|
Dietrich posted:It'd be pretty strange that 99% of modern development proceeds under git and microsoft is moving to git if it weren't better than svn or tfvc. "Better" is relative. Every organization isn't Microsoft. There are plenty of businesses whose primary profit center isn't software but yet employs software developers, and those types of businesses tend to be behind the curve. Git isn't always a good fit for those organizations because they don't need the advanced features, they're comfortable with the tools they're using today, and the cost of disruption/retraining would vastly outweigh the benefits. I see this all the time. Many of them are aware of Git, and many of them are saying "No thanks, we're good."
|
# ? Oct 13, 2017 14:27 |
|
|
# ? Jun 8, 2024 06:18 |
|
Sure, there's lots of companies out there that don't give two shits about being modern or using the best tools, but I don't think that's an effective argument against being modern or using the best tools. Living paycheck to paycheck and eating fast food all the time "works" for plenty of people as well.
|
# ? Oct 13, 2017 14:40 |
|
Dietrich posted:It'd be pretty strange that 99% of modern development proceeds under git and microsoft is moving to git if it weren't better than svn or tfvc. Microsoft moved the Windows team, an operating system development team that's somewhat geographically distributed, to Git, a source control system designed for geographically distributed developers of an operating system? Huh, yeah, I can't think of why they'd do such a thing. Otherwise, Microsoft is embracing Git because it's popular and they realized (later than we'd all have liked!) that they can't just get people to use their products through sheer dint of being Microsoft.
|
# ? Oct 13, 2017 14:44 |
|
Munkeymon posted:Otherwise, Microsoft is embracing Git because it's popular and they realized (later than we'd all have liked!) that they can't just get people to use their products through sheer dint of being Microsoft. Unpack this a bit.
|
# ? Oct 13, 2017 15:04 |
|
Dietrich posted:Unpack this a bit. Am I supposed to come to the realization that Git is popular because it's the best? Hell, it's not even the least bad DVCS. I'm certainly not going to argue that TFS is better, but I would point out that TFS has features that some people see as requirements or just plain like, so I suppose you could say it has different strengths. For the record, I'm not a big fan of TFS and you can look up a rant I went on in the horrors thread about a year ago for my complaints about it, but just because Microsoft is supporting it doesn't mean it's better, just that it's popular. Same as how Linux Subsystem for Windows isn't a sign they think Ubuntu is superior to Windows. Again, it's not even the least bad Linux, it's just popular.
|
# ? Oct 13, 2017 15:31 |
|
Munkeymon posted:Am I supposed to come to the realization that Git is popular because it's the best? Hell, it's not even the least bad DVCS. When it comes to software that you didn't write and that you're going to be using to interact with other computers and people, popularity is incredibly important. If hg was the lingua franca of opensource then hg would be what we're talking about now. Basically, git is the best because it's popular.
|
# ? Oct 13, 2017 15:47 |
|
Dietrich posted:When it comes to software that you didn't write and that you're going to be using to interact with other computers and people, popularity is incredibly important. If hg was the lingua franca of opensource then hg would be what we're talking about now. Basically, git is the best because it's popular. Its popularity make it the best choice for Microsoft to throw resources behind and the best choice to learn if you've only got time to learn one VCS, yes, but those are not what most people mean when they hear "it's the best". Java is more popular than C#, so where's my VS Java integration? (I actually am curious how hard it'd be to get Java to compile to MSIL or C# to JVM but apparently not curious enough to try to write a compiler, heh)
|
# ? Oct 13, 2017 16:10 |
|
Munkeymon posted:Its popularity make it the best choice for Microsoft to throw resources behind and the best choice to learn if you've only got time to learn one VCS, yes, but those are not what most people mean when they hear "it's the best". Java is more popular than C#, so where's my VS Java integration? Your VS Java integration is being held up by Oracle being assholes.
|
# ? Oct 13, 2017 16:26 |
|
Dietrich posted:Your VS Java integration is being held up by Oracle being assholes. That... would not surprise me, actually.
|
# ? Oct 13, 2017 16:30 |
|
While I have yet to see a C# to JVM compiler yet (I'm sure it exists somewhere). I have seen a number of programs that will convert your C# code to Java. I can't imagine being crazy enough to use it in production though.
