Coding Horrors: You can gather all your technical debt into one easy framework!

The Something Awful Forums > Discussion > Serious Hardware/Software Crap > The Cavern of COBOL > Coding Horrors: You can gather all your technical debt into one easy framework!

«‹›1503 »

canis minor: May 4, 2011

I'm currently redoing code that was solely developed by tech lead that has left and am more and more surprised how culture of "let's add more poo poo on top of existing poo poo and never refactor" can result in code that doesn't really do anything.

I really want my next company to not be "hey, we've written this code, but we also took heavy drugs while doing this, so search product functionality is actually done in user class". I guess third time is the charm... although new place doesn't have any unit tests, so lol.

canis minor fucked around with this message at 12:48 on Jun 15, 2018

# ? Jun 15, 2018 12:42

Adbot: ADBOT LOVES YOU

# ? May 31, 2024 04:13

eth0.n: Jun 1, 2012

quote:

To avoid login errors, please do not use any of the following special characters in your password: % & + � � � Also do not use bullet points or the Euro currency symbol.

Bonus points: it's a scrolling marquee. That isn't on the create password page:

# ? Jun 15, 2018 14:38

nielsm: Jun 1, 2009

"You must use symbols in your password, but NO NOT THAT ONE!"

# ? Jun 15, 2018 14:58

Ola: Jul 19, 2004

Jeez, just tell me the exact password phrase you want me to use.

# ? Jun 15, 2018 15:17

Polio Vax Scene: Apr 5, 2009

Why are ' and � two very real things

# ? Jun 15, 2018 15:41

canis minor: May 4, 2011

Polio Vax Scene posted:

Why are ' and � two very real things

Not only two! https://en.wikipedia.org/wiki/Quotation_mark#Unicode_code_point_table

# ? Jun 15, 2018 15:52

Janitor Prime: Jan 22, 2004; PC LOAD LETTER

What da fuck does that mean; Fun Shoe

Polio Vax Scene posted:

Why are ' and � two very real things

Are you programming in MS Word? That poo poo almost never comes outside of MS products.

# ? Jun 15, 2018 17:07

Polio Vax Scene: Apr 5, 2009

Some fancy mcfancypants entered their first name as Le'douche into their data entry (Not actually le'douche but you get the drift) and the data entry does not correct this
e: It is a Microsoft product. Not word though.

# ? Jun 15, 2018 17:14

90s Rememberer: Nov 30, 2017; by R. Guyovich

eth0.n posted:

Bonus points: it's a scrolling marquee. That isn't on the create password page:

Kinda makes you wonder how seriously ADT takes security in the first place.

# ? Jun 15, 2018 17:30

dwazegek: Feb 11, 2005; WE CAN USE THIS

Mooey Cow posted:

For instance, every local variable you use in an iterator function will/can be made into a member variable in a hidden class so it can preserve state between yields. I believe async/await does something similar. If you're careless or don't know about this, causing enormous allocations from what looks like fairly innocuous stuff is very easy.

While it's true that variables can be captured in a generated type, you'd also have to be capturing hundreds of variables before the class ends up in multi-kb territory. Which means that you have hundreds of variables in method scope to capture in the first place, which seems to be the real problem.

Also some of the genetrated types are structs, so instead of a variable on the stack, you end up with a generated struct on the stack with the variable in it, so there's not that much difference.

Your point about templates is true though. There is some tooling out there to generate code for that sort of stuff, but that's kind of a poo poo solution compared to templates.

However, I do wonder if C# is the right tool for your job if you frequently run into that sort of issue. Math heavy stuff has never really been one of C#'s strengths.

# ? Jun 15, 2018 17:47

Pollyanna: Mar 5, 2005; Milk's on them.

What kind of data do you think a column named source_id would store? If you answered �an integer� or �a uuid�, you are wrong! It�s actually a big ol� minimally structured string of arbitrary key-value data. :shepface:

# ? Jun 15, 2018 18:13

rjmccall: Sep 7, 2007; no worries friend; Fun Shoe

Polio Vax Scene posted:

Some fancy mcfancypants entered their first name as Le'douche into their data entry (Not actually le'douche but you get the drift) and the data entry does not correct this
e: It is a Microsoft product. Not word though.

What would need to get �corrected� about this?

# ? Jun 15, 2018 18:16

hailthefish: Oct 24, 2010

Presumably it got stored as Le�douche and trying to do operations on Le'douche doesn't work, or entered as Le�douche but got stored as Le?douche or Le�douche or

code:

Le&rsquo;douche

or whatever.

Or :thejoke:

hailthefish fucked around with this message at 18:27 on Jun 15, 2018

# ? Jun 15, 2018 18:24

canis minor: May 4, 2011

Pollyanna posted:

What kind of data do you think a column named source_id would store? If you answered �an integer� or �a uuid�, you are wrong! It�s actually a big ol� minimally structured string of arbitrary key-value data.

