I'm done with MS and their .NET Core Bullshit
-
@unperverted-vixen said in I'm done with MS and their .NET Core Bullshit:
Asides from being a slow, Java-based pile of crap.
I've not noticed any particular problems with that, but the application is very resource hungry (and that's probably to do with DB caching; that's ever so easy to screw up in ways that don't show up on a beefy machine with a small testing DB).
Right now, I'm using github issues for this sort of thing and they're just a bit too freeform (they're better for managing the pull request flow, but are still imperfect and labels aren't enough with my cow-orkers). We don't do time tracking at all with issues, but that's OK as our metrics aren't tied to that sort of thing in the first place.
-
@stillwater said in I'm done with MS and their .NET Core Bullshit:
The more the magic, the difficult it is to use
Exception: mockito.
-
@stillwater said in I'm done with MS and their .NET Core Bullshit:
another whole bag of WTFs
It always is.
-
@magnusmaster said in I'm done with MS and their .NET Core Bullshit:
jwt tokens
Are those like ATM machines?
-
@dkf said in I'm done with MS and their .NET Core Bullshit:
@unperverted-vixen said in I'm done with MS and their .NET Core Bullshit:
Asides from being a slow, Java-based pile of crap.
I've not noticed any particular problems with that, but the application is very resource hungry (and that's probably to do with DB caching; that's ever so easy to screw up in ways that don't show up on a beefy machine with a small testing DB).
I'll admit that my experience is very much with IIS/.NET, and not Java, so I'm mostly spitting in the wind trying to get and Confluence to suck less. But for some reason, throwing more resources at them doesn't seem to help past a certain point. DB profiling doesn't show anything obvious either.
-
@gąska said in I'm done with MS and their .NET Core Bullshit:
@unperverted-vixen said in I'm done with MS and their .NET Core Bullshit:
@dkf said in I'm done with MS and their .NET Core Bullshit:
But for just tracking the actual state of issues themselves, JIRA isn't too bad; it's better than many of the alternatives that I've tried.
Asides from being a slow, Java-based pile of crap.
Also. Who the fuck made the decision to set time tracking precision to SECONDS!?
It's even better: the day is in work hours; so if you want to represent two days you don't put 24 hours of seconds, you put 2*workhours seconds. Intuitive!
-
@gribnit said in I'm done with MS and their .NET Core Bullshit:
@stillwater said in I'm done with MS and their .NET Core Bullshit:
The more the magic, the difficult it is to use
Exception: mockito.
Never heard of it. Let's take a look at homepage!
http://site.mockito.org/
Um... Who the fuck uses "site" subdomain for their homepage? And "mockito.org" actually redirects to it!
But enough about the site. Let's look at the very first sample they show:
// using mock object - it does not throw any "unexpected interaction" exception mockedList.add("one"); mockedList.clear();
Oh, great. Silent failures in unit tests. Well, at least they added the strict testing option in December 2016 after 8+ years of being stable. The docs are very confusing in how to use it, but it's there!
// selective, explicit, highly readable verification verify(mockedList).add("one"); verify(mockedList).clear();
So I have to check if the calls worked manually? And I have to repeat exact invocation I've made when declaring expected interactions? Talk about automation.
// the following prints "null" because get(999) was not stubbed System.out.println(mockedList.get(999));
Oh, so it doesn't just silently fail - it also returns garbage data that's either completely unexpected by the caller, or worse, expected and handled gracefully! What the fucking fuck. Isn't it supposed to make testing easier, not harder?
-
@gąska Like any bean, this works better planted than jammed in the ear.
-
@gribnit you mean buried in ground? Yeah, I agree.
-
@unperverted-vixen said in I'm done with MS and their .NET Core Bullshit:
I'll admit that my experience is very much with IIS/.NET, and not Java, so I'm mostly spitting in the wind trying to get and Confluence to suck less.
