9])\s*\)|([2-9]1[02-9]|[2-9][02-8]1|[2-9][02-8][02-9]))\s*(?:[.-]\s*)?)?([2-9]1[02-9]|[2-9][02-9]1|[2-9][02-9]{2})\s*(?:[.-]\s*)?([0-9]{4})(?:\s*(?:#|x\.?|ext\.?|extension)\s*(\d+))?$