Hacker Newsnew | past | comments | ask | show | jobs | submit | yrmt's commentslogin

Thanks for yet another good interview!


You are welcome!


Great series of interviews, thanks to the beastie.pl team!


Thanks! Few more to come next!


NetBSD welcomes all SPARC owners


fixed that, thanks for pointing it out


I'm building fresh 64bit packages on Mac OS X Yosemite. They will be uploaded in a few days at http://pkgsrc.saveosx.org/Darwin/2014Q4/x86_64/All/. Oh and they are GPG signed. Documentation at http://saveosx.org



colordiff looks like: http://www.colordiff.org/screen2.jpg

It's just coloring the standard diff output so lines starting with > or + are yellow and ones starting with < or - are green. This is also what "git diff" does by default: http://www.jefftk.com/git-diff-color-2x.png

icdiff is doing a side-by-side diff, and it's only coloring the changed sections of lines: http://www.jefftk.com/icdiff-side-by-side-2x.png


Not all that different than what you can achieve with git diff --color-words

Example: http://i.imgur.com/smubTnH.png


git diff --color-words is good. Here's what that looks like on the same file: http://www.jefftk.com/diff-color-words-2x.png


It doesn't seem to be side by side, while icdiff is. Colors by themselved are not an issue as git supports it natively.

Have I miseed something?

edit: fix typo


diff -y is side-by-side. As far as I know colordiff does per-line coloring, while icdiff seems to have finer granularity. This dude really should have patched colordiff, though. Or just used emacs.


colordiff is just a perl script to colorize diff's output. Patching it to add support for the kind of coloring I want to do here would be a lot of work.

As for using emacs, there are many cases when I want a stand-alone diff tool.


Wow, all these years and I haven't known about diff -y. Thanks.


One more called cdiff: https://github.com/ymattw/cdiff

(works with repositories; not a general diff replacement.)


Correct me if I'm wrong, but most of the things (or maybe just the ideas?) you have implemented seems to be coming from NetBSD. If so, could you share other NetBSD stuff you'd like to see on Bitrig ?

Congratulations on the 1.0 !


For me personally, better armv7 support. Even though many things stem from our hands, NetBSD and FreeBSD are a great resource. I have already started working on SMP support, based on their example.

Otherwise I do not have any NetBSD stuff in mind. Can't speak for our other developers though. ;)

I would love to have FreeBSD's bhyve on Bitrig. I will dedicate some time in January for that goal. DragonflyBSD is an inspiration for us, too. We have an experimental branch (smpns) for revamping the kernel for decent SMP support.


A great alternative to Homebrew is pkgin because of it's speed and the amount of packages available in pkgsrc.

More info here: http://saveosx.org/


This is the work of Jonathan Perkin (from joyent): http://www.perkin.org.uk/

It's great. It really is. I never understood why it's not used by more people. Homebrew is simply awful.


Just curious, what's wrong with Homebrew? I'm just a student, not a full-time developer, but I've never had problems with it and find it pretty pleasant to use.


I use a macbook for the battery life and the no-hassle experience (which is getting worse everyday, but that's a different discussion). Homebrew works against these two principles in multiple ways. Compiling software takes massive amounts of time; if I need to install something I want it now, dammit. In principle, there are some binary packages too, but when I tried homebrew there were not for anything that I needed.

Apart from time, compiling software consumes energy, decreasing the autonomy of my laptop.

But those are not the main reasons I dislike homebrew so much. My main gripe is that the packages are of extremely poor quality. When I needed something, it was not compiled with the options I needed. E.g. packages were missing DTrace integration, QEMU was missing the one target I cared about etc.

I am a Go developer; you have no idea how many problems people have had with the homebrew package. It is garbage. And Go is quite trivial to package compared to other stuff. I simply gave up on helping anyone reporting a problem if he is using homebrew. I ask them to recompile Go from official source first.

And then there's the whole curl into bash things, gah I'll just stop now.

The biggest advantage of pkgsrc (apart from quality binary packages) is that I can use the same package versions, compiled exactly the same (to the extent possible) on OS X, Linux, and Solaris; and these are production-ready releases used in production by Joyent (and others) vetted by actual release engineers.

And with pkgsrc it is trivial to install from source when you need to modify a package locally (unlike, say, with apt).


They don't (as a policy, IIRC), include libraries for dynamic languages, even when those libraries has native dependencies.

Hence, trying to get such libraries installed with homebrew results in dependency hell, as the package manager won't handle those dependencies.


It's slow and it has to build the packages most of the time. pkgin will install a precompiled packages very quickly. It is also a whole C program and Homebrew is just ruby scripts.


> It is also a whole C program and Homebrew is just ruby scripts.

And that's a problem?

I would argue that it's a feature. Due to Homebrew being written in Ruby, it has easily allowed over 4,000 contributors [1]. Recipes are easy to read and inspect for verification purposes.

What is the advantage of C over Ruby?

[1](https://github.com/Homebrew/homebrew/graphs/contributors)


Homebrew now has prepackaged binaries ("bottles") for most of the packages that take awhile to build, which has really helped the speed thing.


Yeah, but they often rely on the homebrew being installed to /usr/local which is just plain bad advice.

At least, not using /usr/local myself, I see a lot of packages building because they have detected I changed the base directory.


So the opaque pre-compiled blob that installs other pre-compiled blobs is somehow better than a collection of scripts that lets you build from source? I don't think so... As an update is something that is done occasionally and in the background "slow" is not exactly a reason to change.


So the opaque pre-compiled blob that installs other pre-compiled blobs...

Why use OSX, then? Shouldn't you be using Gentoo?


Well the largest issue by far is the fact that it overrides system blobs. Homebrew installs into it's own sandbox area and symlinks if requested.


actually the "src" in pkgsrc means that you have the choice of using binary packages (pkg) or building from source. The best of both worlds!

Give it a try, you might like it (because it rocks!)


Homebrew does both as well for popular packages; 'bottles' are binary packages, with the option (--build-from-source) to build it yourself.


What makes C better than ruby?


I'm just guessing here but perhaps runtime speed? (In this context, that is)


I also have a repo of 64bit packages built on Mavericks: http://pkgsrc.saveosx.org/Darwin/2013Q4/x86_64/All/


Because it's relatively new and Homebrew is simply useful? Would be great if you could elaborate some, maybe I'll be the next convert :)


pkgsrc has had OS X support since Oct 2001.

Source : http://www.netbsd.org/docs/pkgsrc/pkgsrc.pdf


Holy shit. NetBSD pkgsrc for OS X? Yes please.

(On a sidenote, I got into Arch because I heard hype, but did not even like it until I recognized the BSD outlook, with ports and pkgsrc style tools. This is very cool, and now I know what to use if I deal with OSX again!)


Of course! pkgsrc runs on a lot of OS :)


Consider applying for YC's Winter 2026 batch! Applications are open till Nov 10

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: