|
this image is everything wrong in the world
|
# ? Jan 1, 2016 13:17 |
|
|
# ? May 24, 2024 21:17 |
|
devenv is great but turn off capital menu bars you psycho
|
# ? Jan 1, 2016 14:11 |
|
i wish that the asp.net 5/webapi were usable on os x but i can't even get the generator to update/install correctly for some reason. thanks npm
|
# ? Jan 1, 2016 17:15 |
|
nothing works on Linux. you have to use windows if you want things to just work
|
# ? Jan 1, 2016 17:25 |
|
i want my asp.net 5 app to spawn a long-running background worker thread. where should i do that? the configure method? sounds like that's kind of the start-up point?
|
# ? Jan 1, 2016 17:27 |
|
the general wisdom is don't put long running background tasks in your webapp, but if you really want to then startup.cs is probably the right place. what is the goal of the thing cause if its something that handles work processing independently of the web app then putting it into a windows service or some other similar long running service application might be more appropriate
|
# ? Jan 1, 2016 17:36 |
|
you can use the new DI stuff to configure it as a service in startup.cs ConfigureServices likeC# code:
C# code:
|
# ? Jan 1, 2016 17:38 |
|
i guess it doesn't really need to be long-running. the app is just pulling poo poo from a few other internet apis and serving up cached results on page load. i could just hit the apis on page load but that sounds like it would be slow so i was gonna have a background worker just hit the apis every #update_period and cache that data in the webapp
|
# ? Jan 1, 2016 17:48 |
|
~Coxy posted:1. lol ensuring that your update scripts produce the same thing as the create-from-scratch script is not difficult. as part of testing every commit that modifies either of them, your ci server just has to run the previous create script, run the newly added update script, and then run the new create script, and verify that they give identical results. if you don't have a ci server testing everything that you can ever dream of testing then you should fix that as it's pretty awesome idk what to say other than that i have debugged sprocs in the range of a hundred or two lines and while having to copy and paste poo poo rather than being able to step over lines was dumb and annoying, proper debugging tools would not have fundamentally changed my debugging workflow in the way that they do for compiled languages. a debugger would be nice to have, but the lack of it does not make things impossible to debug mocking out internal components is an antipattern that leads to useless tests. if your stored procs are making network requests or something then lol? not having useful refactoring tools is the norm, not something unusual about sql. having anything more useful than an "extract method" which doesn't always produce working code and a "rename class" that also sometimes renames some unlreated things is pretty much limited to java and c#
|
# ? Jan 1, 2016 18:01 |
|
Bloody posted:i guess it doesn't really need to be long-running. the app is just pulling poo poo from a few other internet apis and serving up cached results on page load. i could just hit the apis on page load but that sounds like it would be slow so i was gonna have a background worker just hit the apis every #update_period and cache that data in the webapp http://haacked.com/archive/2011/10/16/the-dangers-of-implementing-recurring-background-tasks-in-asp-net.aspx/
|
# ? Jan 1, 2016 18:02 |
|
my broader point is that while there isn't really a good standard ecosystem of tools for applying proper software engineering to your database stuff, that's still no excuse for not doing it anyway. if you're writing tests for your sql and doing absolutely everything you can to version your schema, then it's reasonable to complain about the shortcomings of your tools. it's not reasonable to look at the tools available and just decry that sql is "untestable" and just give up (or give a bunch of money to redgate and then give up) the sad state of software engineering when it comes to database things is a cultural problem, and the tooling problem is a consequence of that, not the cause
|
# ? Jan 1, 2016 18:02 |
|
Bloody posted:i guess it doesn't really need to be long-running. the app is just pulling poo poo from a few other internet apis and serving up cached results on page load. i could just hit the apis on page load but that sounds like it would be slow so i was gonna have a background worker just hit the apis every #update_period and cache that data in the webapp
|
# ? Jan 1, 2016 18:08 |
|
yeah but i dont know how to do any of that
|
# ? Jan 1, 2016 18:54 |
|
literally all i know is i can take an asp.net project and hit publish and then its in the cloud. everything else is unknown to me
|
# ? Jan 1, 2016 18:55 |
|
can i create a separate project that also somehow publishes to the cloud, and then have the first project ask the second project for stuff? how does any of that work? i have no idea. i can come up with a dozen different ways to do this like just on my pc or something but... cloud
|
# ? Jan 1, 2016 18:56 |
|
i guess i could make em talk to each other through a database. thats a thing i know how to do. but then i gotta have a database, and i dont want to
|
# ? Jan 1, 2016 18:58 |
|
Bloody posted:i guess it doesn't really need to be long-running. the app is just pulling poo poo from a few other internet apis and serving up cached results on page load. i could just hit the apis on page load but that sounds like it would be slow so i was gonna have a background worker just hit the apis every #update_period and cache that data in the webapp if you know everything that might be accessed then doing a big round of precaching at startup might be an ok idea, but If the data changes then you should be requesting it during each page load. use async controller actions and async requests to get the data. if you want to cache create a memorycache and check it for the data during the page load, if it doesn't exist grab it from the api and stuff it in the cache. you can configure the cache expiry when you add it to the cache. if the api provides cache headers you can base your caching on that