This reminds me that some time ago I've discovered that basket_id actually contains a hash of the session current user has :v:

(no problem there's an actual baskets table that has nothing to do with this field). Still, something better from today: a join on LocationSearches table and Searches table on varchar2 address field.

# ? Jun 15, 2018 18:47

Pollyanna: Mar 5, 2005; Milk's on them.

Oh yeah, and as part of killing off our old Java monolith, I have to go through its codebase and extract hardcoded values and overridden methods that construct arbitrary maps n poo poo. :shepface:

I�m beginning to brainstorm easier ways to do that. I wonder if we can leverage the code that puts these blobs of data together, and have the monolith send the blob of data along with what it made that blob of data for to a service or something with a line of code after that method call? To kinda automate the poo poo I�d have to do otherwise. Does that even make sense?

# ? Jun 15, 2018 19:16

TooMuchAbstraction: Oct 14, 2012; I spent four years making
Waves of Steel
Hell yes I'm going to turn my avatar into an ad for it.; Fun Shoe

Pollyanna posted:

What kind of data do you think a column named source_id would store? If you answered �an integer� or �a uuid�, you are wrong! It�s actually a big ol� minimally structured string of arbitrary key-value data.

There was a brief period of darkness where we were converting some numeric IDs into strings for transmission to the frontend code, which would then dutifully cast them back to numbers. Fixing that felt good.

(This was more of a project management horror than a coding horror, though; give people more than six weeks to do major software development)

# ? Jun 15, 2018 19:25

dwazegek: Feb 11, 2005; WE CAN USE THIS

One of our applications puts items in a database table for later processing, so basically a queue. A separate scheduled task then calls a wcf service to retrieve the items, batched per 7 items (why 7? who the gently caress knows). After processing an item the task makes another call to the same wcf service to remove the item from the queue. Relatively straightforward stuff.

The stupid thing is that the id for the item is a guid. Which is stored as a char(38) in the db. And is exposed as a string by the wcf service. But the call to dequeue the item takes a guid, so the client has to parse the string it got from the server to a guid in order to pass it back to the same server. Which in turn obviously turns it back to a string when it goes to the db :downs:

# ? Jun 15, 2018 19:46

Hammerite: Mar 9, 2007; And you don't remember what I said here, either, but it was pompous and stupid.; Jade Ear Joe

dwazegek posted:

One of our applications puts items in a database table for later processing, so basically a queue. A separate scheduled task then calls a wcf service to retrieve the items, batched per 7 items (why 7? who the gently caress knows). After processing an item the task makes another call to the same wcf service to remove the item from the queue. Relatively straightforward stuff.

The stupid thing is that the id for the item is a guid. Which is stored as a char(38) in the db. And is exposed as a string by the wcf service. But the call to dequeue the item takes a guid, so the client has to parse the string it got from the server to a guid in order to pass it back to the same server. Which in turn obviously turns it back to a string when it goes to the db

Who cares? unless it's actually the bottleneck in your application, which doesn't sound likely

# ? Jun 15, 2018 20:34

fritz: Jul 26, 2003

Polio Vax Scene posted:

Some fancy mcfancypants entered their first name as Le'douche into their data entry (Not actually le'douche but you get the drift) and the data entry does not correct this
e: It is a Microsoft product. Not word though.

https://www.youtube.com/watch?v=YPOCIlvAsP8

# ? Jun 15, 2018 20:40

reversefungi: Nov 27, 2003; Master of the high hat!

hailthefish posted:

Presumably it got stored as Le�douche and trying to do operations on Le'douche doesn't work, or entered as Le�douche but got stored as Le?douche or Le�douche or
code:
Le&rsquo;douche
or whatever.

Or ?

This is probably Windows 1252 encoding. Had an issue related with this that drove me mad the other day trying to figure it out.

# ? Jun 15, 2018 20:54

dwazegek: Feb 11, 2005; WE CAN USE THIS

Hammerite posted:

Who cares? unless it's actually the bottleneck in your application, which doesn't sound likely

You don't think that it's an issue that a web service exposes an id as a string, but requires the client to post that same id back as a guid? Or that the id could have just been a guid throughout the application, which would have resulted in better performance, and, more importantly, less potential for bugs? Sure, it's not the worst coding horror in the world, but still...

# ? Jun 15, 2018 21:27

Hammerite: Mar 9, 2007; And you don't remember what I said here, either, but it was pompous and stupid.; Jade Ear Joe

dwazegek posted:

You don't think that it's an issue that a web service exposes an id as a string, but requires the client to post that same id back as a guid? Or that the id could have just been a guid throughout the application, which would have resulted in better performance, and, more importantly, less potential for bugs? Sure, it's not the worst coding horror in the world, but still...

