Conversation with #inferno at Tue Jun 23 12:44:22 2009 on powerman-asdf@irc.freenode.net (irc)
(17:18:29) sav [n=savio@jagat.xored.org] entered the room.
(17:29:50) mennis [n=mennis@adsl-155-180-66.asm.bellsouth.net] entered the room.
(17:31:22) j1m left the room (quit: Remote closed the connection).
(17:35:37) mennis left the room (quit: Client Quit).
(17:35:48) mennis [n=mennis@adsl-155-180-66.asm.bellsouth.net] entered the room.
(18:12:21) underspecified_ left the room (quit: Connection timed out).
(20:46:08) j1m [n=jim@unas-228.rsity.ru] entered the room.
(21:28:34) j1m left the room (quit: Read error: 110 (Connection timed out)).
(21:42:27) underspecified left the room (quit: ).
(22:57:26) j1m [n=jim@unas-228.rsity.ru] entered the room.
(23:39:09) arcfide [n=arcfide@h-69-3-100-36.chcgilgm.dynamic.covad.net] entered the room.
(23:40:07) arcfide: I am using Acme SAC, and I would like to be able to use the Acme process to run some commands from my host environment. Is there a way to do this? Right now, I can only run those things that are already built in with inferno, and I can't seem to interact with my host environment beyond reading files from the file system.
(23:40:38) arcfide: Come to think of it, is there a way of shortcuting my way to my home directory? It alwasy pulls up into the acme-home, but this isn't where I want to be.
(23:42:19) maht: os module
(23:42:46) maht: one step closer to running Inferno & wm/wm
(23:43:29) mennis: How do I get native inferno to have a /dev/random ?  I have random in the port section of my recipe.
(23:50:37) mjl-: mennis: that should be in cons(3)
(23:50:39) mjl-: it's not there?
(23:53:40) mennis: should I have a "cons random" under dev then?
(23:56:17) sea-gull left the room (quit: "Lost terminal").
(23:56:59) mennis: no that didn't work.  
(00:05:02) mennis: Oh weird.  I do have one it is osinit.dis that is complaining.
(00:15:31) mennis: Is there such a thing as osinit.dis?  Is it wrapped up in C code somewhere?
(00:16:35) mjl-: you should just have a cons under dev
(00:16:46) mjl-: and then it should be available as '#c/random'
(00:16:47) mjl-: i think
(00:18:13) mjl-: are you sure '#c' is mounted on /dev?
(00:18:36) mjl-: os/port/devcons.c doesn't seem to have any options for leaving just "random" out
(00:26:20) mennis: OK it seems to be there once my init runs. 
(00:26:32) mennis: Oh ho! I think I got it.
(00:27:46) mennis: calling dhcp module load too early I think
(00:30:06) mennis: Yup thanks for the help.  I was trying to put all my loads early in the my init function but there was no /dev/'s yet.
(00:42:56) j1m left the room (quit: Read error: 104 (Connection reset by peer)).
(00:46:24) j1m [n=jim@unas-228.rsity.ru] entered the room.
(00:59:17) underspecified [n=eric@softbank220043052007.bbtec.net] entered the room.
(01:01:58) underspecified left the room (quit: Client Quit).
(01:02:59) mennis left the room (quit: Client Quit).
(01:12:19) underspecified [n=eric@softbank220043052007.bbtec.net] entered the room.
(03:23:05) underspecified left the room (quit: ).
(04:55:52) arcfide left the room.
(04:58:45) underspecified [n=underspe@leopard175.naist.jp] entered the room.
(04:59:40) arcfide [n=arcfide@h-69-3-100-36.chcgilgm.dynamic.covad.net] entered the room.
(04:59:46) arcfide left the room (quit: Remote closed the connection).
(06:23:50) j1m left the room (quit: Read error: 110 (Connection timed out)).
(06:36:14) MrWGW left the room (quit: clarke.freenode.net irc.freenode.net).
(06:36:16) uriel left the room (quit: clarke.freenode.net irc.freenode.net).
(06:36:16) sav left the room (quit: clarke.freenode.net irc.freenode.net).
(06:36:24) MrWGW [n=MrWGW@74.124.206.166] entered the room.
(06:36:40) sav [n=savio@jagat.xored.org] entered the room.
(06:36:40) uriel [n=uriel@li43-28.members.linode.com] entered the room.
(06:52:07) j1m [n=jim@unas-228.rsity.ru] entered the room.
(08:11:45) wardub [n=inferno@sel158.154.und.nodak.edu] entered the room.
(08:14:36) wardub: Not sure if anyones here or not, but I'm trying to install inferno with linux as the host os. I ran the install sh script but nothing seems to be happening.  How long does it usually take to install?  I'm planning on running this overnight
(09:37:21) underspecified_ [n=eric@isa7-dhcp-116-205.naist.jp] entered the room.