|
# ? Oct 13, 2017 16:44 |
|
The JVM to CLR transpiler was called IKVM.NET, it was reasonably popular and was only officially abandoned like a few weeks ago, so hey if you want a side project to pick up... CLR to JVM, I don't think any project went past the proof of concept stage. Unless you count Xamarin Android, that is.
|
# ? Oct 13, 2017 19:04 |
|
NihilCredo posted:The JVM to CLR transpiler was called IKVM.NET, it was reasonably popular and was only officially abandoned like a few weeks ago, so hey if you want a side project to pick up... Which compiles to Dalvik*-specific bytecode, right? *and/or whatever their new runtime is called
|
# ? Oct 13, 2017 19:11 |
|
Pilsner posted:How can you say the benefits always outweight the costs, no matter how great the costs? what i said is that when people argue for switching to git it's not due to a lack of costs, but due to benefits outweighing the costs. if you want to contrive a scenario where any change at all is absurdly costly then feel free. edit: wow, this post looks way ruder in the morning than it did at night. please mentally rephrase it when reading Gul Banana fucked around with this message at 02:08 on Oct 14, 2017 |
# ? Oct 13, 2017 19:13 |
|
NihilCredo posted:The JVM to CLR transpiler was called IKVM.NET, it was reasonably popular and was only officially abandoned like a few weeks ago, so hey if you want a side project to pick up... IKVM is still amazing to me. You throw this thing in with some JARs, and they just work side by side with your .NET assemblies with no friction. IIRC you could even navigate toy reference in VS. After giving it almost no thought, I suspect that JVM -> CLR would be easier than the other way round, since the CLR has much better support for generics.
|
# ? Oct 14, 2017 01:14 |
|
Dietrich posted:It'd be pretty strange that 99% of modern development proceeds under git and microsoft is moving to git if it weren't better than svn or tfvc. I think the same thing, but this is just an appeal to authority which sort of limits the effect. Dietrich posted:Your VS Java integration is being held up by Oracle being assholes. I don't know why you'd want/need this when IntelliJ exists.
|
# ? Oct 15, 2017 03:16 |
|
I think MS chose git because it is popular, that they needed to invent something like LFS for it seems like it might not have been the best choice for them. Same with architecting .Net Core to something like NPM, their obsession with HTTP performance for Kestrel etc. IMHO ofcourse
|
# ? Oct 15, 2017 18:31 |
|
git is actually pretty performant. LFS is a symptom of Windows being one of the largest and longest-running software projects in existence. I couldn't say whether it would have made more sense to have a version control system that's just for Windows-scale stuff.
|
# ? Oct 15, 2017 21:34 |
|
fleshweasel posted:git is actually pretty performant. LFS is a symptom of Windows being one of the largest and longest-running software projects in existence. I couldn't say whether it would have made more sense to have a version control system that's just for Windows-scale stuff. True, they probably did not do it on a lark. Maybe they took a long hard look at Git and others and decided that writing LFS would be worth it compared to other solutions.
|
# ? Oct 15, 2017 21:55 |
|
Pedant note, so people who are unfamiliar with this stuff don't get confused: Git LFS isn't a Microsoft thing, it's Large File Storage -- it handles large binary files by putting the binary in some sort of blob storage and only storing a pointer to the blob in the repo. This reduces repo bloat but allows large, uncompressable binaries to be contained in a repo alongside code. Microsoft wrote Git VFS (Virtual File System) for Windows -- it works by making it look as though all of the files are present in the repo, but only actually retrieves the file data when the file is touched.
|
# ? Oct 16, 2017 00:06 |
|
Got those mixed up....... Point stands though.
|
# ? Oct 16, 2017 05:58 |
|
When implementing user authorization/authentication in a WPF app, is the following article still considered the Right Way to do it, with IIdentity and IPrincipal? It was posted in 2014, last revised in 2016. https://social.technet.microsoft.com/wiki/contents/articles/25726.wpf-implementing-custom-authentication-and-authorization.aspx
|
# ? Oct 18, 2017 16:57 |
|
WPF and non-windowsstore desktop apps in general have not changed much in 10 years and will not do so ever, probably. So yeah, whatever you read that is a few years old is still valid for sure.