Forget it. Confluence sucks pretty hard. I don't think there are any good general replacements, but there are definitely options that are better. Weirdly, JIRA is one of the better general issue trackers despite being built on top of the same pile of but that might reflect just how terrible so many issue trackers really are…
-
@dkf I think one of JIRA's main problems is that it's too customizable. Because it's always customized badly.
-
@gribnit said in I'm done with MS and their .NET Core Bullshit:
Like any bean
Ha. 'Cause it's Java. I get it.
-
@pie_flavor I had kinda dreaded that that would happen, but no, not really, that was accidental. I should have maybe said, like any crayon, this works better used to draw stick figures than inserted nasally.
-
@tsaukpaetra said in I'm done with MS and their .NET Core Bullshit:
@gąska said in I'm done with MS and their .NET Core Bullshit:
@unperverted-vixen said in I'm done with MS and their .NET Core Bullshit:
@dkf said in I'm done with MS and their .NET Core Bullshit:
But for just tracking the actual state of issues themselves, JIRA isn't too bad; it's better than many of the alternatives that I've tried.
Asides from being a slow, Java-based pile of crap.
Also. Who the fuck made the decision to set time tracking precision to SECONDS!?
It's even better: the day is in work hours; so if you want to represent two days you don't put 24 hours of seconds, you put 2*workhours seconds. Intuitive!
Huh? I just enter
2d
and It Works™ ?EDIT: Damn, now you making me doubt myself. I'll have to go and check again, but I thought it turns 60s into an hour (duh), 8 hours into a day and 5 days into a week.
-
@jbert said in I'm done with MS and their .NET Core Bullshit:
@tsaukpaetra said in I'm done with MS and their .NET Core Bullshit:
@gąska said in I'm done with MS and their .NET Core Bullshit:
@unperverted-vixen said in I'm done with MS and their .NET Core Bullshit:
@dkf said in I'm done with MS and their .NET Core Bullshit:
But for just tracking the actual state of issues themselves, JIRA isn't too bad; it's better than many of the alternatives that I've tried.
Asides from being a slow, Java-based pile of crap.
Also. Who the fuck made the decision to set time tracking precision to SECONDS!?
It's even better: the day is in work hours; so if you want to represent two days you don't put 24 hours of seconds, you put 2*workhours seconds. Intuitive!
Huh? I just enter
2d
and It Works™ ?EDIT: Damn, now you making me doubt myself. I'll have to go and check again, but I thought it turns 60s into an hour (duh), 8 hours into a day and 5 days into a week.
I was talking about the API, but yes eight hours per day. Which makes weirdness happen unless you know about that, because of course the API doesn't return their pretty format (if memory serves), only seconds, so on order for my App to display it the same way JIRA does I have to replicate that calculation.
-
@boomzilla said in I'm done with MS and their .NET Core Bullshit:
That was just a joke. @cartman82 posted a thread with ".Net is dying" or similar at one point.
However, I believe that neither .Net nor Java are dying.They're declining slowly on PYPL
-
@stillwater said in I'm done with MS and their .NET Core Bullshit:
@pie_flavor said in I'm done with MS and their .NET Core Bullshit:
@stillwater You mean like F#? I hear it runs on .NET Core!
No I mean like functional programming in python or java or C#. Not a FP language.i tried using Haskell once and tried writing hello world and ended up getting my dick stuck in a fan or however that meme goes. No pure FP for me.
If you want to do functional programming on the side on the JVM I'd recommend Kotlin. It's not a pure FP language, or a pure anything language really, but it picks the good parts that can be squashed together and still work nicely...ish and does it in a fairly coherent way.
Java just hurts when you try to do functional programming stuff in it.
-
@gąska said in I'm done with MS and their .NET Core Bullshit:
@gribnit said in I'm done with MS and their .NET Core Bullshit:
@stillwater said in I'm done with MS and their .NET Core Bullshit:
The more the magic, the difficult it is to use
Exception: mockito.
Never heard of it. Let's take a look at homepage!
http://site.mockito.org/
Um... Who the fuck uses "site" subdomain for their homepage? And "mockito.org" actually redirects to it!
