|
Armauk posted:Oh nO BuT YoU CaN'T SaY bLaCkLiSt AnYmOrE!!
|
![]() |
|
![]()
|
# ? May 29, 2023 15:56 |
|
Armauk posted:Oh nO BuT YoU CaN'T SaY bLaCkLiSt AnYmOrE!! yeah that's right, shouldve said denylist or blocklist, habit of naming inertia, pardon me (that's how you should react to that)
|
![]() |
|
Share Bear posted:why even have config files, why not have just straight up a class/module with the config and that's what gets edited/templated spring boot cloud uses yaml files (or other k/v sources) to fill values in a config class using the @Value annotation
|
![]() |
|
its really hard for me to stop saying "youse guys" though cause of similar reasonsBlinkz0rz posted:spring boot cloud uses yaml files (or other k/v sources) to fill values in a config class using the @Value annotation ![]() Share Bear fucked around with this message at 17:16 on Sep 2, 2021 |
![]() |
|
if you use spring or other server-side stuff with kotlin gradle you can put your config classes in a source file and then reference that same file to type-safely instantiate the class in the kotlin build/test/deploy/run script, which is pretty neat except for the fact that you're using gradle
|
![]() |
|
the problems started when we tried to make computers imitate people by talking to each other.
|
![]() |
|
saw some galaxy-brained poo poo in a code review today:code:
code:
|
![]() |
|
that fuckin' rules
|
![]() |
|
Ciaphas posted:today i was trying to work out why some log messages were getting mixed together line-by-line in 20yo common code. was expecting to find that there were no mutexes or other synch and i'd have to add it, ho hum lmfao. in my own tp status, i learned something new about c++ today: C++ code:
Joe Chip fucked around with this message at 08:04 on Sep 3, 2021 |
![]() |
|
It's good, but I see it only used for constructors and insane tmp bullshit like variant
|
![]() |
|
I think thats good. If you're going to have overloads you might as well be able to implicitly overload an inherited method without needing to repeat it
|
![]() |
|
how do you think management would react if I told them that the project we're working on is doomed to failure due to technical debt
|
![]() |
|
animist posted:how do you think management would react if I told them that the project we're working on is doomed to failure due to technical debt "Technical debt? What's that?" "It's when imperfect decisions accumulate in the code and slow us down everytime we have to interact with it." "Why didn't you do it right the first time?" "But but but" "YOU'RE FIRED! HERE'RE YER WALKIN' PAPERS"
|
![]() |
|
animist posted:how do you think management would react if I told them that the project we're working on is doomed to failure due to technical debt "how do you know?"
|
![]() |
|
EDIT: nevermind I'm dumb and did the test wrong
Zaxxon fucked around with this message at 22:48 on Sep 3, 2021 |
![]() |
|
nvm then
|
![]() |
|
animist posted:how do you think management would react if I told them that the project we're working on is doomed to failure due to technical debt i've been wondering the same thing myself recently. problem is, we are still in the "make some shiny (but inherently flawed) feature we can show to justify our existence" mode which is what lead us to accumulate the debt in the first place. as soon as it feels like I've built some breathing room by reworking some of the decisions of the past, 4 new half-baked features get shoved down the pipeline
|
![]() |
|
happily I get to claim no responsibility for the tech debt because I joined after all the decisions were made. it really is excruciatingly bad. i am getting paid though so really i can't complain animist fucked around with this message at 02:20 on Sep 14, 2021 |
![]() |
|
Share Bear posted:why even have config files, why not have just straight up a class/module with the config and that's what gets edited/templated yeah. nowadays most of the time deploying a "config change" is exactly the same process as deploying a code change. it's just inertia that keeps people using config files.
|
![]() |
|
nah, it’s useful to have configuration clearly separated from implementation, and you literally can’t have that if it can execute arbitrary code inside your application
|
![]() |
|
I must have a different definition of configuration because we do spring and absolutely still have config files at runtime for different environments, and those files are not part of the code repo. There's no getting rid of those (well theoretically you could hardcode all this stuff (e.g. your DB URL) into the app, but... surely not?) Yeah, there's stuff that used to be xml at compile time and is now annotations (@Bean and friends), but that's not even what I'd call "config files" Ocean of Milk fucked around with this message at 09:56 on Sep 4, 2021 |
![]() |
|
animist posted:happily I get to claim no responsibility for the tech debt because I joined after all the decisions were made. hell, ![]()
|
![]() |
|
Got epicly trolled yesterday. Someone removed a test server as a deployment target for a rarely-updated project in Octopus, but didn't bother to remove it from the load balancer. Took way too many hours to figure out why my bug fix only worked 80% of the time.
