|
I met a traveller from an ancient forum who said I came across a great standing post made of stone, and on it these words are written: Soricidus posted:gently caress, I wrote my own json output code because how hard could it be nothing beside remains. Round the decay Of that colossal wreck, boundless and bare The lone and level gigabytes of text stretch far away.
|
#
?
Dec 10, 2019 22:30
|
|
|
|
| # ? Apr 26, 2024 18:51 |
|
|
Soricidus posted:gently caress, I wrote my own json output code because just run it through eval() (hope you don’t have any line or paragraph separators)
|
|
#
?
Dec 10, 2019 23:22
|
|
|
Soricidus posted:gently caress, I wrote my own json output code because if it's only a handful identify them and fix them manually it's only downhill from here, friend
|
|
#
?
Dec 10, 2019 23:25
|
|
|
fix your serializer and run it again
|
|
#
?
Dec 10, 2019 23:30
|
|
|
gonadic io posted:if it's only a handful identify them and fix them manually this is what I actually did but regex was my first idea (and funnier) Nomnom Cookie posted:fix your serializer and run it again this would have been the correct answer but the program that generated the json takes like a day to parse all the input data and I needed it to be ready yesterday Soricidus fucked around with this message at 23:38 on Dec 10, 2019 |
|
#
?
Dec 10, 2019 23:35
|
|
|
MononcQc posted:I just flat out don't remember having a thing go well. What the gently caress is nice tooling supposed to be? this is why you earn a salary i am now yelling at everyone about improving tooling and, if not possible, about writing docs should that nice tooling not exist people are happy in this weird wizard knowledge bullshit world (which is why this thread also exists) and i hate it Share Bear fucked around with this message at 23:57 on Dec 10, 2019 |
|
#
?
Dec 10, 2019 23:49
|
|
|
i proposed that stem-the-compiler-warning-bleeding project to boss yesterday and was told no because a major feature originally estimated for next october is now due in march cool. cool. guess those 236 compiler warnings will be 300 by the time i can be rejected again
|
|
#
?
Dec 10, 2019 23:57
|
|
|
just gotta hide a warning fix or two in each commit
|
|
#
?
Dec 11, 2019 00:01
|
|
|
Ciaphas posted:i proposed that stem-the-compiler-warning-bleeding project to boss yesterday and was told no because a major feature originally estimated for next october is now due in march your boss dgaf and will never approve your pet project
|
|
#
?
Dec 11, 2019 01:01
|
|
|
Ciaphas posted:i proposed that stem-the-compiler-warning-bleeding project to boss yesterday and was told no because a major feature originally estimated for next october is now due in march do it anyway and don’t ask for permission this time
|
|
#
?
Dec 11, 2019 01:19
|
|
|
just thought it'd be nice, is all. a recent late-stage bug was from code emitting a build warning no one saw, mired as it was in the logs; wasted a day instead of half an hour guess it's not my job to care about quality, though. my pay doesn't change.
|
|
#
?
Dec 11, 2019 02:20
|
|
|
Been implementing some networking stuff at work for a phone app feature. I've got classes to parse/build ip4 and udp headers, but that was the easy part and now I have to do most of a tcp stack implementation. This is gonna suck
|
|
#
?
Dec 11, 2019 02:23
|
|
|
Ciaphas posted:guess it's not my job to care about quality, though. my pay doesn't change. and ciaphas was enlightened
|
|
#
?
Dec 11, 2019 02:33
|
|
|
Captain Foo posted:and ciaphas was enlightened that's only part one. part two is working out how not to be annoyed, angry, or otherwise upset when things go awry in ways that reflect badly on me (due to e.g. ignoring warnings and creating a nasty regression months later) (part three is mentally reconciling part one with how much software dev sucks rear end - arguably entirely because of that conclusion!) Ciaphas fucked around with this message at 02:39 on Dec 11, 2019 |
|
#
?
Dec 11, 2019 02:37
|
|
|
could try something like fixing five warnings each day, maybe first thing in the morning or as a lil wind down before the end of the day. you’ll make it through in no time
|
|
#
?
Dec 11, 2019 04:55
|
|
|
suppose it's as good as anything else for the "gently caress i've been staring at this bug too long, need to do something else" role
|
|
#
?
