jpstewart wrote:kubatyszko wrote:Don't encourage him that much or we never get newish firefox on IRIX

"People achieve the best results not knowing something is difficult or even impossible"...
Cheers

It's not my intention to be negative or discouraging.
I find having realistic expectations about the time required helps me immensely. If I go in thinking something will be a quick and easy task, I'll very quickly get discouraged if it's not. If I know it's going to be difficult, I'm more apt to slog through just for the sense of accomplishment that comes at the end.
I prefer the quote, "to be forewarned is to be forearmed". Better to prepared for the hard work than surprised by it!
Well, from looking at previous patches, getting Firefox to compile itself isn't the hard part (there was a successful build of 3.5); the problem is bitrot. Part of it I bet is that Firefox on MIPS seems to only really get regularly compiled for little-endian chips (Loongson mostly) and on Linux, so some of the IRIX assembler code has probably broken. My initial effort is to build libnspr, and make sure its test suite passes; this confirms issues with sockets, threads, and other low-level plumbing is Ok. Then the next issue is checking libnss (not expecting issues here), and then XPI where I expect build-related error message confetti. Once the entire thing builds, then its a matter of checking the irix assembler code in the tree, and make sure it still matches reality.
I was hoping the cray-cyber o2000 would have been online today so I could start working, but its remained off, and I have gotten no reply to an email I sent as of yet. I'll probably be waiting for at least two more weeks until I can get my hands on SGI hardware. Still hoping someone with a box is willing to give me a shell

.