Conversation with #inferno at Mon Jun 22 12:24:48 2009 on powerman-asdf@irc.freenode.net (irc)
(12:27:28) npe [n=npe@212.224.218.143] entered the room.
(12:30:26) npe left the room (quit: Read error: 104 (Connection reset by peer)).
(12:31:24) npe [n=npe@212.224.218.143] entered the room.
(12:37:42) acmeuser left the room (quit: Read error: 110 (Connection timed out)).
(12:41:56) npe_ [n=npe@212.224.218.143] entered the room.
(12:41:57) npe left the room (quit: Read error: 104 (Connection reset by peer)).
(12:42:22) npe_ is now known as npe
(12:43:02) npe left the room (quit: Client Quit).
(12:58:45) acmeuser [n=acmeuser@192.98.33.35] entered the room.
(13:03:01) acmeuser left the room.
(13:03:39) acmeuser [n=acmeuser@192.98.33.35] entered the room.
(13:08:33) acmeuser left the room (quit: Remote closed the connection).
(13:26:34) sea-gull left the room (quit: Read error: 110 (Connection timed out)).
(13:51:13) ***maht would rather be writing Limbo
(13:51:20) maht: even though I'm doing rc shell !
(13:53:59) mjl-: i'm doing limbo!
(13:54:01) mjl-: http://www.ueber.net/code/r/ftpsrv
(13:54:05) mjl-: fresh off the presses
(13:54:50) acmeuser [n=acmeuser@192.98.33.35] entered the room.
(14:04:22) sea-gull [n=sea-gull@95-28-96-88.broadband.corbina.ru] entered the room.
(14:07:24) anss1 [n=app@192.98.33.9] entered the room.
(14:08:03) anss1: Seems to be a bit hard to get all the things in acme-sac README working...
(14:09:37) anss1: On Ubuntu, I can't get back to this channel anymore ("login failed") although I did it once. On Windows, I succesfully join this channl but can't write to it, the Acme window with the IRC write prompt seems to send nothing...
(14:09:49) anss1: So now I am back on Pidgin..
(14:10:49) anss1: Expalanation: I am trying .deb and Windows installation for Acme-sac...
(14:12:48) anss1: On the other hand, I
(14:15:56) acmeuser left the room.
(14:16:27) acmeuser [n=acmeuser@192.98.33.35] entered the room.
(14:16:56) anssiuuser [n=anssiuus@192.98.33.35] entered the room.
(14:17:07) acmeuser left the room.
(14:21:11) underspecified_ [n=eric@softbank220043052007.bbtec.net] entered the room.
(14:21:12) anssiuuser left the room (quit: Remote closed the connection).
(14:38:47) acmeuser [n=acmeuser@90-231-71-114-no64.business.telia.com] entered the room.
(14:41:11) uriel: anss1: don't use acme-sac's irc client, mjl-'s ircfs is much better
(14:41:21) uriel: (it is sad that acme-sac doesn't use ircfs *sigh*)
(14:41:41) acmeuser left the room.
(14:44:30) dagle: This is strange. I can't resolve hostnames and such in my native os but it works fine in inferno...
(14:45:09) dagle: Inferno should use the same nameserver right?
(14:50:54) uriel: not necessarily, it has its own dns resolver
(15:29:02) gualteri left the room (quit: "leaving").
(15:32:01) mennis [n=mennis@adsl-155-180-66.asm.bellsouth.net] entered the room.
(15:33:44) gualteri [n=unknown@crespins.disca.upv.es] entered the room.
(16:04:28) mennis is now known as relocator
(16:04:37) relocator left the room (quit: ).
(16:40:14) jas left the room (quit: ).
(16:47:11) gualteri: mjl-, it's a good idea to keep the history 
(16:48:03) mjl-: yeah, it was silly of me to throw it away at first...
(16:48:14) mjl-: be sure not to commit from an old checkout :)
(16:48:27) mjl-: but i think hg will error out if you even try
(16:48:42) gualteri: yes, probably revisions won't match
(16:49:26) gualteri: anyway i still have to do a cleanup of the patches i've got lying around and merge them
(16:51:20) gualteri: we've received the plug, but it will still take a week until i can get my hands on it (it's on a friends house atm)
(16:52:16) mjl-: argh, a week!
(16:52:22) mjl-: is he far away?
(16:52:53) gualteri: a hundred km, but it's busy right now
(16:53:06) mennis [n=mennis@adsl-068-016-104-079.sip.asm.bellsouth.net] entered the room.