Dec 11, 2019 05:22
|
|
|
Ciaphas posted:just thought it'd be nice, is all. a recent late-stage bug was from code emitting a build warning no one saw, mired as it was in the logs; wasted a day instead of half an hour alternatively, it is your job to care about quality because the thing the people above you care about are the second-order effects of quality. your boss doesn't give a gently caress about compiler warnings, but your boss might give a gently caress about the problems caused by not fixing them. if they're slowing down your ability to work on new feature development, then fixing warnings in the area that you're touching is an obvious first step in adding a new enhancement. if you just had a bug related to a warning that went ignored, then obviously part of fixing that bug would be to address that warning everywhere else it shows up too to make sure you fully fixed the bug and didn't just stop it from manifesting in one place. getting time allocated to do things like fix compiler warnings is about making a case for why doing so has actual business benefits bigger than the other things you could be working on rather than just "i don't like that this number is big and would be happier if it was small". managers can't make that evaluation, but convincing them that you can and that they can trust your evaluation will do good things for your career.
|
|
#
?
Dec 11, 2019 07:33
|
|
|
I fixed ~500 warnings during the last two months without a specific mandate to go fix bugs, I just roll it under "continuous improvements to the codebase that are required to keep adding features". It helps that I generally only spend "downtime" on them, e.g. when I am waiting on benchmarks for a new PR to finish.
|
|
#
?
Dec 11, 2019 08:41
|
|
|
MononcQc posted:It's all over various shapes of "download this visual studio code plugin, and press the following buttons" so it works kind of okay until you maybe want to do a thing that is not in the tutorial and then who the gently caress knows. redleader posted:lotta docs always seem to miss out on the tiny little detail of "how the gently caress am i actually supposed to run this in production"
|
|
#
?
Dec 11, 2019 19:12
|
|
|
cant imagine trying to build something using a microsoft toolchain in a CI/CD system
|
|
#
?
Dec 11, 2019 19:14
|
|
|
Soricidus posted:gently caress, I wrote my own json output code because well, you're certainly posting in the correct thread
|
|
#
?
Dec 11, 2019 19:14
|
|
|
Tbf vscode is an ide that understands the concept of version controlled configuration better than most. So you can do things like recommend extensions and also version control the config for said extensions The config files are all hand edited json but hey you can't have everything
|
|
#
?
Dec 11, 2019 19:19
|
|
|
Sapozhnik posted:The config files are all hand edited json
|
|
#
?
Dec 11, 2019 19:41
|
|
|
CRIP EATIN BREAD posted:cant imagine trying to build something using a microsoft toolchain in a CI/CD system just use the microsoft CI/CD system, the whole stack screams for "use everything developed by Microsoft from the computer up to the cloud components" and any place in the production chain where you want to use something different is going to work shoddily and any support you'll get from them will scream for "use the microsoft thing they provide". If you're a big enough business with a big enough partnership, you can impact the microsoft product timelines and get their sales engineers to write a solution using the microsoft stack for your problem. The rest of your developers can then be employed into clicking various Azure dashboards all day instead.
|
|
#
?
Dec 11, 2019 19:46
|
|
|
well i also cant imagine ever using a microsoft product in general
|
|
#
?
Dec 11, 2019 21:00
|
|
|
azure devops CICD sucks cause it uses procedural yaml based configuration. everyone is so against declarative build systems which is so loving aggravating.
|
|
#
?
Dec 11, 2019 21:04
|
|
|
lmao at procedural yaml config what a poo poo show
|
|
#
?
Dec 11, 2019 21:05
|
|
|
out of the box current Msbuild can sort of (but not really) do declarative builds but azure devops doesn't use that directly. it uses it via the yaml build config which makes it such a pain in the rear end.
|
|
#
?
Dec 11, 2019 21:08
|
|
|
CRIP EATIN BREAD posted:well i also cant imagine ever using a microsoft product in general outlook.com is good Shaggar posted:procedural yaml based configuration this is the opposite of good in conclusion, microsoft is a land of contrasts
|
|
#
?