(10:09:06) soul9: wardub: it takes a while, though i don't know how up-to-date the tarballs on vitanuova.com are. We mostly use the code checked out from code.google.com/p/inferno-os, lately inferno has had a surge of development made on it, and there are quite a few fixes, especially for linuxes with newer libcs and X servers
(10:09:41) soul9: (even more broken libcs and Xservers)
(10:12:15) wardub: hmm, well do you think it will be done when I wake up this morning/
(10:12:18) wardub: ?
(10:12:57) uriel: soul9: the vn.com crap is only two or three years outdated... 
(10:13:01) uriel: no big deal!
(10:13:14) soul9: ☺
(10:13:16) uriel: wardub: do not use any of the itiodic install scripts
(10:13:27) uriel: only sane way to install inferno is from source at google-code
(10:13:42) soul9: wardub: i think that that tarball is hugely outdated, just install from inferno-os svn
(10:13:46) soul9: wardub: what distro?
(10:13:55) wardub: slackware 12.2
(10:14:05) wardub: I'll check google-code
(10:16:28) soul9: ok, no packages for that ☺
(10:16:45) soul9: mjl-: wow, i just discovered the middleclick-scroll in wm/irc, that's pretty awesome!
(10:20:03) wardub: So how and what exactly do I need to download?
(10:20:40) mjl-: soul9: the middleclick scroll is standard tk behaviour!
(10:20:53) mjl-: wardub: this might help:  http://www.xs4all.nl/~mechiel/inferno/getting-started.html#installing
(10:22:03) wardub: thanks, I'll try this in the morning.  Need sleep.
(10:22:04) mjl-: if you're in a hurry you can skip the "file tree" & "fetching fonts" sections, and just do "fetching inferno" and "compiling"
(10:22:15) mjl-: cool
(10:22:16) mjl-: sleep well
(10:25:05) soul9: ah..
(10:25:19) soul9: well, nonetheless it's great, i didn't know about it
(10:35:25) soul9: hmm, if it worked in the channels bar i wouldn't even care that there are so many names in my sidebar i need to scroll a lot
(10:41:52) j1m left the room (quit: "Leaving.").
(10:42:03) j1m [n=jim@unas-228.rsity.ru] entered the room.
(11:54:05) underspecified_ left the room (quit: ).
(12:01:23) wardub left the room (quit: Read error: 110 (Connection timed out)).
(13:10:53) sea-gull [n=sea-gull@95-28-109-134.broadband.corbina.ru] entered the room.
(13:18:30) sea-gull: mjl-: hi!
(13:46:58) j1m left the room.
(14:29:08) mjl-: yo sea-gull!
(14:30:22) sea-gull: mjl-: have you read the thread about gitfs?
(14:33:31) mjl-: not yet
(14:33:33) mjl-: will do in a bit
(14:53:28) mjl-: read it
(14:53:35) mjl-: i don't really understand what it's about
(14:53:43) mjl-: there are many ideas loosely mentioned
(14:53:49) mjl-: and not very finely explaiend
(14:54:50) mjl-: oh, i guess that "webfs" means the program webfs(4), i was interpreting it as <negative-connotation>-fs
(14:55:46) sea-gull: hm. Moreover, I don't see why they think that cp, rm won't work?
(14:56:41) mjl-: well, i have the vague idea that fgb wants to "cp /some/where /mnt/git/new/file", and make it end up in the git repo
(14:57:20) sea-gull: and some/where is in git mntpt?
(14:57:34) mjl-: but such designs are a bit tricky:  where should gitfs leave the file until the commit?  you really don't want a create & each write to cause a commit (without commit message).  and what if the gitfs is killed and started again?
(14:57:43) mjl-: no, /some/where just a random file in the file system
(14:59:00) sea-gull: yes, it's more flexible to seperate commit from adding to the working tree.
(14:59:04) mjl-: hmm, git can store uncommitted files somewhere in its repo right?  perhaps that is possible then.  for mercurial it doesn't make sense
(14:59:20) mjl-: yeah, exactly, i think you want creating/adding files & committing separate
(14:59:42) sea-gull: it stores in active dir and in object repo, if you added it to the index.
(15:00:05) mjl-: hmm, on first add already?
(15:00:12) sea-gull: yes
(15:00:19) MrWGW left the room (quit: clarke.freenode.net irc.freenode.net).
(15:00:33) MrWGW [n=MrWGW@74.124.206.166] entered the room.
(15:00:44) mjl-: would make more sense if it only did that when updating another revision
(15:00:44) mjl-: hmm, "update" is not the right word in git for that.
(15:01:09) mjl-: i mean getting a different revision to the current working directory.  if you do that, then it puts uncommitted changes in some temporary place, right?
(15:01:42) sea-gull: you can use git stash for that.
(15:03:21) sea-gull: do you think I wasn't clear enough in describing ideas?
(15:03:21) mjl-: ah, ok, then it may at least technically be feasible to allow random file tree modifications to a /mnt/git/ (where a gitfs is mounted), and then apply all styx operations (create, write, remote, etc) in the stash
(15:04:01) mjl-: of all the mails, i think yours is the clearest of them
(15:04:32) mjl-: but i think we can talk more (on the mailing list as well) about how gitfs is going to be used
(15:04:38) mjl-: so that at least we understand what we are all talking about
(15:04:55) sea-gull: agree 100%. 
