Usinagaz : just connect your HP 4x to the Internet !
Some History(reverse order)
- 09/2007: Brandon Del Bel reported that it works fine on 50g
- 07/2004: Got it ! IrDA seems to work on 49g+ :D
- 10/2003: I got an 49g+, and tested Usinagaz on it. It didn't work at first (TTRM), because the interrupt handler changed a little bit. After some trickering about it, I could do everything I used to do on hp49g, but so much faster !!
- 07/2003: well, not much work in between: my compiler project took me much time... Some fixes here and there. It might work on emulators now. I also cleaned the SysRPL interface.
- 02/2003: A HPParty took place in Paris, so that I wrote an applet frame for 38/39/40 series. It should work, although not really tested yet. apparently, wget's speed actually overcame 700o/s... Wget now supports http/1.0, and Yoann's navigator uses it. How nice this is :)
- 01/2003: I rewrote buffering, and wget in assembly, one can now download http files at the terrible speed of 450B/s !
- 12/2002: I eventually took some time to rewrite Tftp in assembly, so that it is actually now very fast ! This also means that the assembly interface does work fine :)
- 10/2002: at last, got some time to write down what I did during summer... I implemented PAP (Thomas Rast also did :o), as well as dynamic DNS. Thomas Rast wrote an irc client, and I eventually wrote a SysRPL and an assembly entry. On Thomas' prompt, I shifted Usinagaz' library number to free 1772, so version 0.3 will *not* be compatible with any previous one...
- 04/2002: well, indeed, I didn't have much time... But I still hope :) I wrote everything in the TGV, I'll code it some day...
- 16-17/02/2002: another HPParty took place in Paris, so I began writing my own interrupt handler for those silly hp49s... It should work, but I don't have much time to spend on it
- 27/01/2002: at last ! it seems that 0.2 can be released : no really noticeable bug (some retransmission bug though, maybe) but much improved speed since applications can now read line by line if they want. And much improved robustness for tcp thanks to retransmission handling using a tcp window.
- 11/11/2001: an HPParty took place in Paris this week-end, so that I wrote a web server and tuned the rpl server.
- 04/11/2001: (05/11/2001 ? =) ) I'm getting back from (very) small holidays, and I see "245 unread posts" on comp.sys.hp48 ! Yeah ! Hp users are alive !... Uh oh... No, that's not the point... Oh... big troubles... took me some hours to read everything !... So ACO is dead, and HP will only make calcs, not develop them... I'm quite sad... Well ! Jonathan is right ! We must take it over, and go on developping by ourselves, as JYA & co did for years ! (I'm sad I'm too young to have seen all this being built...) So here is a more precise idea about future version 0.2. Yoann is about to release Navigator, and will begin what he calls "Internet Prime", a package of internet related software for hp ! It will begin with icq, which may better suited for hp than www =)
- October 2001: Version 0.2 is progressively appearing. Still beta, although no TTRM noticable. It now has a standard Socket system, which *must* be correctly used. It seems to be a bit quicker. It also saves some memory.
- August 2001: Rewriting the whole stuff. Chat looks quite nicer ! Version 0.12 released : better Chat.
- 28/06/2001: Version 0.11 released : No User-rpl in Usinagaz any more, some cleaning in the sources, and minor bug fixes
- 22/04/2001: Version 0.1 released, still quite beta, though
- I rewrote RESOLV in assembler and moved it into usinagaz, since it is quite essential.
- 02/2001: OK, now TCP is implemented, I wrote mail so as to easily send mails and post news from an hp4x :-)
- Great : tftp was able to download usinagaz ;-)
- Resolv also works : I resolved www.yahoo.fr !
- Chat is written ! One can now connect to a isp via a modem ! (not tested yet)
- Argnnn... I now understand why ping flood is worst : a continuous data stream locks the hp :
Used to x86 programming, I thought that my prog would see the serial buffer getting filled, and empty it in the meanwhile, without having to do some software flow control. But it locks while receiving ! I eventually hand-debugged the interrupt handler, and found a quite nasty thing, but... no other way : to avoid loosing chars, it locks the hp looping until there is nothing more to receive, or... the buffer is full ! (256 bytes). Fortunately, it supports XON/XOFF, but that's a bit sad : I think that with a 128 mru (the minimum accepted by pppd), it may have full-worked. I'll have to escape those chars.
- the third client : nettime : enable the time port in /etc/inetd.conf on the remote host, set up the timing zone, set the -4 flag if it is summer time, and voilą !
- the raw socket is done, the second client : ping !
- the first client : tftp ! It brilliantly downloaded using udp a little 64 bytes file.
- udp layer is done, so is the socket.
- ip layer is done, with icmp : Ping works ! (but ping flood, worst :-( )
- 09/2000: For now, ppp negiociation works well, messages can pass through
by Samuel Thibault