Dec 11, 2019 21:11
|
|
|
ms exchange is a bag of incompatible poo poo-piss (so what MononcQc said pretty much)
|
|
#
?
Dec 11, 2019 21:56
|
|
|
about 2 years ago we had a ticket to change the compiler to treat warnings as errors and also fix all existing warnings and honestly it was a really really good loving idea
|
|
#
?
Dec 11, 2019 22:02
|
|
|
Azure DevOps is bad until you have to teamcity
|
|
#
?
Dec 11, 2019 22:03
|
|
|
Ciaphas posted:just thought it'd be nice, is all. a recent late-stage bug was from code emitting a build warning no one saw, mired as it was in the logs; wasted a day instead of half an hour Just do it! I spent the better part of a day last week changing all of our imports in a TypeScript project so that we have no use of tree shaking. And removing fixing up dependencies to remove cycles and stuff. Because jest, the test framework, doesn't tree shake and it was way to slow. Went from 120s+ to run the all the tests to about 50s. Getting the first test results after 5s instead of 30s. And being able to run a single test suite in less than 10s. I still think it's a bit slow, but holy poo poo this made a huge impact for me. I didn't ask permission, it was needful so I did it.
|
|
#
?
Dec 11, 2019 22:31
|
|
|
MononcQc posted:just use the microsoft CI/CD system, the whole stack screams for "use everything developed by Microsoft from the computer up to the cloud components" and any place in the production chain where you want to use something different is going to work shoddily and any support you'll get from them will scream for "use the microsoft thing they provide". this is totally how they do things though. if you stick within their ecosystem, the tools all work very well and mostly painlessly with each other if you're half in their ecosystem, then the friction from using non-ms tooling is supposed to gradually push you towards adopting more of their poo poo* if you're outside their ecosystem, they don't care because you're not a customer * and some of their poo poo is actually really good
|
|
#
?
Dec 12, 2019 01:46
|
|
|
i run mostly netcore apps on azure i use "dotnet build" in my dockerfiles and i use azure to provision each resource group with a ssh-enabled vm scale set, a blob storage, and a pgsql database. that's pretty much the extent of my reliance on ms stuff for everything else - from testing to ci/cd to ssl termination, logging, oauth - i use open source non-ms stuff, gitlab, elk, keycloak, you name it hell, i wouldn't even be using the azure managed sql databases - I'd have just deployed postgres + pgbouncer + a backup cronjob inside the swarm cluster - if it weren't for the fact that postgres doesn't work with nfs storage (technically it can, but you need to disable so many features that you lose acid) (notable exception: we're starting to work on some mobile apps, and we're likely to go with azure app center since it apparently provides a lot of integrations that are a pain to handle on one's own) the upside is that I'm currently working on another app which I'm doing in kotlin/springboot instead of netcore for various reasons, and it integrates flawlessly into our existing infrastructure NihilCredo fucked around with this message at 02:53 on Dec 12, 2019 |
|
#
?
Dec 12, 2019 02:50
|
|
|
dotnet core on aws is pretty easy and seamless, including serverless apps
|
|
#
?
Dec 12, 2019 07:01
|
|
|
you know what, I have decided I'm OK with mixed line endings in a single project. hell, in a single file. join me; set yourself free.
|
|
#
?
Dec 12, 2019 07:27
|
|
|
NihilCredo posted:i run mostly netcore apps on azure well yeah, if you've got your poo poo together and use things like 'more than one resource group' and put things into 'containers' and know how to make a 'vm' then it won't be as painful
|
|
#
?
Dec 12, 2019 08:24
|
|
|
zokie posted:Just do it! I spent the better part of a day last week changing all of our imports in a TypeScript project so that we have no use of tree shaking. And removing fixing up dependencies to remove cycles and stuff. can you elaborate on what you did here? I've been poking at tree shaking in our build but I don't think I understand what you did. how did avoiding tree shaking make it faster? did you just split stuff into smaller files to compile less?
|
|
#
?
Dec 12, 2019 08:57
|
|
|
|
| # ? Apr 26, 2024 18:51 |
|
|
How can you tell someone does too much webshit: treeshaking.
|
|
#
?
Dec 12, 2019 09:25
|
|