Never mind automated linters, how many times have you seen this exchange?
"You should always use strict and warnings."
"I used to have those but they caused a lot of errors and my code wouldn't run so I took them out."
(This has a relative: "Didn't the CPAN shell make test?" "Yeah, but it gave a lot of errors so I had to do a force install.")
| [reply] |
Andy Lester graciously wrote:
Never mind automated linters, how many times have you seen this exchange?
“You should always use strict and warnings.”
“I used to have those but they caused a lot of errors and my code wouldn't run so I took them out.”
How many times have I seen that particular exchange, Andy? As many times as there are grains of sand in the sea! Especially if you count including -Wall in your Makefile’s CCFLAGS.
But I can top that. I’ve actually seen this, and not just once, either:
Manager: “We can’t have any more coredumps from our C programs. It makes us look bad and interferes with production.”
Programmer: signal(SIGSEGV, SIG_IGN);
I dearly wish I were kidding; I’m not.
I see stuff like this every day, stuff that doesn’t so much move you to righteous ranting as it does to being stricken dumb with sheer disbelief. I bet I now have keyboard-shaped wedges permanently imprinted on my forehead.
It’s code that leads first to apoplexy and thence to syncope. The thing is, after they bring around the code-smelling salts and still dazed you awaken from your nightmare, it’s all still there: holy terrors of truly Biblical proportion.
And by Biblical, I’m specifically thinking of the venerable Books of Job and of Lamentations and of Ecclesiastes: no Good News to be found. :)
We’re well into sackcloth-and-ashes territory here.
“Weeping may endure for an entire dark release, but joy cometh in the fresh light of refactoring.” ― Psalms 30:5, the Programmers’ edition
| [reply] |
