|
Is it viable to split the tests in two or more test suites? Running all tests once per release may be necessary, but perhaps it's sufficient to run a smaller subset on every commit?
|
#
?
Sep 18, 2023 16:09
|
|
|
|
| # ? May 10, 2024 14:47 |
|
|
Apoffys posted:Is it viable to split the tests in two or more test suites? Running all tests once per release may be necessary, but perhaps it's sufficient to run a smaller subset on every commit? I doubt "but it passed half the tests!" is going to fly.
|
|
#
?
Sep 18, 2023 16:47
|
|
|
There is probably a much smaller set of tests which would be good enough to run per-commit, but identifying that set is an extremely difficult problem.
|
|
#
?
Sep 18, 2023 17:00
|
|
|
Obviously the whole test suite needs to pass before the code is released to production. I assumed the scenario was "commits happen daily, releases are more rare". As long as the full test suite is run before release, any issues should be found. It sounds like a simplified test suite would be useful in any case, just for development purposes.
|
|
#
?
Sep 18, 2023 17:22
|
|
|
Agreed, I wonder if their CI system has the ability to run The Big Test Suite on merge, and The Small One on commit. Or hey, maybe no tests on commit, “it’s builds!” is enough for us on feature/bugfix branches. It does sound like the eyebrow-raising part here is doing a big thing on every commit.
|
|
#
?
Sep 18, 2023 18:40
|
|
|
If you've got a test suite that takes twelve hours to run, I wonder if many of the test cases can be converted from integration tests to unit tests. Reduce the number of dependencies that need to be spun up in the tests whose purpose isn't to verify that they are integrating correctly, that sort of thing. If you've got a matrix of test cases that result in a combinatorial explosion, parameterize them and narrow the focus of the test into as close as you can get it to a single pure function. Your codebase will benefit from having the appropriate abstractions that make this possible. Tests taking forever is a signal, too: if there's a lot of detail which you're only able to use integration tests to verify, then that's a sign that you've got too much logic embedded in the architecture of the system rather than in the parts that are easy to isolate and change.
|
|
#
?
Sep 18, 2023 18:56
|
|
|
Has there ever been anything as huge a waste of time as the performance review?
|
|
#
?
Sep 19, 2023 11:21
|
|
|
Mega Comrade posted:Has there ever been anything as huge a waste of time as the performance review?
|
|
#
?
Sep 19, 2023 11:25
|
|
|
Jamus posted:However, it is also the kind of domain where correctness issues can be statistical due to unexpected CPU branch predictor results or similar factors. Alright you got me there, I guess testing for unexpected CPU branch optimization necessitate bombarding huge amount of inputs that must be run directly on a specific hardware… However, if you are not writing the firmware or controller for some embedded component it could be argued that you are testing some other organisation code rathe than yours.
|
|
#
?
Sep 19, 2023 13:35
|
|
|
AskYourself posted:However, if you are not writing the firmware or controller for some embedded component it could be argued that you are testing some other organisation code rathe than yours. This isn’t a great soapbox to stand on, realistically. Squirrelly branch prediction might be a problem in microcode but no end user has ever had their app crash, or their device not do what it was supposed to do, and thought “gently caress I’m never buying a device with a microcontroller from NXP again!” It’s still incumbent on the developer to make sure their product works on a given vendor’s crap if that’s how it’s being sold. If I knew getting lost in branch prediction was a mission-critical problem my code had I’d sure as poo poo be regression testing that case however possible - while also calling up the hardware vendor and trying to figure out what the gently caress
|
|
#
?
Sep 19, 2023 13:45
|
|
|
Mega Comrade posted:Has there ever been anything as huge a waste of time as the performance review? Mega Comrade, great job writing a bunch of nice sounding bullshit about the bullshit goals we set in Q1 that became irrelevant a week later. I like you, so here's a 5% raise. I don't like your coworker so you get their 2.5%. Congrats on the huge raise!!
|
|
#
?
Sep 19, 2023 14:07
|
|
|
a dingus posted:Mega Comrade, great job writing a bunch of nice sounding bullshit about the bullshit goals we set in Q1 that became irrelevant a week later. I like you, so here's a 5% raise. I don't like your coworker so you get their 2.5%. Congrats on the huge raise!! Oh hey boss, what's up
|
|
#
?
Sep 19, 2023 14:59
|
|
|
Mega Comrade posted:Has there ever been anything as huge a waste of time as the performance review? Every meeting where somebody just reads you a document you could have read yourself is indeed a complete waste.
|
|
#
?
Sep 19, 2023 16:02
|
|
|
We are doing the worst kind too, self reviews 
|
|
#
?
Sep 19, 2023 16:21
|
|
|
I left mine blank 
|
|
#
?
Sep 19, 2023 16:37
|
|
|
Mega Comrade posted:We are doing the worst kind too, self reviews Oh that's easy, say you made progress towards ${GOAL} but you wish you could do more
|
|
#
?
Sep 19, 2023 16:59
|
|
|
We just set our department and personal goals for 2023 in the last month, as an indication of how worthwhile that whole endeavour is.
|
|
#
?
Sep 19, 2023 17:42
|
|
|
We just rearranged our teams to put our good engineers in one team to deliver something and everyone else in another. This will be fine.
|
|
#
?
