>>54962615
Why does NetBSD's pkgsrc developer documentation kinda suck? I'm talking about this:

https://raw.githubusercontent.com/jsonn/pkgsrc/trunk/doc/pkgsrc.txt

When I was a new pkgsrc user, I found it overly verbose for just trying to jump in and understand how to begin writing my own pkgs or modifying existing ones to suit my needs. Some common stuff was either undocumented, or embedded in a random paragraph within it.

I barfed some of my random thoughts into a text file as I was learning and working with pkgsrc if anyone's interested. Though it's missing a bunch of other stuff that I was too lazy to write down at the time I encountered issues.

I'll paste it in the next reply since it's too long for this post.

>>54962764
2/2

These were my raw thoughts at the time. Too lazy to clean it up.

Copy-pasta powers, ACTIVATE.


pkgsrc complaints

- `make show-depends' is a pretty useful command, but isn't documented anywhere in pkgsrc.txt.
- It shows runtime dependencies by default, so I would've never assumed `make show-depends BUILD_DEPENDS', for example, would've been used to show build deps.
- I also wouldn't have guessed to run `make help topic=show-depends' because I would've thought showing runtime and buildtime deps would've been different commands.


- I don't like how `make makepatchsum' is only mentioned once in pkgsrc.txt, in some random paragraph. It's a useful, time-saving command that people new to working with pkgsrc should know about from the get-go.

- If you run `make show-depends BUILD_DEPENDS', the command runs as expected, and shows output as expected. But afterwards, Make complains BUILD_DEPENDS is an invalid command. This can be be confusing if one is working with a broken/WiP pkg that doesn't have its dependencies correctly defined and prints nothing when `bmake show-depends [DEP_TYPE]' is ran, because it just looks like a straight-up error.


## Makefiles
- *Having* to specify SUBST_STAGE feels redundant in the Substitition Framework. In the vast majority of use-cases, substition is probably done right after post-extraction, so if it's ommitted, it should be implied. Right now, if it's ommitted, substition doesn't work at all.

- The Makefile example isn't really clear about how to specify dependencies? What is this "buildlink" thing? And how come it only documents BUILD_DEPENDS, for what I presume is for build-time dependencies? Wouldn't most people want to know about *runtime* deps the vast majority of the time?

>>54962785
With this said, I love pkgsrc, and it's been very useful for me. Thank you for all the hard work pkgsrc folks!

You post this every day now looking for attention since the thread fails and isn't bumped and dies within 3 hours

Even if it does get posts, everyone here knows how full of contrarian shit you are, how you don't really give a fuck about BSD or computers at all, so that thread sucks for you too. You cannot win either way. It's time to stop posting and try somewhere else.

I'mma try and post in these threads regularly once I got OpenBSD going in a VM. good work OP

>>54962911
How do you figure, anon? What'd OP do in the past?

>>54962615
FreeBSD is best girl

>>54963555
I think the fact you never see *BSD arguments here in /g/ is probably a testament to the maturity of their userbases.

Or it could just be that nobody here actually uses *BSD...

>>54962911
are you the anon who gets so triggered by these threads that he shitposts in them

>>54963338
Did ya install OpenBSD yet, Anonafriend?

Also, this: http://www.openbsdjumpstart.org/#/

Anybody know any good BSD hardware?
Thinkpads are supposedly the "GNU/Linux PC", So what would be the best BSD hardware around?

>>54965717
Thinkpads are BSD hardware. I think the X series especially.