[defaultrouteuk]

Good Day to you all,

I’ve always been more of an ‘ops’ than a ‘dev’. I dabble with scripting and used to be fairly ok in a shell. I was a *nix admin for a few years but as you can imagine, working with binaries and distributions does not make a developer.

Anyway, I never really got into software development but I believe I wish to try. With all the hype surrounding my daily grind with ‘python this’ and ‘automate yourself out of a job’ that it’s fair to say I’m concerned that my lack of focus on software development has left me rather exposed to obsolescence.

My passion for SGI and love of IRIX however has given me an opportunity to do two things I hope to enjoy. First I’d like to learn C/C++ to a degree of competence where I can be proud. Secondly I’d like to try to help any developers porting stuff for IRIX. Of course the later is aspirational but that’s what aspirations are for right?!

I therefore throw myself to the wolves and ask kindly and without pride if there is anyone here who would assist my path in some way. I’ve started doing a few MOOC courses and I’m not averse to study. If there are student guides you’ve had or used in the past, particularly related to IRIX, then I’d love to have access to those.

If anyone has advice or can point me in the right places I’d appreciate any help on the road. If there is any grunt work I can help with, pass it on.

Best regards
Rich

[Raion]

Hey Rich!

So there's a few areas we should cover ground with when it comes to programming on IRIX:

1. Knowing C/C++

C and C++ concepts are moreorless universal. I can give you a PDF of Zed A Shaw's "Learn C the Hard Way" to get into coding more. That'll cover many foundationals of C programming itself.

2. Programming using a modern toolchain

As much as I advocate MIPSPro, most tutorials, including Zed's, assume you're using a modern C compiler like LLVM or GCC. Thus, it's often advantageous to learn on a more modern toolchain. IRIX does not have LLVM, but there's two "GCC" moreorless:

GCC "Official" which ended at 4.7.4, almost 10 years ago
GCC "SGUG" which is a fork of GCC that patched IRIX support back in.

The SGUG port of GCC is pretty competent, you can get GCC 8 or 9 easily. didbs is probably the easiest way to do that, but forewarning, they rely on patched system headers and stuff and that is all fine and dandy, but it can lock you into that system. It's just a different approach, though.

3. Programming using the native toolchain.

MIPSPro supports C99 and C++03, and it's a relatively good compiler that will force you to learn very good C programming practices, though the code of IRIX isn't often super portable. It does have some MIPSProisms, just like any other compiler does.

If you want to start here, I suggest three sources:

IRIX Wiki (https://wiki.irixnet.org) I got some good info here.
Techpubs (http://techpubs.jurassic.nl) There's tons of programming manuals here
O'Reilly Practical C Programming, 3rd Edition or another older C programming manual.

[defaultrouteuk]

Thanks Raion,

Your feedback is appreciated. I have access to a bunch of texts (including your recommendation of Shaw) but I struggle to 'create' applications if that makes sense. I understand it's a common thing to deal with and I need to challenge myself to find a task I can attempt without frustrating myself unnecessarily. I'll take a look at this links. The Techpubs one I think on the face of it is more where I want to be. The GCC/GNU dev tools themselves lend a lot to portability but as you mentioned the MIPSPro support and the 'learn very good C programming practices' is where I'd like to be aiming. I'm particularly interested in understanding network/socket development.

I went to a Solaris Network Programming course once in London (20-something years ago). I rocked up and was literally the only person who was *not* a developer. So I basically sat there wondering what was going on for 5 days. There wasn't an easy entry point and I got frustrated and shelved the whole idea of trying again. So nowadays in my field (networking) there is so much python with their 'easy win' libraries that I feel it's lost a lot of the gleam and kudos associated with the feeling of understanding the underlay. I'm not trying to say I know python either, I just don't. But I see a lot of publications pushing the narrative of automation. Anyway, I digressed.

I'll take a look - thank you again!
Rich