|
# ? Oct 18, 2017 20:44 |
|
LongSack posted:Thanks for the response. While I was puzzling over it this morning, I thought harder on the problem and realized that I was unable to come up with a single use case where I would actually need an "or" search. When searching for a ticket, I generally either have one concrete piece of information (say a Task or a Work Order number) or else I have a date range and possibly a Requester. My existing search handles all that as is, so I'm not gonna mess with it. If you change your mind, check out LinqKit. It does what you're after.
|
# ? Oct 19, 2017 14:24 |
|
What am I doing wrong here, it says numPlayers doesnt exist. It doesn't consider it an error when I don't make it private but I have to.
|
# ? Oct 23, 2017 04:35 |
|
underage at the vape shop posted:
code:
|
# ? Oct 23, 2017 04:39 |
|
yeah i figured that out and came back to edit, and felt dumb. I haven't done any C sharp for a while, been snowed under with C microcontroller stuff instead.
|
# ? Oct 23, 2017 04:54 |
Is there any way to tell Visual Studio to be a pedantic piece of poo poo to you about the structure of its build file. Apparently it doesn't raise as much as a warning when you're missing and EndProject marker in a .sln, but production systems aren't quite as lenient.
|
|
# ? Oct 23, 2017 17:25 |
|
I've got an app that's actually sort of a main app, and sub applications. I like to keep it together because they share a lot of things, and I like to stick with click once installs because it makes updating a breeze. Recently, I've wanted the ability to add a direct shortcut to one of the sub apps, so the end user can click an icon from the desktop and not get the full app, but one of the child windows. Worked great with command line arguments in visual studio, but then I realized that click once is sort of a piece of poo poo as far as command line arguments go. Ideally, installing the app on a pc would add it's main icon, and then I could place an icon manually for one of the sub windows. Is there an easy way to do this where the click once auto update feature would still work? As is, I'm just changing the assembly name and installing the app twice with a small change in the startup code, but it's sort of annoying to have to do that every time I want to push an update to one of the apps.
|
# ? Oct 23, 2017 22:24 |
|
Joda posted:Is there any way to tell Visual Studio to be a pedantic piece of poo poo to you about the structure of its build file. Apparently it doesn't raise as much as a warning when you're missing and EndProject marker in a .sln, but production systems aren't quite as lenient. Sort of comedy option is to create a test to read the .sln file and check for it. I've never used it for this specific case (reading and checking a solution file) but I sometimes would create a separate little folder or project of "not unit tests" to check for special snowflake things for interacting with [legacy system].
|
# ? Oct 24, 2017 08:54 |
|
Xik posted:Sort of comedy option is to create a test to read the .sln file and check for it. I think a set of sanity check tests are a good idea too. I've created tests before for example that use reflection to check that I haven't forgotten to add specific attributes to class members. I've also used T4 to when starting a new project to bootstrap all the library and exe project files with all the options, references and basic files that would otherwise be manually added. I'm thinking about adding templates to generate tests as well (again, possibly using reflection to create the test scaffolding for each of the types in the projects). It's all about having the code do the bulk of the manual, repetitive, error prone work and to ensure consistency across the solution.