But enough about the site. Let's look at the very first sample they show:
// using mock object - it does not throw any "unexpected interaction" exception mockedList.add("one"); mockedList.clear();
Oh, great. Silent failures in unit tests. Well, at least they added the strict testing option in December 2016 after 8+ years of being stable. The docs are very confusing in how to use it, but it's there!
// selective, explicit, highly readable verification verify(mockedList).add("one"); verify(mockedList).clear();
So I have to check if the calls worked manually? And I have to repeat exact invocation I've made when declaring expected interactions? Talk about automation.
// the following prints "null" because get(999) was not stubbed System.out.println(mockedList.get(999));
Oh, so it doesn't just silently fail - it also returns garbage data that's either completely unexpected by the caller, or worse, expected and handled gracefully! What the fucking fuck. Isn't it supposed to make testing easier, not harder?
Mockito requires you to orchestrate what should happen when the mock is used. It always returns null for things you have not told it what to do with, or 0 or other similar emptyish values. This usually results in the test generating NPEs, or if the code prudently checks for nulls, failure codes or specific exceptions.
I wouldn't say it's magically easy to use or even pretty, but it's a lot easier than rolling your own reflection based decoupling black magicks. And it works, and does so in a repeatable way. As far as the Java ecosystem goes, Mockito is not half bad.
-
@carnage said in I'm done with MS and their .NET Core Bullshit:
@gąska said in I'm done with MS and their .NET Core Bullshit:
@gribnit said in I'm done with MS and their .NET Core Bullshit:
@stillwater said in I'm done with MS and their .NET Core Bullshit:
The more the magic, the difficult it is to use
Exception: mockito.
Never heard of it. Let's take a look at homepage!
http://site.mockito.org/
Um... Who the fuck uses "site" subdomain for their homepage? And "mockito.org" actually redirects to it!
But enough about the site. Let's look at the very first sample they show:
// using mock object - it does not throw any "unexpected interaction" exception mockedList.add("one"); mockedList.clear();
Oh, great. Silent failures in unit tests. Well, at least they added the strict testing option in December 2016 after 8+ years of being stable. The docs are very confusing in how to use it, but it's there!
// selective, explicit, highly readable verification verify(mockedList).add("one"); verify(mockedList).clear();
So I have to check if the calls worked manually? And I have to repeat exact invocation I've made when declaring expected interactions? Talk about automation.
// the following prints "null" because get(999) was not stubbed System.out.println(mockedList.get(999));
Oh, so it doesn't just silently fail - it also returns garbage data that's either completely unexpected by the caller, or worse, expected and handled gracefully! What the fucking fuck. Isn't it supposed to make testing easier, not harder?
Mockito requires you to orchestrate what should happen when the mock is used.
Like every other mocking library. But what every other mocking library doesn't do is fail silently.
It always returns null for things you have not told it what to do with
It could have thrown exception, but nooooo, let's waste days of development time on tracing nulls in unit tests!
or 0 or other similar emptyish values.
Which is even worse, because while null usually never happens, 0 and friends usually do happen, and happen quite often. Which makes silent failure so much harder to detect and debug.
This usually results in the test generating NPEs or if the code prudently checks for nulls, failure codes or specific exceptions.
Or some non-failure non-exception behavior when null is within the range of valid values. Exception has 100% guarantee to stop the test right there and then, with full information of what happened and where. Null has a low chance of never getting reported and letting the test being marked as passed, and a higher chance of being reported in some random place down the line with a random call stack.
I wouldn't say it's magically easy to use or even pretty, but it's a lot easier than rolling your own reflection based decoupling black magicks.
But compared to other available libraries, it looks like a terrible choice.
And it works, and does so in a repeatable way.
Veeeeery repeatable, if you catch my drift. And while it technically works, it doesn't look like it's doing what I expect from a mocking library.
As far as the Java ecosystem goes, Mockito is not half bad.
Is the thought that it could have been so much worse than it already is, supposed to put me at ease? If so, it's not working.