(16:54:37) gualteri: i'll try to tease him about letting the plug lying idle at home for one week
(17:19:22) eno: mjl- and gualteri: i suggest to expand the code to also support orion5x NAS devices
(17:21:40) eno: google for MV88F5182-usermanual.pdf
(17:29:21) gualteri: at first glance looks similar to the kirkwood, was that feroceon soc a previous desigin to the kirwood soc?
(17:29:31) gualteri: *design
(17:29:37) eno: yes, exactly
(17:30:00) eno: there's more devices available, mostly NAS
(17:32:23) eno: so maybe have the google code of inferno-feroceon ?
(17:34:49) eno: newer devices with SATA are appearing
(17:35:42) eno: qnap ts119/219, synology ds209/409, and linkstation XHL (not sure about the model)
(17:35:52) eno: they are more expensive though
(17:36:21) eno: (i mean newer kirkwood devices )
(17:37:42) gualteri: i've seen too the openrc-{client,server} from marvell/globalscale
(17:37:56) eno: that's right
(17:54:54) mjl-: at least for now i'm going to write code to make use of the sheevaplug.  new devices cost more money.  and there is lots of code to write (network services)
(18:01:31) eno: i understand, what i'm asking is to structure the codebase in a way that allow reuse
(18:02:17) eno: you've done the hard work, and lots of these can be leveraged
(18:13:35) stu8ball left the room (quit: "leaving").
(18:14:23) stu8ball [n=stuart@aurora.ossified.net] entered the room.
(18:15:08) mjl-: yeah, that makes good sense
(18:15:17) mjl-: do you have concrete ideas on how to do that?
(18:15:25) mjl-: i guess it requires me to read some of the orion docs
(18:15:31) mjl-: and get a feeling of where things are differently
(18:16:17) eno: i wonder if it's possible to build for different targets in the same directory
(18:16:17) mjl-: if we're lucky, the kernel config file can be all that's needed to compile for a different target
(18:16:43) eno: mostly kirkwood is a superset of orion5x
(18:17:10) eno: rtc is not included, for example
(18:17:20) eno: so main.c needs to be different
(18:17:34) eno: i think uart should be exactly the same
(18:17:51) eno: i don't know how to deal with io.h
(18:19:27) eno: i got a mssii with serial console
(18:19:36) eno: ready for some experiment
(18:20:31) eno: if you know anyone visiting SF bay area, i can part with my 2nd mssii
(18:21:01) eno: only 32MB RAM though
(18:21:14) soul9 left the room (quit: Excess Flood).
(18:22:38) soul9 [n=none@unaffiliated/johnnybuoy] entered the room.
(18:23:27) mjl-: io.h could be the tricky one
(18:23:55) mjl-: i bet we're not lucky enough that the registers for the various devices are at the same offsets
(18:24:39) mjl-: rtc isn't even in main.c i think.  it's in sheevainit.b, but we can probably system time setting to devrtc.c's poweron function
(18:24:41) eno: need careful check, yes
(18:24:52) mjl-: is ethernet about the same you think?
(18:25:01) eno: i think so
(18:25:57) mjl-: and then we need a sata driver :D
(18:27:08) eno: 1st step would be able to repeat what you've done for sheeva, tftp boot it
(18:27:32) eno: sata should be common to both
(18:27:47) eno: so if one is solved, the other as well
(18:29:24) mjl-: that's pretty useful actually
(18:29:42) mjl-: i'm not around sf bay area unfortunately.  but i might be able to get one of those devices second hand here.
(18:32:10) eno: yeah, since these are previous gen, they can be had not very expensively
(18:34:15) eno: some would require a bit soldering work to get serial access
(18:34:23) eno: some already have serial header on
(18:35:35) eno: (part of the reason i have mssii, not very handy in terms of soldering myself)
(18:37:02) mjl-: i have zero soldering experience knowledge, so i should get something with serial available :)
(18:38:58) eno: if you know some hardware engineer, it's a piece of cake for them
(18:41:03) eno: gotta go to work now, later
(18:50:43) j1m left the room.
(18:50:49) j1m [n=jim@unas-228.rsity.ru] entered the room.
(19:52:19) maht left the room (quit: Read error: 54 (Connection reset by peer)).
(19:53:30) maht [n=maht__@85-189-31-174.proweb.managedbroadband.co.uk] entered the room.
(20:17:54) sea-gull: hi all! Did anybody have problems with path resolution while writing file server?
(21:18:07) maht: nope
(21:18:37) maht: it is fun though
(21:18:57) sea-gull: maht: I don't know why, but my modules can't be load even they're pointed as absolute path.
