I was a full time test engineer / QA person for a while. My motto quickly became “nothing ever works”.
Pretty much any ticket behind a static copy change would have some problem or oversight. Sometimes even those would (did you account for very narrow view ports?)
Good developers would take this feedback gracefully. “Shit, you’re right, I need to account for mobile users.”
Bad developers would get defensive and upset. “We barely have any mobile users (me: did you check?). Alan already approved so I’m merging. I don’t want to waste time on this”
As a dev I’ve been on both sides to be honest. Especially when there is pressure to finish the next task. I think it needs good planning to create enough time for these things.
In the end bad devs will still shut you up about things they are not interested in fixing…
I’ve done a lot of “then go get approval from the stakeholder to go ahead with this bug/problem”.
If product wants it out now now now they can sign off on it not working on mobile, so when their boss has a fit about it I can point to the conversation where Ryan said it was fine.
I’ve mostly worked at smaller companies though.
I’ve caught problems in code review and had to do this even.
Often it’s reading it and realizing there’s a complicated edge case or they missed something entirely.
Sure we can make a different ticket for that to move this along, but we’re getting product to agree first.
Me neither buddy, me neither…
Falsehoods About Time: … Time always moves forwards.
I had to learn this the hard way… I was working at a platform that pulled measurements from sensors. The sensors did not declare the timezone for the timestamps of the measurement and the platform broke down twice after daylight saving. The first time there were duplicated records which caused conflicts and the second one we weren’t handling impossible timestamps.
I had a client whose clock was just a few milliseconds behind the server’s, but due to timezone crap one hour in the past. And the signature was valid for one hour.
If the network just happened to be too congested, the validation failed. The next request went through just fine. Took us forever to find out.
Epochs aren’t that simple either.
First of all, local time can be relevant, so you have to store timezone information somewhere anyway.
Epochs are also somewhat iffy in regards to leap years or seconds.
And finally: write me an SQL to retrieve all entries submitted in 2022 using just epochs.
Timezones are annoying as fuck, don’t get me wrong, but simply ignoring them isn’t a solution either.
One from JWZ: mysqldump writes out a date that it cannot parse (and more)
Something I find helpful is to PR review my own code before I create the actual PR. It’s surprising how giving it a once over in a different setting to the comfort zone of your code editor can save you a bit of unnecessary back and forth.
Yep. This is the way. Also, you’d be surprised how many devs don’t run through their own QA steps before asking other people to verify.
Yep, shit devs will rely on those reviewing the PR to do their work for them.
I like to take any text and copy / paste it into a different editor. Inevitably, this changes the layout because of different font settings or window size or whatever else. Reading it in a slightly altered layout helps me catch a lot of tiny errors that my eyes otherwise glide past.
It’s so fucking goddamn sad how rare this is. Personally I review changes when I push commits and review all changes personally before publishing the PR and requesting reviews. I’m a senior dev so I guess it’s just something that comes with time (if you care about the code you produce)
I swear to fucking god every PR review I do for younger devs and contractors we have to use (guess what country they’re from), I swear they don’t even look at their own fucking code once they’ve written it, let alone perform any sort of critical analysis. thoughts about any missed use-cases/optimization/future application and evolution? Nah!! That’s what the PR is for right? 🤬
As you said even just a quick once over make a big difference and somehow it’s not even close to common.
Damn I need to quit
This is my typical experience as well, too many people don’t do a code review of their own PR first.
When I was a junior, I had this coworker who did all my reviews. I was doing my absolute best and wanted to show that I was learning, so I would review all my work before submitting it and think, how would he review and respond to this code.
That just stuck with me and it’s my normal practice now.
I eventually learned that’s not as normal as I thought. I also tend to give better code reviews than others.
Edit: the other thing I do is check in with who will be reviewing my code well before I submit anything someone might think is weird and have a discussion about it before the reveiw. If it’s weird, there might be a better way unless were stuck due to technical debt or something, and doing that early vs at the end usually saves time.
I wouldn’t do that, too much tunnel vision and biases. I just skim through and make sure everything makes sense. Especially naming and comments.
I wouldn’t do that, too much tunnel vision and biases.
Absolutely not. Self-reviews are very productive. I can confirm this from my own work and my colleagues, who also find it so.
You’re of course free to vary the degree and depth of self-review, but tunnel vision and bias is definitely not overbearing and diminishing in those situations for us.
Someone else will of course see more, what you may not see due to tunnel vision. But that’s besides the point.
Weird, never heard of anyone doing this. Aren’t your team self reviewing the code while writing it?
I just skim through and make sure everything makes sense. Especially naming and comments.
What do you think a self review is?
Falsehoods About Time
Having a background in astronomy, I knew going into programming that time would be an absolute bitch.
Most recently, I thought I could code a script that could project when Easter would land every year to mark it on office timesheets. After spending an embarrassing amount of…er…time on it, I gave up and downloaded a table of pre-calculated dates. I suppose at some point, assuming the code survives that long, it will have a Y2K-style moment, but I didn’t trust my own algorithm over the table. I do think it is healthy, if not essential, to not trust your own code.
Falsehoods About Text
I’d like to add “Splitting at code-point boundary is safe” to your list. Man, was I ever naive!
I’m sure many smaller companies had their own internal Y2K moment as they scaled and became a big hit, and realized they used a wrong datatype like int instead of long or something and shit was gonna break by XYZ date if they did nothing heh.
Give it long enough and somehow the person who decided on IPv6 will feel the same as every piece of matter we want to interact with can be networked.
Hey I know this place! I filmed there it’s in northern Ireland, behind the camera there’s an Observatory kind of building (round thing).
