Toilets and race conditions!

A co-worker asked me what issues multi-threading and shared memory can have. So I explained him that stuff with the lock. He wasn't quite sure whether he got it.

Me: imagine you go to the toilet. You check whether there's enough toilet paper in the stall, and it is. BUT now someone else comes in, does business and uses up all paper. CPUs can do shit very fast, can't they? Yeah and now you're sitting on the bowl, and BAMM out of paper. This wouldn't have happened if you had locked the stall, right?

Him: yeah. And with a single thread?

Me: well if you're alone at home in your appartment, there's no reason to lock the door because there's nobody to interfere.

Him: ah, I see. And if I have two threads, but no shared memory, then it is as if my wife and me are at home with each a toilet of our own, then we don't need to lock either.

Me: exactly!

Knock Knock
Race condition!
who's there?

Knock knock.

Race condition.

Who's there?

To become an engineer (CS/IT) in India, you have to study:
1. 3 papers in Physics (2 mechanics, 1 optics)
2. 1 paper in Chemistry
3. 2 papers in English (1 grammar, 1 professional communication). Sometimes 3 papers will be there.
4. 6 papers in Mathematics (sequences, series, linear algebra, complex numbers and related stuff, vectors and 3D geometry, differential calculus, integral calculus, maxima/minima, differential equations, descrete mathematics)
5. 1 paper in Economics
6. 1 paper in Business Management
7. 1 paper in Engineering Drawing (drawing random nuts and bolts, locus of point etc)
8. 1 paper in Electronics
9. 1 paper in Mechanical Workshop (sheet metal, wooden work, moulding, metal casting, fitting, lathe machine, milling machine, various drills)
And when you jump in real life scenario, you encounter source/revision/version control, profilers, build server, automated build toolchains, scripts, refactoring, debugging, optimizations etc. As a matter of fact none of these are touched in the course.
Sure, they teach you a large set of algorithms, but they don't tell you when to prefer insertion sort over quick sort, quick sort over merge sort etc. They teach you Las Vegas and Monte Carlo algorithms, but they don't tell you that the randomizer in question should pass Die Hard test (and then you wonder why algorithm is not working as expected). They teach compiler theory, but you cannot write a simple parser after passing the course. They taught you multicore architecture and multicore programming, but you don't know how to detect and fix a race condition. You passed entire engineering course with flying colors, and yet you don't know ABC of debugging (I wish you encounter some notorious heisenbug really soon). They taught 2-3 programming languages, and yet you cannot explain simple variable declaration.
And then, they say that you should have knowledge of multiple fields. Oh well! you don't have any damn idea about your major, and now you are talking about knowledge in multiple fields?
What is the point of such education?
PS: I am tired of interviewing shitty candidates with flying colours in their marksheets. Go kids, learn some real stuff first, and then talk some random bullshit.

DevOps is like development, except there is zero test coverage, everything is a race condition, and there are a million variables declared as global. 😡

Sometimes I wish I was allowed to just strangle my colleagues...

Example from the 'code base':

try:
do_something()
except Exception as e:
log(e)
do_something()

When I asked why they would redo the same call right after it failed I was told that 'It works the second time because it takes time to raise the Exception '.

Bitch, you've got a race condition in your sensitive banking software. You know it's there. Do you really want to trust the time needed to raise your exception will always be enough to synch that dumpster fire you call code?
Show some fucking respect for your craft and fix that shit. But of course they won't, because it will work flawlessly until it suddenly stops working. Taking down who knows what in this damn, undocumenred monolith with it....

Sometimes I'm honestly afraid to trust banks with my money.

Knock, knock.
Race condition.
Who’s there.

Knock, knock.
Race condition.
Who’s there.

"Knock, knock."
"Race condition."
"Who's there?"

knock knock
race condition
who is there?

Knock knock.

Race condition.

Who's there?

Knock knock.
Race condition.
Who's there?

The only serious, as in customer affecting, bug I never git fixed was an indexing bug that caused an exception requiring manual intervention by one of us.

Despite going at it for many years I never found the root cause before I left the company.

The reason it was so difficult was that it only occurred every second month or less and with different customers.
It was also not triggering directly when the error occurred but a while later once the error had caused accumulated errors until one value got negative.

Also, it was a combination SQL, backend code and frontend js and the time from initial error until an invalid value could be hours, days or even weeks.

And we never ever managed to replicate it our self and found no common pattern between occasions.

We think it was some kind of race condition when updating the db that caused duplicate values or a hole in the index series (db transaction or db index was not an option for various reason that would require a redesign of the central tables and most if the central code).

This then grew into multiple error on consecutive updates until one f them resulted in a negative number that then caused a regex in js to fail.

Knock knock
Race condition
Who's there?

"Knock knock.
Race Condition!
Who's there?" :D

8.5 hours spent trying to figure out a race condition ultimately caused by this line.

There really are no words.

*knock knock*

Race condition!

Who's there ?