|
![]() |
|
animist posted:i am getting paid though so really i can't complain careful, you might eventually be blamed
|
![]() |
|
speaking of config, I had a wonderful midnight call about some files not getting published that, after an hour and a half trying to work out where they were going and giving up and pushing manually, I realised the next day that these two jobs had some old setup where the publish location was in a separate config file that wasn't environment aware and was hard coded to non prod and then I dug up old emails and there was one from 5 years ago from me finding the same problem and going "well, these jobs are ancient and the one consumer needs to stop using them, I'm sure by the next time we deploy they'll have migrated to the replacement" ![]()
|
![]() |
|
this is perennially relevant re config http://mikehadlow.blogspot.com/2012/05/configuration-complexity-clock.html
|
![]() |
|
Bloody posted:this is perennially relevant re config http://mikehadlow.blogspot.com/2012/05/configuration-complexity-clock.html needs a station for 'buy a 3rd party tool that promises to end the need to code your workflows!' then spending years trying to migrate/upgrade/replace it idk if we still have it but we had an absolute spaghetti nest of dctm workflow stuff and right now, some consultant jackoffs are building a workflow in sharepoint 2013 for the low low price of $50k!
|
![]() |
|
Soricidus posted:nah, it’s useful to have configuration clearly separated from implementation, and you literally can’t have that if it can execute arbitrary code inside your application I admit one of the things I dislike about nodeJS is the lack of instant clarity on whether the code I'm writing is running on the server or the client, having different languages helps there, but I don't really buy this for configuration. It's pretty obvious if you're in a file that's just a set of configuration. a real reason to use xml/yaml/json is if multiple different applications need to consume the same configuration, but I don't see that too much in practice either
|
![]() |
|
configuration is something read in at runtime that establishes parameters for an env and sets up really important constants/flags and that functionally should be independent from the implementation however, that can still be accomplished by writing it in the target lang and loading it at runtime like any other dep injection methodology xml/json/yaml make this boundary more apparent, and likely reduce compile time, but if youre bouncing the application anyway (or triggering a ci/cd redeploy anyway) it just seems like a micro dsl i argue keeping it as text vs code isnt as advantageous or youre assuming your operations people will never have to know how your app is built and works (hahahaha)
|
![]() |
|
config contains the stuff that differs between environments
|
![]() |
|
all I know is if you have to rebuild your application artifact to change its currently deployed config you done hosed up
|
![]() |
|
some people like that but those people are nuts. the big risk with cute names like "immutable infrastructure" is that the english definition of the words becomes an explicit goal because some guy read a blog article, and the benefit gained from following whatever cute name becomes a secondary concern i had a job interview once where i described a process for deterministic deploys on aws autoscaling groups and my interviewer asked "why not just build an ami" and i had to be like "because i dont want each deploy to take 45 minutes".
|
![]() |
|
12 rats tied together posted:some people like that but those people are nuts. the big risk with cute names like "immutable infrastructure" is that the english definition of the words becomes an explicit goal because some guy read a blog article, and the benefit gained from following whatever cute name becomes a secondary concern We're at the "deploys take too long, let's do a project to speed them up" phase of the cycle.
|
![]() |
|
our deploy is some guy copying and pasting poo poo either into a remote SQL session, or to some random network share and the transferring the MSI to an RDP session because they can't work out how to download artefacts from nexus it's exactly as bad as it sounds and I genuinely didn't know how bad it was until this week when I watched it in real time and went "woah woah, this is what they've been doing"
|
![]() |
|
why do we keep writing new code we have code at home!
|
![]() |
DrPossum posted:why do we keep writing new code we have code at home! code at home: PHP code:
|
|
![]() |
|
cinci zoo sniper posted:code at home: is that the e-mail address regex
|
![]() |
|
Captain Foo posted:is that the e-mail address regex lol I've copied and pasted and hosed up that one so many times. 'oh but now we need to paste multiple addresses!' etc. another good one is an LDAP search. First + last, last, uid or email and all in the Garbo LDAP query syntax e: actually I guess it's just rpn or similar. Which also explains why it feels wrong because I never used rpn when I actually did calcs e2: I should also add that I generally only handle internal emails addresses so can home roll and modify a much simpler one. Until someone wants multiples. Or a mix of firstname.lastname@org and edls. Powerful Two-Hander fucked around with this message at 23:48 on Sep 4, 2021 |
![]() |
Captain Foo posted:is that the e-mail address regex yep https://emailregex.com
|
|
![]() |
|
![]()
|
# ? May 29, 2023 15:56 |
|
People who want legit email address instead of .*@.* Are thought criminals. Stop validating my inputs you jerks
|
![]() |