PPC NetPositive+ for PowerPC BeOS

[ClassicHasClass]

Who says you can't do TLS on BeOS ... in NetPositive itself? It's a little rocky, but it works! Downloads in the article.

It builds with Metrowerks cc/BeOS CodeWarrior, but not sure which was the more appropriate prefix ...

Going where BeOS NetPositive hasn't gone before: NetPositive+

BeOS browser : TLS apocalypse Won't keep it off line. (How do you pronounce BeOS ?) This is a real 133MHz BeBox running otherwise s...

oldvcr.blogspot.com

 

[fogWraith]

This is incredibly nice to see, thank you for the awesome work that you do!

I'll have to get this installed and give it a spin, as far as I can remember, NetPositive wasn't a very stable product to begin with ... even back then it was very crash-prone and also froze quite a lot regardless of system.

Definitely an excellent addition to the toolbox.

 

[mmu_man]

In theory it should be possible to build NetSurf, but I stopped maintaining C89 support on their codebase…

 

[ClassicHasClass]

NetPositive wasn't a very stable product to begin with

Well, the rockiness here is the TLS bolt-on -- it really strains the BeOS' stack allocations and its multithreaded nature means things can stomp on it. Stripping out the crashtastic JavaScript and a few other changes improved the core browser's stability markedly.

 

[mmu_man]

The biggest problem was the network stack, which was written in a week because it was required for a demo, and it was in userland (net_server). It was later replaced with a much better in-kernel one called BONE (BeOS Network Environment, IIRC), except it never shipped in an official release as BeOS itself was canned in favor or BeIA for Internet Appliances, shortly before Be went down. You can find BONE as a separate package for R5 (x86 only though), or in the non-official Dano0 variant.

 

[ClassicHasClass]

Oh, I ran into that when developing inetb. When threads terminated abnormally (like, I forcibly killed them) it made a mess in net_server, which inevitably needed to be restarted (but at least it could be restarted). I had to figure out a way to gracefully tell threads to go away even when they were in an indeterminate state. I can well imagine that crashing threads in Net+ would similarly unsettle it.