|
# ? Jan 1, 2016 19:09 |
|
you might need some kind of persistence if you're going to do results caching so you may as well use a db if your data is structured or a queue or something use azure webjobs and write a exe that does your querying and stores the data in a db/queue then hit that datastore from your webapp
|
# ? Jan 1, 2016 19:11 |
|
eh thats probably easy enough
|
# ? Jan 1, 2016 19:17 |
|
Shaggar posted:nothing works on Linux. you have to use windows if you want things to just work i will admit that i have had to manually hunt down, download and install more drivers in the couple months i've been using my work macbook than i had to in the near decade or so since windows was able to do that for you and linux came with all drivers ever made built in. and these are drivers for new things like modern printers and scanners too, not even some ancient parallel port garbage or some bullshit like that. "it just works" my rear end
|
# ? Jan 1, 2016 19:18 |
|
NihilCredo posted:give paket a try maybe? it's basically "nuget, but a bit better" jesus not a single version number/git tag, just pull and hope
|
# ? Jan 1, 2016 19:27 |
|
move fast, break dependencies
|
# ? Jan 1, 2016 19:28 |
|
Parallel Paraplegic posted:i will admit that i have had to manually hunt down, download and install more drivers in the couple months i've been using my work macbook than i had to in the near decade or so since windows was able to do that for you and linux came with all drivers ever made built in. and these are drivers for new things like modern printers and scanners too, not even some ancient parallel port garbage or some bullshit like that. "it just works" my rear end lol if you print in 2016
|
# ? Jan 1, 2016 19:42 |
|
AWWNAW posted:lol if you print in 2016 actually most often it's been scanners, OSX seems to loving hate scanners without installing the 200MB
|
# ? Jan 1, 2016 19:54 |
|
hackbunny posted:jesus not a single version number/git tag, just pull and hope its the web "development" way
|
# ? Jan 1, 2016 19:56 |
|
what's the typical lifespan of a web "app" these days anyway? 1 year?
|
# ? Jan 1, 2016 20:02 |
|
Parallel Paraplegic posted:actually most often it's been scanners, OSX seems to loving hate scanners without installing the 200MB also linux driver support for gpu, sound, drawing tablets, etc is atrocious + buggy
|
# ? Jan 1, 2016 20:04 |
|
like shaggar says, if you want something to work, use windows
|
# ? Jan 1, 2016 20:05 |
|
MeruFM posted:That just sounds like scanner makers being poo poo and not adhering to some supposed standard. things linux developers dont use.txt
|
# ? Jan 1, 2016 20:05 |
|
hackbunny posted:jesus not a single version number/git tag, just pull and hope i assume they're at least recorded in the paket.lock file
|
# ? Jan 1, 2016 20:05 |
|
AWWNAW posted:what's the typical lifespan of a web "app" these days anyway? 1 year? less than 6 months or longer than 5 years. theres nothing inbetween
|
# ? Jan 1, 2016 20:12 |
|
Bloody posted:things linux developers dont use.txt GPU is terrible, i'll concede that, but sound's been pretty fine for a while now unless you have some bitchin' $200 sound card or something. also i have a wacom intuos that i originally got for drawing cute anime boys gently caress you don't judge me that works fine* on linux *if you don't mind that 95% of the settings can only be accessed via a command line utility and it resets everything to defaults every time you unplug it
|
# ? Jan 1, 2016 20:13 |
|
the test for sound is if you can get 2 midi devices to play samples at the same time with sub 50ms lag windows still cannot unless you buy ASIO supported hardware
|
# ? Jan 1, 2016 20:18 |
|
the people who say "just werks" are usually people who know little enough about computers to say it completely sincerely and without knowledge of the unironic version.
|
# ? Jan 1, 2016 20:21 |
|
MeruFM posted:the test for sound is if you can get 2 midi devices to play samples at the same time with sub 50ms lag like physical midi devices? because i've gotten stacks of software synths + my midi keyboard working fine with pretty low latency before but it's not like i actually use that poo poo professionally or anything.
|
# ? Jan 1, 2016 20:34 |
|
ffs guyssuffix posted:i assume they're at least recorded in the paket.lock file yes, the paket.lock file is autogenerated and it will include all version numbers / git commits, plus transitive dependencies and such. if you don't specify any versioning in the dependencies file, when you *manually choose* "Update packages" the paket.lock references will update to the latest version. you normally DO specify versioning in the paket.dependencies file, though. this is how: https://fsprojects.github.io/Paket/dependencies-file.html posted:The file specifies that Paket's NuGet dependencies should be downloaded from nuget.org and that we need:
|
# ? Jan 1, 2016 20:38 |
|
NihilCredo posted:ffs guys oh wow this poo poo is soooooo much better than node
|
# ? Jan 1, 2016 20:50 |
|
Power Ambient posted:oh wow this poo poo is soooooo much better than node i'm the special snowflake tilde arrows
|
# ? Jan 1, 2016 20:52 |
|
Parallel Paraplegic posted:i'm the special snowflake tilde arrows lots of things (like gems/bundle in rails) do that? maybe node too?
|
# ? Jan 1, 2016 22:22 |
|
|
# ? May 24, 2024 21:17 |
|
people actually give a poo poo about midi in tyol 2016?
|
# ? Jan 2, 2016 01:49 |