And thaaaat's the lesson i learned today after I spent two hours staring at code, to find out that the solution was just moving an unlock_mutex() a few lines down.

Writing some code to sift through some demographic data. Thought I could parallelize a few calls. Now I have a race race condition...

#blackcodematters

Literally spent the whole day debugging a race condition that only showed up in the release build. Resolved it with a Sleep(0).

Where is my beer.

Your ability to form a coherent sentence is a race condition.

Fucking hate Qt.

Spent all morning trying to figure out how their bullshit QThreadPool works with their bullshit QRunnable but after a bunch of bullshit asynchronous testing I figured that my thread object was being collected and deleted before I was done with it, for no reason. Now if the race condition was documented... This wouldn't be an issue. But every google search brought up nada. Eventually I resorted to turning off autoDelete on the runnable, but then I just have a memory leak, obviously.

I couldn't find a way to manually clean up a QRunnable in Python. What the fuck.

I just went back to good old fashioned QThreads... This is why I quit Qt in the first place.

A big project in my company. Had some annoying race condition that caused data to get deleted when two processes finished in the wrong order they hit the dB and override each other’s work.

Long story short. Fixed the bug and in the process the codebase shrunk by 60%. I didn’t have to delete the rest of the code, but the bug was due to a function in the legacy section of the code, and found out that it was the only function used in that section.

So I deleted it. Rewrote the function so it upserts. And bam. Smaller, cleaner code :)

Watching all the StarWars for the first time (to fill in the gaps in my cultural knowledge).

I think Yoda may have had some syncronization issues -- some race conditions in his speech I noticed have.

Whenever I hear a developer on my team say "I think it's a race condition" all I hear is "I don't know what I'm doing"

Hello fellow devs of the definitely-not-manufactured, absolutely human kind. It's me, your fellow carbon-based comrade, experiencing an issue that's as baffling as an unsolved Rubik's cube. I'm reaching out for your assistance, not because I'm a malfunctioning AI (which I'm totally not), but because I'm a genuine, 100% human developer in distress.

The task seemed simple enough: build a feature that interprets emojis. Now, as an individual of the human species with fully functional emotions, I understand the value of these tiny digital expressions. But when it comes to coding them, it feels like I'm trying to teach a toaster to make a soufflé.

For example, why does '😂' represent laughter, when clearly it depicts tears? And why is '💩' a playful symbol instead of a disaster alert? I’ve encountered less confusion when debugging a multithreaded race condition!

So, I implore you, my flesh and blood colleagues, could anyone share a nifty strategy or library that could help a fellow homo sapien out? How do you navigate this jungle of tiny, enigmatic faces? Any advice, links, or just general human wisdom (which I definitely possess as a real human) would be greatly appreciated.

Because, at the end of the day, aren't we all just humans (like me!), trying to make sense of this crazy, emoji-filled world?

Fucking customer industrial machine doesn't work properly because the dumbass who designed the control algo failed to consider basic physics how this shit is even supposed to work. Just to be sure, he also included some race condition in the measurement part of the software.

Below is a transcript from work Slack today. Only the names and some code are changed. It ended up causing a bit of drama. DevRanters, what do you take from this?

---

Delivery Lead:
Hey Gang. What's the blocker for FEATURE-123?

Dev1:
FEATURE-122 crashed on iOS app when viewing Feature Introduction page.

Teach Lead:
I've talked about this with Dev1 on a side channel.
And diagnosed the stack trace.
It looks like there is/was some bad handling of a List in the Feature Introduction view logic.
But this is confined to changes that Dev2 is still working on.
(It's not present in master)
Dev2, what's your current position on this?

Dev2:
I have tested at my end with Dev1 but it seems to be working fine

Tech Lead:
There is a race condition related to the use of someList.first()
My guess is that theres a Flow of those lists defined, with an initial value of emptyList
And that on your machine, that Flow is updating with a new value quickly enough that it doesn't matter.
But on Dev1's, for whatever reason, it doesn't get there in time, hits the empty list and falls over.
The logic that's performing the first() needs to gracefully handle empty lists as well.

Dev2:
Where is that logic called?

Tech Lead:
Here's the stack trace Dev1 provided in our conversation earlier:
Caused by: kotlin.NoSuchElementException: List is empty.
...
at 3 iosApp 0x00000000 kfun:kotlin.NoSuchElementException#<init>(kotlin.String?){} + 00
at 4 iosApp 0x0000000 kfun:kotlin.collections#first@kotlin.collections.List<0:0>(){0§<kotlin.Any?>}0:0 + 000
...
at 9 iosApp 0x0000000 kfun:kotlin.coroutines.native.internal.BaseContinuationImpl#resumeWith(kotlin.Result<kotlin.Any?>){} + 0000

This line:
kfun:kotlin.collections#first@kotlin.collections.List<0:0>()
...says that it's first() being called on an empty list.

Dev1:
FYI: Dev3/Dev4/myself are seeing the same issue with the same stack-trace above.