-
@gąska said in I'm done with MS and their .NET Core Bullshit:
Um... Who the fuck uses "site" subdomain for their homepage? And "mockito.org" actually redirects to it!
It's like in the 90s when companies started using domains like "companyname-online.com".
Like of course it's online. All websites are online. Don't be redundant.
-
@jbert said in I'm done with MS and their .NET Core Bullshit:
@tsaukpaetra said in I'm done with MS and their .NET Core Bullshit:
@gąska said in I'm done with MS and their .NET Core Bullshit:
@unperverted-vixen said in I'm done with MS and their .NET Core Bullshit:
@dkf said in I'm done with MS and their .NET Core Bullshit:
But for just tracking the actual state of issues themselves, JIRA isn't too bad; it's better than many of the alternatives that I've tried.
Asides from being a slow, Java-based pile of crap.
Also. Who the fuck made the decision to set time tracking precision to SECONDS!?
It's even better: the day is in work hours; so if you want to represent two days you don't put 24 hours of seconds, you put 2*workhours seconds. Intuitive!
Huh? I just enter
2d
and It Works™ ?EDIT: Damn, now you making me doubt myself. I'll have to go and check again, but I thought it turns 60s into an hour (duh), 8 hours into a day and 5 days into a week.
I vaguely recall that the default time unit (in the UI entry boxes) is seconds, but it's admin-configurable and I got our admin to change it to either hours or days. They've stopped making us do time tracking though, so it's a vague memory.
-
@carnage said in I'm done with MS and their .NET Core Bullshit:
Java just hurts when you try to do functional programming stuff in it
It feels fine, I say, festooned with lambdas.
-
@carnage said in I'm done with MS and their .NET Core Bullshit:
Java just hurts
when you try to do functional programming stuff in it.FTFY
-
Alrighty went from Idea to a working website with authentication, forms and a DB on the cloud in about 8 hours using Django with VS Code.
Off the bat glaringly obvious things about devloping an entire application from scratch on Django
- There is far less magic than ASP .NET Core 2.1
- The authentication system seems far easier than I expected it to be.
- The Admin Dashboard is a huge fucking help.
- The extensible template mechanism is pretty cool, not neccessarily better or worse than ASP .NET
The things I do have problems with are
- No static typing. I feel so dirty right now doing this for a whole day I feel like taking a shower with sandpaper for soap
- The conventions for putting templates in the right places is sometimes trippy when you call a template from another that's not in the same folder.
- The absolute lack of static typing is a huge PITA when there are typos in your model properties in the html template and also if you forget 'return' somewhere, the system does not alert you and I ended up debugging this for a good hour before I found out. Maybe I'm TRWTF.
Maybe few of these things get better with practice and with the right editor plugins. All things equal, this was a pretty okay experience for building a website quickly.
8/10 would Django again.
-
@stillwater if I wanted just a website, I think I'd go with Wordpress.
Python is an absolute PITA to host, unlike PHP.
-
-
@gąska said in I'm done with MS and their .NET Core Bullshit:
@carnage said in I'm done with MS and their .NET Core Bullshit:
@gąska said in I'm done with MS and their .NET Core Bullshit:
@gribnit said in I'm done with MS and their .NET Core Bullshit:
@stillwater said in I'm done with MS and their .NET Core Bullshit:
The more the magic, the difficult it is to use
Exception: mockito.
Never heard of it. Let's take a look at homepage!
http://site.mockito.org/
Um... Who the fuck uses "site" subdomain for their homepage? And "mockito.org" actually redirects to it!
But enough about the site. Let's look at the very first sample they show:
// using mock object - it does not throw any "unexpected interaction" exception mockedList.add("one"); mockedList.clear();
Oh, great. Silent failures in unit tests. Well, at least they added the strict testing option in December 2016 after 8+ years of being stable. The docs are very confusing in how to use it, but it's there!
// selective, explicit, highly readable verification verify(mockedList).add("one"); verify(mockedList).clear();
So I have to check if the calls worked manually? And I have to repeat exact invocation I've made when declaring expected interactions? Talk about automation.