|
# ? Oct 24, 2017 21:00 |
|
Is there a way to get back to a window's DataContext in a Binding from a child control? I can get back to the Window using RelativeSource, but that doesn't get me back to the DataContext. What I was trying to do us use a behaviour to handle double-clicking on an item in a ListBox, but when I bind to the attached property it doesn't work because the data context of the ListBoxItem is the ItemsSource. My Behaviour: C# code:
C# code:
C# code:
|
# ? Oct 24, 2017 23:05 |
|
I feel stupid. I'm trying to create a new ASP.NET Core 2.0 project and the only option available in my Visual Studio is targetting .NET Framework. Am I misunderstanding something here? Edit: Nevermind, I found the right Google-fu and it turns out that dropdown means nothing for .NET Core projects. Not confusing at all, cheers Microsoft... putin is a cunt fucked around with this message at 11:17 on Oct 26, 2017 |
# ? Oct 26, 2017 11:13 |
|
Okay I have another dumb question. I'm trying to integrate Auth0 into a brand new project, so no legacy code or anything I would consider likely to mess with things. I literally created the project for this purpose. I followed the Auth0 tutorial for an ASP.NET Core website to the letter, but somehow I've wound up with an error on my Home/Index action that says "Correlation failed." Googling for the exception seems to show that 90% of the time it's some kind of cookie issue, but my cookies are completely clear and I have no idea what else to even try. I can login fine to Auth0 by manually inputting my login page URL - it redirects to Auth0 and I sign in fine, and get returned to my app. But the app just shows this frustratingly vague message. Anyone ever encountered this? If it makes any difference this seems to be something that can occur with ANY Open ID Connect provider, not just Auth0.
|
# ? Oct 26, 2017 12:21 |
|
So I got help in here previously in making a model definition work when serializing XML to JSON. I have another question regarding conditional logic with the models. Say I have an XML tag that may/may not be present. Within that tag, it could be formatted in one of 8 different ways. How do I set up my models to handle this and not pass along a bunch of blank data? Really dumbed down XML example: XML code:
|
# ? Oct 30, 2017 19:20 |
|
Having a really weird issue here, that I'm not even sure how to google for or dive into to see what is going on. Have a recordset being returned from MSSQL that I'm loading into a DataTable. When I try to populate a string with one of the data fields the DataRow is showing that field is an empty array (so, db NULL I think) so it throws an error. The thing is when I run that stored procedure in SSMS the returned record most certainly has data in that field. If I wrap that output field in the stored procedure with an ISNULL(result, ' ') then the DataRow shows the data = ' ', even though in SSMS that data is not NULL in the result table. Not really sure where the error is occurring. The ISNULL check makes me think in MSSQL but there's data being returned in the result set, so maybe it's in the DataRow? Any insight or where I can start looking for a solution would be most appreciated.
|
# ? Oct 30, 2017 22:16 |
|
excidium posted:So I got help in here previously in making a model definition work when serializing XML to JSON. I have another question regarding conditional logic with the models. You don't mention which serializer you're using so I'm assuming Newtonsoft https://stackoverflow.com/questions/6507889/how-to-ignore-a-property-in-class-if-null-using-json-net
|
# ? Oct 31, 2017 16:37 |
|
Anyone know of a good CMS-type component that can be dropped into an existing MVC application? I guess I'm thinking some sort od document editor, Views for rendering them, etc, but not like a full scale application. Everything I'm finding is more like a full-on framework with its own routing, core application etc.
|
# ? Oct 31, 2017 16:41 |
|
Just-In-Timeberlake posted:Having a really weird issue here, that I'm not even sure how to google for or dive into to see what is going on. Your database driver might be setting an option that causes some operator to act differently than SSMS. Try running select @@options through both your driver and SSMS and see if they're different. SQL code:
Munkeymon fucked around with this message at 16:58 on Oct 31, 2017 |
# ? Oct 31, 2017 16:55 |
|
What exactly is the selling point behind using an ORM? I keep running across stuff like EF as the choice way to interact with data sources, but all of the selling points seem to be stuff like "never use SQL again." If I don't have any issues coming up with SQL queries on the spot are they just not for me? I've got thousands of DB tables that I could potentially be pulling and aggragating data from. What benefit is there for me to build some secondary abstraction of them to link an ORM to when I could just write a quick SQL query and get the result back as a data structure that is super simple to parse? I'm trying really hard not to be a close minded goon, but whenever I look at ORM's all I see is enough overhead that I'll probably be finished with my project before I implement the architecture needed to make the ORM function with our databases.
|
# ? Oct 31, 2017 21:17 |
|
|
# ? Jun 8, 2024 06:18 |
Two major advantages off the top of my head is the ability to multi-thread your database queries and doing data pre- and post-processing in a language that isn't SQL. Also, you can have a single server codebase, in stead of one for database queries and one for all the other stuff.
|
|
# ? Oct 31, 2017 21:28 |