I missed that the id is exposed as different types on the same interface, yeah that's a bit ugly. But still, if that's the biggest wart on the application then you're doing ok.

# ? Jun 15, 2018 21:36

CPColin: Sep 9, 2003; Big ol' smile.

code:

entry.inTime = xml.InPunch?.Punch?.'@Time'?.getAt(0)

Huh, that getAt() call looks weird, but maybe Groovy's XML support does weird things and this is how you get the Time attribute of the first Punch element, in case there's more than one? Whatever, time to cover this super-old code with a test before I change anything:

code:

assertEquals("12:30", entry.inTime)

 junit.framework.ComparisonFailure: expected:<12:30> but was:<1>

God loving drat it.

Edit: Oh gently caress, Groovy's XML support returns attribute values as lists:

http://grails.asia/groovy-xmlparser-examples-for-parsing-xml

So the previous developer did need those getAt() calls after all and my stupid test has to wrap that value in a list for no reason! Never mind that well-formed XML disallows multiple attributes with the same name in the same element!

Now I need to use actual XML as my mock data, because this is getting stupid.

CPColin fucked around with this message at 23:00 on Jun 15, 2018

# ? Jun 15, 2018 22:21

necrotic: Aug 2, 2005; I owe my brother big time for this!

You should've been using real xml for the mock data from the start.

# ? Jun 16, 2018 00:30

CPColin: Sep 9, 2003; Big ol' smile.

Yep. Groovy's dynamic typing is letting me get away with getting sloppy.

# ? Jun 16, 2018 00:39

Absurd Alhazred: Mar 27, 2010; by Athanatos

Janitor Prime posted:

Are you programming in MS Word? That poo poo almost never comes outside of MS products.

TeX wants you to distinguish starting from ending quotes, and from double quotes. So you don't write "this", you write ``this''.

# ? Jun 16, 2018 01:00

Eggnogium: Jun 1, 2010; Never give an inch! Hnnnghhhhhh!

dwazegek posted:

You don't think that it's an issue that a web service exposes an id as a string, but requires the client to post that same id back as a guid? Or that the id could have just been a guid throughout the application, which would have resulted in better performance, and, more importantly, less potential for bugs? Sure, it's not the worst coding horror in the world, but still...

Assuming a message protocol is SOAP or some other text-based protocol then it�s being converted to and from a string to cross the wire anyway, so probably not any real performance loss. Mildly annoying that you have to convert in your own code instead of letting WCF do it for you.

# ? Jun 16, 2018 16:29

LOOK I AM A TURTLE: May 22, 2003; "I'm actually a tortoise."; Grimey Drawer

Eggnogium posted:

Assuming a message protocol is SOAP or some other text-based protocol then it�s being converted to and from a string to cross the wire anyway, so probably not any real performance loss. Mildly annoying that you have to convert in your own code instead of letting WCF do it for you.

You can easily do binary encoding over TCP with WCF, so this may not be a good assumption, but even in that case the loss of type safety is probably a bigger issue than performance.

# ? Jun 16, 2018 17:56

ToxicFrog: Apr 26, 2008

Having a weird issue with some Discord client software, I eventually tracked the problem down to this initialization code (pared down a bit):

code:

for (const channel in server.channels) {
  this.channels[channel.name] = channel;
}

Discord channel names are not guaranteed to be unique. channel.id is the unique key. If the server has multiple channels with identical names, only the last one in server.channels makes it into this.channels. It's unpredictable from connection to connection which order they're iterated in, so this doesn't even give you consistently wrong results.

# ? Jun 16, 2018 18:17

brap: Aug 23, 2004; Grimey Drawer

Hell, for-in doesn't even give you consistent iteration order, does it?

# ? Jun 16, 2018 23:25

Bruegels Fuckbooks: Sep 14, 2004; Now, listen - I know the two of you are very different from each other in a lot of ways, but you have to understand that as far as Grandpa's concerned, you're both pieces of shit! Yeah. I can prove it mathematically.

brap posted:

Hell, for-in doesn't even give you consistent iteration order, does it?

in js, for in returns properties in the order they were defined - which is consistent, but arbitrary.

# ? Jun 17, 2018 00:37

putin is a cunt: Apr 5, 2007; BOY DO I SURE ENJOY TRASH. THERE'S NOTHING MORE I LOVE THAN TO SIT DOWN IN FRONT OF THE BIG SCREEN AND EAT A BIIIIG STEAMY BOWL OF SHIT. WARNER BROS CAN COME OVER TO MY HOUSE AND ASSFUCK MY MOM WHILE I WATCH AND I WOULD CERTIFY IT FRESH, NO QUESTION

dwazegek posted:

You don't think that it's an issue that a web service exposes an id as a string, but requires the client to post that same id back as a guid? Or that the id could have just been a guid throughout the application, which would have resulted in better performance, and, more importantly, less potential for bugs? Sure, it's not the worst coding horror in the world, but still...