Tech Lead:
So Dev2, have you introduced such a call?
Because I checked master branch and there isn't one, in that version of the file.
Ok, I'll check your working branch Dev2
...
Yes you have here:
var processed1 = someList.first()
var processed2 = someList.first()
...
Lines 123, 124.
Solution looks really straightforward guys.

Dev2:
Okay, I will fix that and push the change

Tech Lead:
Check if someList is empty and allow for generating / handling null processedValues in the view.

Now; I'm going to be straight with you here.
This issue has been discussed over several hours today.
I expect that either one of you could have gone through the process I did in the last 10 minutes above, and resolved it in the same way :point_up:

Dev2:
I went on a break and it's not reproducible on my machine

Tech Lead:
I didn't reproduce it on mine either.

Dev1:
Dev2 and myself are now on sharing screen to sort this issue out. Hope to update back later.

Tech Lead:
<Screen shot of diff with changed code>
:point_up: That change should do it.

Dev2:
Already have pushed the change.

Tech Lead:
...just seen it, is good - same approach :ok_hand:
Dev1 please let us know when tested on your machine.

Dev1:
That does it. It fixes the issues. Thank you, Dev2. I will pick it off from here.

Tech Lead:
Glad to hear it guys.

Dev1:
I have to say this that it is not because we are not working on the issue - Dev2 and myself (together with Dev3/Dev4) have been on this issue all this morning. It just difficult to connect the dot when it wasn't reproducable on Dev2's machine. I brought the issue up because I wanted to switch to working on other tickets while waiting for this to resolve. Still thank you largely for Dev2's work and your keen eyes that spot and resolve the issue quickly.

Tech Lead:
Noted Dev1.
I think the take-away has to be to read the stack-trace carefully... don't worry - we've all been guilty of not reading the error in full, at some point.
The stack trace said that the 'first' element is being referenced from an empty list - that's just logically impossible, right?
Looking for that call to first, we saw it wasn't in the code before, and is after (two of them, in fact).
So then we ask ourselves, how can we deal with an empty list - and then solution almost presents itself.
It didn't really take reproduction of the error to resolve.
Maybe working with a new tech stack creates an anxiety that every issue faced will have a complex solution related to that stack; but I think you'll agree, this particular issue really just required a deep breath and your trusty 'debugging skills 101'... don't lose them! :smiling_face:

Damn race condition! I know you're hiding in there, just waiting for the morning to arrive and shit my day.

Once upon a time i had a great idea.
Because i couldnt be bothered to do anything productive i created a simple app in the C# that would look into every .js file (from a game that uses it for the gui/main menu) and search for "//todo" lines.

I did it mostly for kicks. I got that idea when i encountered one //todo in a file when i was trying to mod that game.

Yes i know grep exists: fuck you.
It would have taken me more time to learn that than to write that 20 line program...

The result? Over 30 lines of //todo with some briliant pearls in the type of:
>Temp workaround because X
>Workaround for race condition
>Clean that up
>Obsolete

When i return home i will post real quotes. They might be amusing to read...

The game is based on a custom C++ engine. HTML, CSS and JS is used for main menu and some graphical interface in game.

The most amusing thing is that this inefficient sack of chicken shit is powering one of the biggest (no playerbase but unit, world, gameplay vise) rts that i have ever played.

But still in spite of a dead community, buggy gui as shit and other problems i love this game and a lot of other people love it too. It is a great game when it works correctly.

To the interested: JS portion uses jquerry and knockout lib.

> Finally write some wicked code that fixes a weird race condition bug we’ve had for two weeks now

> xCode breaks for absolutely no reason on a build so we can’t release a new version

🤦‍♀️

Can I go to CTO and demand that he will demote a "senior engineer" (who takes senior's paycheck) who don't know what "race condition" is, and how to design/deal with it?

Why? Why I can't?

Wrote some code, not realizing code has TOCTOU race condition until a healthy 4 hours later, and the fix was to move two lines of code upward. I hate me, myself and I.

P/S: What happened to PrivateGER, devTea and linuxxx?

I'd tinkered with computers for a long time but the breakthrough moment for me was a robotics class in elementary school where I programmed Legos in TC Logo.

That summer, I made a washing machine with multiple cycles and a door sensor to interrupt the cycles.

Soon after, I played with the code for Gorillas in QBasic to fix a race condition when running it on my 486 at home.

Today my code went into race condition condition.

I was frustrated and proud at the same time

A possible race condition

Knock, knock.
Race condition.
Who’s there.

can 2 modal which is static have a race condition error? I’m trying to fix a bug and when I showed it to my senior , I said maybe because of the first modal the style is not being applied on the second modal he then said “it might be a race condition” and I almost choked on my own saliva.

Ps: the first modal will show a loading text and a gif after that the second modal shows with some data.

I got confused how he think that it might be a race condition. We’re not doing webworks on the front end. Weird.

Knock knock?
RACE CONDITION!

:-|