(21:19:43) maht: oh, try using the stepping debugger, that's usually helpful
(21:20:08) sea-gull: maht: to be more precise they get loaded if path is "/usr/...", but in case of "/dis/.." it fails
(21:23:00) mjl-: sea-gull: what is the error message?
(21:23:31) mjl-: e.g. mymod = load Mymod Mymod->PATH; if(mymod == nil) sys->print("load: %r\n");
(21:23:32) sea-gull: mjl-: no error msg, it freezes.
(21:23:40) sea-gull: yes, it's null
(21:23:45) sea-gull: I've checked that
(21:23:54) mjl-: yeah, when it can't load it it's null
(21:23:59) mjl-: but the error string will have a useful message
(21:24:11) mjl-: probably mentioning that the file doesn't exist
(21:24:18) maht: try sys->print("%s\n", Mymod->PATH);
(21:24:19) mjl-: or that the module definition is different
(21:27:04) sea-gull: error msg: file doesn't exist
(21:28:52) mjl-: ok, so maht's line will print what it is trying to load
(21:29:13) sea-gull: PATH is /dis/git/init.dis. And that file really exists.
(21:29:55) mjl-: hmm ok
(21:29:57) mjl-: strange
(21:30:03) sea-gull: it loads when path begins with "/usr/..."
(21:30:25) mjl-: and the file is the same?
(21:30:27) maht: ls -l  /dis/git/init.dis
(21:30:34) sea-gull: yes
(21:32:22) sea-gull: ; ls -l /dis/git/init.dis
(21:32:22) sea-gull: --rw-r--r-- U 0 manzur manzur 1166 Jun 22 14:53 /dis/git/init.dis
(21:32:22) sea-gull: ; 
(21:32:51) sea-gull: btw, why it's so big?!
(21:32:52) mjl-: and the version in /usr/...?
(21:32:55) mjl-: 1166 bytes?
(21:33:00) mjl-: is that big?
(21:34:04) sea-gull: ; ls -l /usr/manzur/gitfs/init.dis
(21:34:04) sea-gull: --rw-r--r-- U 0 manzur manzur 1154 Jun 22 18:40 /usr/manzur/gitfs/init.dis
(21:34:04) sea-gull: ; 
(21:34:23) sea-gull: it's an old version. 
(21:36:08) mjl-: very strange... is the code online somewhere?
(21:36:12) mjl-: i haven't seen it before
(21:36:19) sea-gull: not yet.
(21:36:41) sea-gull: I had the same problem with ordinary files while mapping them to repo.
(21:37:02) sea-gull: then I added repopath to the pathname.
(21:41:11) mjl-: what happens if you put a different file in /dis/git/init.dis?
(21:41:16) mjl-: does the error message change?
(21:41:24) sea-gull: the same
(21:41:34) mjl-: then i guess there is something wrong with reading the file
(21:41:37) mjl-: you can cat it?
(21:41:46) mjl-: wc /dis/git/init.dis
(21:41:48) mjl-: or something
(21:42:06) sea-gull: yes
(21:42:23) sea-gull: ; wc /dis/git/init.dis 6      60    1166 /dis/git/init.dis
(21:42:40) gualteri: i've sometimes misconfigured the namespace and afterwards wasn't able to list/open files, maybe it's happening something similar
(21:43:21) sea-gull: gualteri: I've restarted it, but didn't help.
(21:43:57) sea-gull: I'll put that online today. No need to guess. Thanks you all!
(21:50:19) sea-gull: mjl-: how to make "cd .." work correctly with my file server? 
(21:51:00) mjl-: you have to handle ".." explicitly
(21:51:05) mjl-: in walks
(21:51:26) mjl-:                 Walk => 
(21:51:26) mjl-:                         if(op.name == "..") {
(21:51:26) mjl-:                                 op.reply <-= (dir(Qroot, 0), nil);
(21:51:26) mjl-:                                 continue again;
(21:51:26) mjl-:                         }
(21:51:29) mjl-: something like that
(21:51:40) mjl-: in a navigator
(21:51:48) sea-gull: aha :)
(21:52:41) mjl-: but that assumes that there is only one level of depth in the root, so .. of anything goes to Qroot.  if you have more levels of depth, you'll need to do something smarter
(21:52:44) sea-gull: mjl-: doesn't it handle ".." for the root dir?
(21:52:54) mjl-: no
(21:52:57) mjl-: inferno handles that
(21:53:41) mjl-: if inferno see a "cd .." at a mountpoint, it won't query your fs it, but it goes directly to the "higher" (..) file server
(21:53:44) sea-gull: by it I meant Inferno.