It's untidy and I certainly wouldn't design something that way, but I don't think it's likely to increase bugs as you said. If it was the other way around (input expected a string that it subsequently assumes to be a guid) then I could see potential bugs arising from that for sure.

# ? Jun 18, 2018 05:54

TheresaJayne: Jul 1, 2011

My latest one, setting up my old iphone as a company phone,
they force you to have a 6 digit pin unlock code,
The rules are as follows

You cannot have repeating digits (11 or even 123451 )
You cannot have 2 adjacent numbers (129874 is not allowed as 12 and 98 and 87 are adjacent)

# ? Jun 18, 2018 06:47

Loezi: Dec 18, 2012; Never buy the cheap stuff

TheresaJayne posted:

My latest one, setting up my old iphone as a company phone,
they force you to have a 6 digit pin unlock code,
The rules are as follows

You cannot have repeating digits (11 or even 123451 )
You cannot have 2 adjacent numbers (129874 is not allowed as 12 and 98 and 87 are adjacent)

If my quick calculations are correct then that leaves about 12% of the 6-digit pins as valid. Nice going.

# ? Jun 18, 2018 08:09

LOOK I AM A TURTLE: May 22, 2003; "I'm actually a tortoise."; Grimey Drawer

Loezi posted:

If my quick calculations are correct then that leaves about 12% of the 6-digit pins as valid. Nice going.

I'm getting 4.6%.

Python code:

def has_repeats(n):
  seen = set()
  for c in n:
    if c in seen:
      return True
    seen.add(c)
  return False

def are_consecutive(c1, c2):
  n1, n2 = int(c1), int(c2)
  return n1 - n2 == 1 or n2 - n1 == 1

def has_consecutives(n):
  for c1, c2 in zip(n, n[1:]):
    if are_consecutive(c1, c2):
      return True
  return False

def is_valid(n):
  return not has_repeats(n) and not has_consecutives(n)

numbers = list(range(0, 1000000))
valid = list(filter(is_valid, map(lambda n: '{0:06d}'.format(n), numbers)))
print(len(valid) / len(numbers))

All valid codes: https://pastebin.com/ivx1692s

Some assumptions:
1. Codes may start with 0.
2. 0 and 9 are not considered adjacent.

# ? Jun 18, 2018 09:40

Loezi: Dec 18, 2012; Never buy the cheap stuff

LOOK I AM A TURTLE posted:

I'm getting 4.6%.

Yeah, looking at it again, this seems correct.

# ? Jun 18, 2018 11:16

Vanadium: Jan 8, 2005

I mean, that still leaves more choices than industry standard four digit pin codes.

# ? Jun 18, 2018 11:34

TheresaJayne: Jul 1, 2011

Vanadium posted:

I mean, that still leaves more choices than industry standard four digit pin codes.

But a lot of people use patterns on the Keypad like a T L P O etc. so most patterns are not allowed anything going across the pad is not allowed and in some cases up down as well 0147 is not valid

# ? Jun 18, 2018 11:41

Ola: Jul 19, 2004

LOOK I AM A TURTLE posted:

I'm getting 4.6%.

This means 6 and 7 are considered adjacent, but they're not adjacent on the numpad. And perhaps vertical adjacency is invalid as well.

# ? Jun 18, 2018 11:53

Adbot: ADBOT LOVES YOU

# ? May 31, 2024 04:13

LOOK I AM A TURTLE: May 22, 2003; "I'm actually a tortoise."; Grimey Drawer

Ola posted:

This means 6 and 7 are considered adjacent, but they're not adjacent on the numpad. And perhaps vertical adjacency is invalid as well.

Nice, this lets us get rid of even more entropy. I'll assume we have this keypad layout:

pre:

With horizontal and vertical adjacency both disallowed we can now discard sequences with 12, 14, 21, 23, 25, 32, 36, etc. On average each digit is now adjacent to 2.5 other digits. This brings the number of permitted codes down to 2.66%.

But why stop here? We should obviously block direct diagonal adjacency as well. If we include diagonal adjacency we can almost completely eliminate the digit 5 from the list of valid codes, as it would then only be able to occur at the beginning or end of the code, and only next to a 0. If e.g. 15 and 07 are also considered adjacent we can bring the number of possible codes down to 0.52%.

# ? Jun 18, 2018 12:35

The Something Awful Forums > Discussion > Serious Hardware/Software Crap > The Cavern of COBOL > Coding Horrors: You can gather all your technical debt into one easy framework!

«‹›1503 »