Dynamically typed languages aren't out in the cold like they used to be with IDE support. Most of the advanced LSP work going on isn't based on explicit types signatures yet they work well. See https://railsatscale.com/2024-07-18-mastering-ruby-code-navi...
TruffleRuby supports many C extensions like DB adapters out of the box like CRuby. JRuby has Java ports for most popular gems that ship alongside them. Nokogiri is one gem that's a Rails dependency and stands out as a bit of a pain, but both JRuby and TruffleRuby put in substantial effort to support it too. All three generally work with web servers, including Puma, which ships with Rails. There are still gaps in C extensions that TruffleRuby covers and lack of Java ports for less popular but still used gems. They may have some things like gRPC C extensions that don't work and they don't want to port them either. Generally the ecosystem just works between these popular Ruby implementations.
MRuby uses MGems and is just a different ecosystem entirely. It has parallel libraries but they're not shared with the above implementations.
So why the relative lack of popularity for JRuby/Truffle? I know about startup time and a degraded dev experience (compile/restart server etc is slower). Are these the major pain points?
Or is it simply that most Ruby shops have a good enough performance and they're not looking for anything else?
Very slow startup and the compatibility isn't 100%. You don't get the latest Ruby features and every once in a while, code that would work on CRuby doesn't work out of the box. It's kind of a similar situation with PyPy as far as I can tell. It's been around for a long time but not much adoption.
Not to nitpick, but a quick check shows Square has way more stars overall. (233,343 stars for Square repos and 30,982 stars for Stripe repos.) If you’re talking about Square’s brand new SDKs, they of course start with zero stars and some are just weeks old.
Wow, looks like Square is updating its github repository in response to this thread! Just saw they pinned their SDKs. Good job Square, that helps.
Square if you're listening, also work on the docs and dev experience. For example I noticed that Stripe's docs immediately show you how to work with charges. Whereas with Square the first step is... setting a location. It's just another little dev experience hurdle. What if I don't have a location? What if I'm an online store that spans all my locations? Stripe has put a lot of attention to detail here. Pinning repos is just the tip of the iceberg.
This is important because developer mindshare matters. For instance it seems like lots of devs here think this stripe capital product is novel/brilliant, a testament to wunderkind founders, etc. etc. That's all Stripe doing an A+ job grabbing and holding developer mindshare. GET ON IT SQUARE.
P.S. I think my characterization about the SDK stars is correct. They have some other open source projects that are very popular, but their SDK is trailing way behind Stripe by this measure (even years old repos), by like a couple orders of magnitude.
We are most definitely listening. We appreciate the feedback and are always looking to improve our developer experience. Stripe has done an excellent job in that regard.
If you would be open to it, we'd really like to chat with you on your thoughts about our SDKs & documentation. Feel free to shoot an email to moot@squareup.com or message our Twitter @SquareDev.
Under the NSA's definition of 'foreign', a 51% likeliness of foreignness (a coin flip plus 1%), they only target foreigners. Use tor, there is the 1%. Speak a language other than English, gosh that must be worth 10%.
This goes on word after word. The NSA statements are blatantly false. They can only even pretend it is true by twisting words so far they no longer resemble their actual meaning.
OpenOffice isn't the OpenOffice you used to know (just a fork of LibreOffice with Oracle's TMed name). If I recall, Oracle just handed it off to the Apache Foundation after losing developer support and being dropped by prominent linux distros. Oracle called LibreOffice a fork, but that is nonsense since it is the original maintainers and code with only a new name. I'm actually curious why Apache Foundation took on OpenOffice following Oracle's misteps and what the rationale is for its existence alongside LibreOffice. (There is one, right?) To be fair, I haven't used them side by side or followed OpenOffice since the Oracle debacle.
TL;DR: Oracle forked OpenOffice and kept the name. Long live LibreOffice!
It doesn't matter which people do or do not work on the project at this point. The original owners of OpenOffice donated the project to the Apache foundation (under less restrictive licensing terms I might add).
So at this point, it doesn't matter what happened to the project in the past, OpenOffice is now an Apache project owned by the Apache foundation.
TL;DR: The OpenOffice project (now part of the Apache Software Foundation) announced the second release candidate for the soon to be released OpenOffice 4.0.
> The accepted definition of a software fork is a copy of
> an existing project that is independently developed.
> ...
> It doesn't matter which people do or do not work on the
> project at this point
Surely if the definition is "independently developed", then which people are developing it is the only thing that matters?
The independently developed (in my mind) is not referring to the people developing it, but rather the fact that the new copy is developed independently of the old project.
As an example, you can have the same set of developers on a new project, but as long as it's a copy of an old project, most people still expect that to be called a fork since the development of that copy is independent of the original.
From what I can tell, OpenOffice is now the successor to IBM's Lotus Symphony (which was previously an OpenOffice fork). IBM donated their code to Apache, so anyone coming from that software can maintain UI elements that LibreOffice does not have.
OpenOffice is now owned by the Apache Software Foundation; don't you think the Apache Software Foundation wanted to be able to release under the Apache License?
Complaining about the license being incompatible is a bit silly; the choice of license by the Apache Software Foundation should not be a surprise.
It's not really incompatible with LibreOffice. The difference is that you OO.o can't take code from LibreOffice, but code from OO.o can find it's way into LO without any issues.
IMO, the main problem for LibreOffice is not OpenOffice but compatibility with the industry standard. Like it or not, you need to reasonably read and write the MS document formats.
I'm questioning it. The behavior in question seems alcohol-spawned. If alcohol is tolerated it seems bizarre to treat cannabis draconianly.
If cannabis use is legal in the state, there is no prosecutable crime. Federal cannabis drug law is not enforced against cannabis users, only those who sell it and thereby more directly touch on interstate commerce.
If breaking unenforced federal law is ban-able behavior, then better ban everyone.
I think the unfortunate thing is that we all break federal law every day. (Most of us at least violate the Computer Fraud and Abuse Act. Ever made a GET request potentially in violation of an API's ToS?) Scary thing is, the feds actually prosecute under the CFAA.