// the following prints "null" because get(999) was not stubbed System.out.println(mockedList.get(999));
Oh, so it doesn't just silently fail - it also returns garbage data that's either completely unexpected by the caller, or worse, expected and handled gracefully! What the fucking fuck. Isn't it supposed to make testing easier, not harder?
Mockito requires you to orchestrate what should happen when the mock is used.
Like every other mocking library. But what every other mocking library doesn't do is fail silently.
It always returns null for things you have not told it what to do with
It could have thrown exception, but nooooo, let's waste days of development time on tracing nulls in unit tests!
or 0 or other similar emptyish values.
Which is even worse, because while null usually never happens, 0 and friends usually do happen, and happen quite often. Which makes silent failure so much harder to detect and debug.
This usually results in the test generating NPEs or if the code prudently checks for nulls, failure codes or specific exceptions.
Or some non-failure non-exception behavior when null is within the range of valid values. Exception has 100% guarantee to stop the test right there and then, with full information of what happened and where. Null has a low chance of never getting reported and letting the test being marked as passed, and a higher chance of being reported in some random place down the line with a random call stack.
I wouldn't say it's magically easy to use or even pretty, but it's a lot easier than rolling your own reflection based decoupling black magicks.
But compared to other available libraries, it looks like a terrible choice.
And it works, and does so in a repeatable way.
Veeeeery repeatable, if you catch my drift. And while it technically works, it doesn't look like it's doing what I expect from a mocking library.
As far as the Java ecosystem goes, Mockito is not half bad.
Is the thought that it could have been so much worse than it already is, supposed to put me at ease? If so, it's not working.
Since my general thoughts of null is that they should be abolished altogether from any code since they dirty things every bloody time, I agree wholeheartedly. And for the primitives, it really shows how silly it is to return emptyish values. But, it's still better than handrolling your own mocking framework, which I've done once in my silly youth.
I have others I prefer over Mockito, but for some odd reason, every gig I do seems to have picked Mockito as teh One True Mocking Framework, so I guess I've just grown used to it. But I really do agree with every single point you've made. Possibly not the days, but I'm pretty sure you're hyperboling it out of the park there.
-
@cartman82 Azure app service, AWS Lightsail and the likes solve the hosting part of any web app these days
-
@carnage said in I'm done with MS and their .NET Core Bullshit:
But, it's still better than handrolling your own mocking framework, which I've done once in my silly youth.
But is it better than not using any framework at all and just doing the stubs yourself? It's not a lot of work, and there's zero magic involved, and it works exactly like you want to.
-
@anonymous234 said in I'm done with MS and their .NET Core Bullshit:
@gąska said in I'm done with MS and their .NET Core Bullshit:
It's like in the 90s when companies started using domains like "companyname-online.com".
Like of course it's online. All websites are online. Don't be redundant.We had to, someone else had already taken <tla>.com.au. And <unabbreviatedcompanyname>.com.au would have been way too long. So <tla>online.com.au it is.
-
@scarlet_manuka what about <tla>.com or <tla>.au?
-
@gąska said in I'm done with MS and their .NET Core Bullshit:
@carnage said in I'm done with MS and their .NET Core Bullshit:
But, it's still better than handrolling your own mocking framework, which I've done once in my silly youth.
But is it better than not using any framework at all and just doing the stubs yourself? It's not a lot of work, and there's zero magic involved, and it works exactly like you want to.
Yeah, when you don't need to hop through dependency injection hoops to put the stubs in there. then you need to hack up a spiffy reflection library of some sort to do the magicky bits. Not that that part is hard either, it just gets real ugly, real fast. Better to hide it in a framework or library, and then you might as well use one someone else did, that has all (well, most) of the bugs worked out.
Of course, it's even better to have a class design where you can feed the dependencies through a constructor.
I am in agreement with pretty much everything you say today it seems.
-
@gąska Not sure if <tla>.au would fly even now. Certainly it wouldn't have back in the '90s. The .au domain was rather rigidly controlled back then, from what I heard at the time. Looks like <tla>.com was also taken by the time this company was started.