(21:54:08) mjl-: ah
(21:54:19) mjl-: so it does handle that
(22:05:58) maht: the next bit for me to say is sys->stat("/dis/git/init.dis")
(22:06:21) maht: before the load, make sure it has the same ns as you think it does
(00:26:48) mennis left the room (quit: Client Quit).
(00:38:01) backnight [n=backnigh@189.113.96.13] entered the room.
(00:39:42) backnight left the room (quit: Client Quit).
(00:57:33) mennis [n=mennis@adsl-155-180-66.asm.bellsouth.net] entered the room.
(01:06:49) sea-gull left the room (quit: "Lost terminal").
(01:38:21) mennis left the room (quit: Client Quit).
(05:25:00) mennis [n=mennis@adsl-155-180-66.asm.bellsouth.net] entered the room.
(05:39:58) mennis left the room (quit: Client Quit).
(05:49:48) eno_ [n=eno@adsl-70-137-131-139.dsl.snfc21.sbcglobal.net] entered the room.
(06:01:53) eno left the room (quit: Success).
(06:26:09) eno_ is now known as eno
(06:58:55) sea-gull [n=sea-gull@95-28-14-65.broadband.corbina.ru] entered the room.
(07:07:17) fwiffo [n=user@unaffiliated/fwiffo] entered the room.
(07:55:30) anss1 left the room (quit: Read error: 104 (Connection reset by peer)).
(07:56:16) anss1 [n=app@192.98.33.9] entered the room.
(07:58:30) anss1 left the room (quit: Read error: 104 (Connection reset by peer)).
(08:15:25) anss1 [n=app@192.98.33.9] entered the room.
(08:15:33) fwiffo left the room (quit: Remote closed the connection).
(09:18:58) mjl-: ah, the inferno projects on ohloh have now been recounted, and structured basic has been replaced by limbo
(10:01:46) uriel: awesome
(10:01:59) uriel: maybe they will replace perl with rc for werc soon :
(10:02:00) uriel: )
(10:12:20) mjl-: you'll probably have to send a patch for that :)
(10:15:23) uriel: ah, you sent a patch for limbo? nice :)
(10:18:44) mjl-: yeah and i'm pretty happy it got in :)
(10:19:02) mjl-: i'm now making hgfs talk the hg http pull protocol, so ohloh can index my code too
(10:19:39) uriel: ah, really great!
(11:38:46) arcfide [n=arcfide@h-68-165-56-221.chcgilgm.dynamic.covad.net] entered the room.
(11:39:11) arcfide: Hello everyone. I'm using Acme SAC, and I'm kind of wondering if I have the latest version (0.11 Linux), and if there is a way to build a native x86_64 version?
(11:40:47) mjl-: there is no 64-bit inferno afaik.  you can compile inferno yourself, on linux with gcc -m32
(11:41:15) mjl-: i don't know much about acme-sac, i bet it's mostly a matter of "mk install" in the root, just as on normal inferno
(11:42:23) arcfide: mjl-: Okay, it's just that I'm confused whether I have the right version of Acme SAC and Inferno running.
(11:43:01) gualteri: arcfide: the instructions are at http://code.google.com/p/acme-sac/ & http://code.google.com/p/acme-sac/source/browse/sys/Linux/build-Linux.sh
(11:44:52) arcfide: gualteri: Where can I get the latest sources? I'm a bit confused about what is the latest version, since I see different versions for each of the platforms.
(11:45:34) gualteri: on the googlecode project are the instructions: http://code.google.com/p/acme-sac/source/checkout
(11:48:01) arcfide: Do any of you use Acme? I'm playing around with getting a different editor for my programming, and I am wondering if you have any experiences for a Vi/NEdit/Emacs guy?
(11:50:40) mjl-: i use acme and vim.  on unix (openbsd) i still mostly use vim.  part of the reason is that i have a sucky mouse.  and related, part of the reason is that acme doesn't have enough keyboard shortcuts for me.  but acme's use of the mouse is very good.
(11:50:57) mjl-: can't say anything about users coming from emacs
(11:51:33) arcfide: mjl-: Okay, I'm an OpenBSD user stuck on RHEL for the moment, so I figured I would give something crazy a try. I am not afraid to use the mouse.
(11:51:36) mjl-: in short:  some work loads work better for acme, some better for vim (or your current editor)
(11:52:02) arcfide: I'm most used to working with nvi.
(11:54:46) arcfide: Thanks for the help.
(12:00:57) arcfide left the room.