Sep 19, 2023 18:36
|
|
|
The moment of realization that you’re in the bad team.
|
|
#
?
Sep 19, 2023 19:36
|
|
|
Gangsta Lean posted:The moment of realization that you’re in the bad team. https://twitter.com/Shreyash_Jain_/status/1697678370761478418?s=20
|
|
#
?
Sep 19, 2023 20:27
|
|
|
Do "the good engineers" become bad engineers when they fail to hit arbitrary deadlines set by clueless mgmt?
|
|
#
?
Sep 19, 2023 20:35
|
|
|
Xguard86 posted:Do "the good engineers" become bad engineers when they fail to hit arbitrary deadlines set by clueless mgmt? All engineers are bad, so by definition they were always bad in the first place.
|
|
#
?
Sep 19, 2023 20:39
|
|
|
We do self reflections but according to both my manager and skip level no one actually reads them so I spend about five minutes total writing them.
|
|
#
?
Sep 20, 2023 06:48
|
|
|
We have three Matts on our team and we have an applicant named Matt who is actually a good candidate A real life Matt-le Royale, if you will
|
|
#
?
Sep 20, 2023 16:37
|
|
|
Macichne Leainig posted:We have three Matts on our team and we have an applicant named Matt who is actually a good candidate So, how do they go around calling you? Is it Matt, Matthew and...? Was in a similiar situation a few years ago, sharing a name with a team member. A fellow colleague from another team proposed calling us "<Name> one" and "<Name> two". I refused, and pointed out that software engineers start counting from zero, so they should call the other team member "<Name> one", and call me "<Name> zero". It didn't stick, but at least we avoided the unfortunate implication of ranking.
|
|
#
?
Sep 21, 2023 08:30
|
|
|
At my previous job we had three, at one point four people with the same first name. We usually referred to them by their last name. Apparently the three of them were some of the first people hired by the founders and it was up to them to demo the very first version of the product to the MT of a large company who would become the first customer. They told me they went on 'stage' and were like "hi, I'm <firstname>" "and I'm <firstname>" "and I'm <firstname>".
|
|
#
?
Sep 21, 2023 17:28
|
|
|
We do Matt plus first initial of last name, which worked until we got four Matts with two with one shared last initial and two with a different shared last initial
|
|
#
?
Sep 21, 2023 17:57
|
|
|
Mecca-Benghazi posted:We do Matt plus first initial of last name, which worked until we got four Matts with two with one shared last initial and two with a different shared last initial Unfortunately you're just going to have to kill one of them.
|
|
#
?
Sep 21, 2023 18:01
|
|
|
Changing names here, but we have two guys named Joe, two guys named Rob, two guys named Sam and two guys named Seth. We had two guys named Matt and another Rob up until about a month ago. Our stakeholders include two guys named Rob, two guys named Seth, a guy named Matt, and a guy named Joe. I'm probably forgetting a few.
|
|
#
?
Sep 21, 2023 18:04
|
|
|
Worked at a company that had 3 Daniels out of 12 total engineers. Dan, Danny, Daniel man software development has way too many white dudes in america
|
|
#
?
Sep 21, 2023 18:11
|
|
|
We have this issue with Zach's There are 4 of them across 3 related teams and they're constantly in meetings together
|
|
#
?
Sep 21, 2023 18:12
|
|
|
Our office had two Jonathans and a corner with our two interns, John and Nathan
|
|
#
?
Sep 21, 2023 18:34
|
|
|
You should only refer to people by their UUIDs.
|
|
#
?
Sep 21, 2023 18:36
|
|
|
Shut up 89131
|
|
#
?
Sep 21, 2023 18:51
|
|
|
prom candy posted:You should only refer to people by their UUIDs. Unfortunately, we never backfilled anybody born before we added the UUID column. So everyone born before the 80s is basically the same person.
|
|
#
?
Sep 21, 2023 19:05
|
|
|
prom candy posted:You should only refer to people by their UUIDs. I mean we use SSN for everything else may as well use it for that too.
|
|
#
?
Sep 21, 2023 19:41
|
|
|
Just add a unique index on the first_name field and let the problem solve itself.
|
|
#
?
Sep 21, 2023 20:06
|
|
|
Cup Runneth Over posted:I mean we use SSN for everything else may as well use it for that too. Shockingly, SSNs are not unique! Multiple people can have the same one and sometimes a person gets more than one.
|
|
#
?
Sep 21, 2023 20:20
|
|
|
I think this came up earlier, but what do folks use for work notes/etc? I have been using Joplin but it's kind of clunky and I'm not that happy with it, so looking for alternatives. I like the markdown driven approach of Joplin, so Obsidian seems like a good fit, but the pricing is a bit high for me to justify at work. (50/year). I work for a Big Tech so I can't really use anything like Notion/etc, which is generally what I use personally, and I'd rather use notepad than OneNote. I loving hate OneNote.
|
|
#
?
Sep 21, 2023 20:43
|
|
|
|
| # ? May 10, 2024 14:47 |
|
|
I use a paper notepad and every time I fill up a sheet, I tear it off, put it in a drawer, and never look at it again.
|
|
#
?
Sep 21, 2023 20:47
|
|