-
@stillwater said in I'm done with MS and their .NET Core Bullshit:
@cartman82 Azure app service, AWS Lightsail and the likes solve the hosting part of any web app these days
Yeah, PAAS made it possible to deploy app-based sites just as easily as file-based, I guess. I personally still prefer to stick to the VPS level of abstraction.
-
@scarlet_manuka said in I'm done with MS and their .NET Core Bullshit:
The .au domain was rather rigidly controlled back then
Oh. <tla>.tk then.
Filed under: I wonder if that's still a thing as I remember it
-
@gąska Can you as easily say "at no time do my fingers leave my hands" if you've handrolled your stubs?
-
@gribnit I can easily say fuck you.
-
@heterodox said in I'm done with MS and their .NET Core Bullshit:
Filed under: I wonder if that's still a thing as I remember it
Yes.
-
@tsaukpaetra said in I'm done with MS and their .NET Core Bullshit:
@heterodox said in I'm done with MS and their .NET Core Bullshit:
Filed under: I wonder if that's still a thing as I remember it
Yes.
It's an annoying domain to work with. The main registrar isn't particularly competent, and there's a lot of anti-spam web portals that just block the whole TLD.
-
@stillwater said in I'm done with MS and their .NET Core Bullshit:
Are you using .NET Core at work?!
I am, for important infrastructure at a major, international organization. It's working quite well.
-
@masonwheeler said in I'm done with MS and their .NET Core Bullshit:
@stillwater said in I'm done with MS and their .NET Core Bullshit:
Are you using .NET Core at work?!
I am, for important infrastructure at a major, international organization. It's working quite well.
If you don't mind me asking. Are you on .NET Core 1.x or 2.x? If it's the latter how do you manage updates and patches without messing up things?
-
@stillwater said in I'm done with MS and their .NET Core Bullshit:
This is the last fucking straw. .NET Core 1.1 to 2.0 was forgivable. Fucking around with the documentation like this is just wasting my fucking time.
Anytime-- literally anytime-- a software company decides to "reboot" a product, it is guaranteed to be complete utter shit. The only reason it's done is because some small-dicked Chief Sociopath Officer gets into power, and since he has absolutely no talent or ability to do anything except fail upwards, he goes about tearing down everything that came before him JUST so he can make his own shit-stained mark.
Why is there no .Net 4.x ? Because "OMG CORE IS CORE!"
What about all the meticulous backwards compatibility and forwards compatibility the .Net framework had? Fuck you, CORE OMGZ!
What about the decades of documentation, FAQs, examples and forum posts people could refer to? LOL CORE IS NEW OLD IS BAD! I JIZZ EVERYWHERE WITH POO@!!
What about the millions of hours of learning that went into making the framework-- and all the hundreds of thousands of "lessons learned" and "roadblocks discovered" the .Net team went through? NEW ENW NEW JUST MAKE IT IT IS MIIIIINE!
but hey... "[in 2019] With .NET Core 3 the framework will get support for development of Desktop, Artificial Intelligence/Machine Learning and IoT apps."
ITS AMAZING BECAUSE IoT IS TEH BESY WHAT EVERYONE WANTS BUZZ BUZZ BUZZ WHERE IS MU BONUS?!??
-
@lorne-kates said in I'm done with MS and their .NET Core Bullshit:
fail upwards
TIL that can be a thing.
@lorne-kates said in I'm done with MS and their .NET Core Bullshit:
I JIZZ EVERYWHERE WITH POO@!!
@lorne-kates said in I'm done with MS and their .NET Core Bullshit:
What about the millions of hours of learning that went into making the framework-- and all the hundreds of thousands of "lessons learned" and "roadblocks discovered" the .Net team went through? NEW ENW NEW JUST MAKE IT IT IS MIIIIINE!
