|
St Evan Echoes posted:most startups are, ultimately
|
#
?
May 21, 2018 15:16
|
|
|
|
| # ? Jan 23, 2025 20:10 |
|
|
Shaman Linavi posted:sry but i will respect the integrity of the challenge and instead post of a meme of similar quality a no-name poster i've never seen before dumping reddit bullshit into an otherwise useful thread this is cool and good and definitely constructive
|
|
#
?
May 21, 2018 15:17
|
|
|
Pollyanna posted:and then they turn around and get mad when they can’t find a “perfect pick” and the closest ones to that are missing X technology or have Y-2 years of experience this is a temporary problem some day you will be the one with the experience and the desired technology background. there is no substitute for time and effort
|
|
#
?
May 21, 2018 15:18
|
|
|
Notorious b.s.d. posted:a no-name poster i've never seen before dumping reddit bullshit into an otherwise useful thread yeah on sa people actually read the posts they are responding to sometimes
|
|
#
?
May 21, 2018 15:19
|
|
|
Sapozhnik posted:yeah on sa people actually read the posts they are responding to sometimes Not reading the posts he responds to is nbsd's latest gimmick.
|
|
#
?
May 21, 2018 15:39
|
|
|
i have atoned for my sins against memes and would like to say i feel fukken great about my interview tomorrow at house of zuck and am looking forward to turning into a gibbering man-ape when I pick up a marker
|
|
#
?
May 23, 2018 04:03
|
|
|
I’ve heard they’re a house built on php and c++, is there anything to this?
|
|
#
?
May 23, 2018 05:38
|
|
|
I'd just like to revel in the fact image macros got popular enough that posting them is generally frowned upon in these forums
|
|
#
?
May 23, 2018 12:24
|
|
|
Shaman Linavi posted:i have atoned for my sins against memes and would like to say i feel fukken great about my interview tomorrow at house of zuck and am looking forward to turning into a gibbering man-ape when I pick up a marker GL, show those zuckers what you're made of.
|
|
#
?
May 23, 2018 13:01
|
|
|
facebook created a terrifying unaccountable panopticon privy to everybody's deepest darkest secrets but otoh they created react so i mean it's a bit of a wash really
|
|
#
?
May 23, 2018 13:55
|
|
|
Sapozhnik posted:facebook created a terrifying unaccountable panopticon privy to everybody's deepest darkest secrets but otoh they created react so i mean it's a bit of a wash really Facebook is bad; but it also made a javascript framework which is also bad, so its impossible to say if its bad or not
|
|
#
?
May 23, 2018 13:58
|
|
|
Got an onsite interview at the "we release to production 3 times a day" joint downtown. I'm pretty keen to hear how frequently they are fighting fires.
|
|
#
?
May 23, 2018 17:00
|
|
|
i had stuff roll to prod everyday before, but that was with a day of soaking before each release. worked fine
|
|
#
?
May 23, 2018 17:14
|
|
|
Progressive JPEG posted:i had stuff roll to prod everyday before, but that was with a day of soaking before each release. worked fine I think it could work if you have lots of automated tests, I'm getting the feeling they might be on the dark side with that though. Good chance to find out though.
|
|
#
?
May 23, 2018 17:17
|
|
|
qhat posted:I think it could work if you have lots of automated tests, I'm getting the feeling they might be on the dark side with that though. Good chance to find out though. who cares? if their ci/cd pipeline is good enough bugfixes can be rolled out in a matter of minutes.
|
|
#
?
May 23, 2018 17:35
|
|
|
One of our clients has me coming in to help them with interviewing candidates and it's been an interesting experience being on the other side of the table for the first time. Also this is probably the best way to do it because I'm literally not allowed to give any feedback on the interviews, they just sit in on them and judge how the candidate responds to my questions. So there's zero pressure on me.
|
|
#
?
May 23, 2018 18:03
|
|
|
the whole premise of ci/cd and multiple daily deployments is that you're rolling out small changes very often. in theory the smaller changes are easier to review, and thus less likely to have bugs. a side aspect is that if you do break something, it should take only a few minutes more than it takes to fix the bug in the code to get it rolled out. there's a whole bunch of things that go with it (lots of automated tests, vetting things in a staging environment, people actually reviewing prs instead of auto-merging) to reach that point. but ultimately, I really prefer it to most of the alternatives I've experienced. one of the places I interviewed apparently does after hours deploys once every couple weeks, and the whole process takes an hour or so. that's a pretty big negative to me really
|
|
#
?
May 23, 2018 21:30
|
|
|
EVGA Longoria posted:the whole premise of ci/cd and multiple daily deployments is that you're rolling out small changes very often. in theory the smaller changes are easier to review, and thus less likely to have bugs. a side aspect is that if you do break something, it should take only a few minutes more than it takes to fix the bug in the code to get it rolled out. What do you do when you need to roll out something more substantial?
|
|
#
?
May 23, 2018 21:43
|
|
|
qhat posted:What do you do when you need to roll out something more substantial? incremental rollouts with feature flags or reconsider the "thing" entirely and determine if it needs to actually be deployed all at once. usually producing services and persistence layers can be rolled out and validated before consumers, etc. this process requires really good build tooling for a monolith or decent tooling for microservices
|
|
#
?
May 23, 2018 21:57
|
|
|
yeah the feather in my cap on a big client i worked last year was cracking the whip over a team of about 50 to get them over to a ci/cd pipeline. it was a colossal pain in the rear end dealing with all the whining and stubbornness from both the qa and developers with a shitload of set backs. once we actually got there though they realized that the crazy increased reliability and predictability of our deployments meant no more all-night emergency frenzies plugging shitloads of holes after a massive deployment. their management noticed a large upswing in productivity as well due to the huge drop in reworking lovely code. i'm not gonna downplay how much of a pain it is getting a team that's not used to it switched over though. it took over half a year of dealing with lovely cowboy coders, qas who refused to learn how to automate, and business analysts who refused to actually write down AND UPDATE their requirements. it's definitely worth it though- just try to make it someone else's responsibility if you can
|
|
#
?