BUT IS THE CROSSPLATFORMZ ON LINUX AND MAC .NET DEVELOPMENT EMBRACE OPEN SOURCE MAKE LOVE WITH IT AND FUCK IT IN THE ASS
All these years the only thing that was decent about MS was how dev-friendly the .NET/C# experience was and now they embraced open source and fucked the only thing that was decent about MS and fucked it in its tiny little asshole and now gotta use other open source shit that has stable documentation and works. Such a fucking shame. SMHFACEPALM
-
@lorne-kates said in I'm done with MS and their .NET Core Bullshit:
What about the millions of hours of learning that went into making the framework-- and all the hundreds of thousands of "lessons learned" and "roadblocks discovered" the .Net team went through
I reaaaaallly hope they don't throw most of it away.
-
@lorne-kates said in I'm done with MS and their .NET Core Bullshit:
Anytime-- literally anytime-- a software company decides to "reboot" a product, it is guaranteed to be complete utter shit. The only reason it's done is because some small-dicked Chief Sociopath Officer gets into power, and since he has absolutely no talent or ability to do anything except fail upwards, he goes about tearing down everything that came before him JUST so he can make his own shit-stained mark.
Alas, no. Sometimes you have to do it because you've hit the limits of what is possible because of a bad decision taken long ago (e.g., “let's write a multi-threaded CPU-intensive application in Python!”) and the only way out is a complete rewrite in something less retarded.
-
@dkf said in I'm done with MS and their .NET Core Bullshit:
let's write a multi-threaded CPU-intensive application in Python!
Writing an application with one thread in python seems like a PITA. What do you mean multithreaded?
Why is everyone okay with python? What the fuck? I get it is very readable and easy to pick up and all that crap but you gotta leave the bicycle and start using a car for a 50 mile drive at some point in your life. FUcking python everywhere disgusting.
I love how some startups say "we are disrupting the XXX industry" and in reality it is a bunch of python coder type people who don't shower piling hacks on top of each other and duct-taping the shit out of the codebase.
-
@stillwater said in I'm done with MS and their .NET Core Bullshit:
Why is everyone okay with python?
It's superficially very nice. If you're just hacking out some scripts interactively, it's easy to get going. There are some pretty decent tutorials. Some of the support libraries are very useful. A lot of code begins life as small scripts or interactive hacking, and Python does well at that.
It's only once you get into doing a larger body of work that the shittiness rears its head. Or if you need performance. Or threads. Or if you need complete, precise and unambiguous documentation.
-
@dkf A lot of tiny utilities that people wanted at work to make their life a little easier, people wrote them in python. Pretty understandable. The people who wanted very robust utilities threw in a Tests folder filled with a million unit tests. Fair enough. Easy, gets the job done, and development is pretty fast for the required purposes.
What I don't get is this, How the fuck would a 50 member team disrupting an industry with a language that is not typed, working independently with a fuck-ton of merge issues on GIt happening literally every week cos startups are fast paced and documentation is close to zero and tests are not needed cos we are disrupting everything like everything is being disrupted all the time we dont have time for unit testing bitches. ship.ship.ship. How is this a sustainable infrastructure? How long can you hack on top of hacks everytime something breaks? This is absolute madness.
Also, the hiring here in startups, the very hipster ones do not care if you've worked in large scale statically typed languages like Java or C#. You need to have expertise in any of these cool OSS languages. An older colleague who wrote huge-ass Java Integration stuff for probably one of the largest banks in the world was denied a job in a startup because they needed someone with python/django experience. And I don't know in what world is picking up django considered impossible for someone who's worked in typed large scale systems.
Fucking hispters have completely ruined everything. I don't understand when it suddenly became cool to write extremely large codebases solely in languages like python.
-
@stillwater said in I'm done with MS and their .NET Core Bullshit:
TIL that can be a thing.
@stillwater said in I'm done with MS and their .NET Core Bullshit:
I reaaaaallly hope they don't throw most of it away.
You must be fairly new to the profession. So unspoiled.
-
@carnage said in I'm done with MS and their .NET Core Bullshit:
You must be fairly new to the profession. So unspoiled.
I'm probably naive AF.