May 23, 2018 22:41
|
|
|
i was doing well until someone said something about hundreds of millions of dollars and the whole scale of everything hit me and then during the last coding segment i forgot about the advanced cs concept of CompareTo so if you saw some loser walking up bayfront expressway in an attempt to become one with nature that was me
|
|
#
?
May 23, 2018 23:45
|
|
|
double posting to say I was super ready to talk about GDPR and zuck69s talk in europe yesterday but it never really came up probably for the better
|
|
#
?
May 24, 2018 00:20
|
|
|
Lol at being scared of deploying, you cowards
|
|
#
?
May 24, 2018 00:32
|
|
|
Continuous deployment is fine for 99.9% of web apps because it doesn't matter at all if they're broken.
|
|
#
?
May 24, 2018 00:39
|
|
|
Rex-Goliath posted:yeah the feather in my cap on a big client i worked last year was cracking the whip over a team of about 50 to get them over to a ci/cd pipeline. it was a colossal pain in the rear end dealing with all the whining and stubbornness from both the qa and developers with a shitload of set backs. once we actually got there though they realized that the crazy increased reliability and predictability of our deployments meant no more all-night emergency frenzies plugging shitloads of holes after a massive deployment. their management noticed a large upswing in productivity as well due to the huge drop in reworking lovely code. My case is the opposite. Getting the developers on board was easy, but getting the management to even recognise that CI is more than just a luxury was really difficult and honestly the people that matter still aren't fully on board. To them, anything that distracts developers from producing the fastest possible release right now is not worth spending time on. The irony is that's exactly the thing the developers are trying to do lol.
|
|
#
?
May 24, 2018 00:49
|
|
|
qhat posted:What do you do when you need to roll out something more substantial? like it was mentioned, this goes along with an iterative approach. “thin slice” is the term people love to throw around. that said if you DO need a big change, because it’s foundational or structured in such a way that it’s all or nothing, you just add more process time. I replaced our session framework at work, which really just meant putting together an in person code review and ordering lunch for the group. nothing about ci or cd requires that you deploy 3x a day or anything.
|
|
#
?
May 24, 2018 03:49
|
|
|
Shaman Linavi posted:i was doing well until someone said something about hundreds of millions of dollars and the whole scale of everything hit me and then during the last coding segment i forgot about the advanced cs concept of CompareTo yeah this happens to me when i think about the thing i am working on already being sold to customers
|
|
#
?
May 24, 2018 04:42
|
|
|
I seriously bombed my 2nd tech screen for a company yesterday. I knew it was a valley company and they had some of that ridiculous algorithm interview stuff in them, but this was just especially rough for a front end developer.
|
|
#
?
May 24, 2018 11:42
|
|
|
EVGA Longoria posted:like it was mentioned, this goes along with an iterative approach. “thin slice” is the term people love to throw around. This sounds like a super healthy working environment. I will totally crib the idea of in person, bribed (with food) code reviews for a few devs at once.
|
|
#
?
May 24, 2018 17:57
|
|
|
Although, CI is continuous integration which is easy to do several times a day even with large codebases. CD is the continuous deployment part, which does suggest several daily deploys. Which is why I've seen CI thrown around as a great option for all cases, and CD be less universal. I still haven't heard a good answer to how you do CD when you have literally eight hours of automated tests or, good forbid, flakey tests peppered throughout your eight hour automated test suite.
|
|
#
?
May 24, 2018 18:03
|
|
|
TimWinter posted:Although, CI is continuous integration which is easy to do several times a day even with large codebases. CD is the continuous deployment part, which does suggest several daily deploys. if the tests are taking too long you need to beef up your test server or parallelize the tests. if your tests are flakey then write better tests
|
|
#
?
May 24, 2018 18:28
|
|
|
e: woops this isn't the terrible programmer thred5
|
|
#
?
May 24, 2018 18:32
|
|
|
Rex-Goliath posted:if the tests are taking too long you need to beef up your test server or parallelize the tests. if your tests are flakey then write better tests
|
|
#
?
May 24, 2018 19:39
|
|
|
that sounds like something a really bad programmer would say hth
|
|
#
?
May 24, 2018 19:59
|
|
|
Rex-Goliath posted:if the tests are taking too long you need to beef up your test server or parallelize the tests. if your tests are flakey then write better tests I'll see you better tests and raise you more tests. Cause I always got more tests.
|
|
#
?
May 24, 2018 20:16
|
|
|
Corla Plankun posted:that sounds like something a really bad programmer would say hth
|
|
#
?
May 24, 2018 20:19
|
|
|
Lol at 8 hours of tests Those are bad tests
|
|
#
?
May 24, 2018 20:19
|
|
|
Mao Zedong Thot posted:Lol at 8 hours of tests
|
|
#
?
May 24, 2018 20:29
|
|
|
Is the idea that you don't need to compile or deploy your software before you do integration or e2e tests? I'm bringing my ignorance to the bad programmer thread
|
|
#
?
May 24, 2018 20:37
|
|
|
|
| # ? Jan 23, 2025 20:10 |
|
|
Mao Zedong Thot posted:Lol at 8 hours of tests in hardware sims 8 hours is actually pretty short for a full soc. most of our turnin gate regression lists on a big project take 24-48 hrs. do a turnin, work on something else for two days, see if it got through the turnin pipeline, repeat.
|
|
#
?
May 24, 2018 20:40
|
|