(15:05:25) sea-gull: what about file-specific commands? Which of the method do you prefer?
(15:05:27) mjl-: the git fs code is not yet in your git repo, right?
(15:05:36) sea-gull: it's there.
(15:05:52) sea-gull: I uploaded it that day(Monday).
(15:05:55) mjl-: hmm, which file is it?
(15:06:25) mjl-: git://github.com/manzur/gitfs.git is still the one, right?
(15:07:21) sea-gull: yes.
(15:08:44) mjl-: ah, sorry
(15:08:47) mjl-: was using the wrong commands
(15:08:58) mjl-: i need a cheat sheet for git
(15:09:50) sea-gull: imho, the best cheat sheet is git help git.
(15:10:37) soul9: hgfs only displays the last revision of everything?
(15:11:07) mjl-: sea-gull: hah, ok, i'll use that.  i used it just now to find the right command too
(15:11:20) mjl-: soul9: yeah, it only lists the last revision
(15:11:32) soul9: ok, thanks
(15:11:33) mjl-: you can access other revisions by number though.  they are just not listed
(15:11:40) soul9: ah
(15:11:51) soul9: so instead of last/ you'd use rev/
(15:11:55) mjl-: yeah
(15:11:59) mjl-: try ls -ld last
(15:12:08) mjl-: you'll see the number of the last revision if all is well
(15:12:19) soul9: ah!
(15:12:22) soul9: yep
(15:12:31) sea-gull: mjl-: so, which method do you prefer: extra ctl files or read only files under <files> ?
(15:12:40) mjl-: sea-gull: i don't know yet
(15:12:53) mjl-: i think there need to be some read-only files anyway
(15:13:00) mjl-: e.g. for the various branches
(15:13:34) mjl-: ron wants to access any revision/branch by cd'ing to or cat'ing e.g. /files/somebranch/some/dir/in/thebranch
(15:13:53) MrWGW left the room (quit: clarke.freenode.net irc.freenode.net).
(15:13:59) MrWGW [n=MrWGW@74.124.206.166] entered the room.
(15:14:04) mjl-: i think it's a bit too implicit to have all those files writable too
(15:14:10) sea-gull: at the current it's possible to cat files from objects store.
(15:14:34) mjl-: yeah, like this right?  cat /mnt/git/files/some/file
(15:14:38) sea-gull: selecting files from specific branch could be done.
(15:14:58) sea-gull: no, cat objects/1e/...
(15:15:10) mjl-: ah
(15:15:15) sea-gull: files under <files> dir are files from working dir.
(15:15:33) mjl-: hum, but which version?
(15:15:48) sea-gull: working dir hasn't any version.
(15:16:00) mjl-: but where do the contents of the files come from?
(15:16:17) mjl-: is it just mirroring the contents of the working dir?
(15:16:28) sea-gull: correct.
(15:16:37) mjl-: ok
(15:16:40) sea-gull: you can checkout to change the version.
(15:17:13) mjl-: so why would you use the files in files/* instead of just the current working directory?
(15:17:25) mjl-: "instead of just the files in the current working directory"
(15:17:34) sea-gull: I don't want to mix special files with ordinary ones.
(15:18:09) mjl-: ah i see.  yeah that makes sense
(15:18:21) mjl-: you mount no /n/git, right?  not /mnt/git
(15:18:27) mjl-: in your testing i mean
(15:18:44) sea-gull: you should point the dir.
(15:18:50) sea-gull: no default dir.
(15:18:55) mjl-: ok, what do you use?
(15:19:20) sea-gull: /usr/manzur/gitfs/mntpt
(15:19:23) mjl-: ah, ok
(15:19:34) sea-gull: you can use anything. Or is there a problem?
(15:19:56) mjl-: i thought i'ld use the same thing as you, so would be using the same paths in examples
(15:20:02) mjl-: no problem though
(15:20:07) mjl-: i'll keep using /mnt/git :)
(15:20:28) sea-gull: ok
(15:20:49) mjl-: for adding files to the to-be-committed list, gitfs uses the same format as normal git, right?
(15:21:27) mjl-: so you can see .git/ in the current working dir (not /mnt/git), and see them change after you've added a new file to to-be-committed list
(15:22:41) sea-gull: yes. What do you mean by to-be-committed list? Index file? 
(15:22:57) mjl-: yeah
(15:22:59) mjl-: i think so at least
(15:23:03) mjl-:        git-add - Add file contents to the index
(15:23:12) sea-gull: No it uses its own index format.
(15:23:25) mjl-: ah, that's good to know
(15:23:33) sea-gull: as git clone, afair, doesn't need index file.
(15:23:57) sea-gull: but commits, trees, and blobs should be compatible with original git.
(15:24:06) mjl-: yeah, i bet the index file is normally only used when you modify the repo?
(15:24:09) mjl-: ok
(15:24:25) mjl-: and gitfs does have its own index in the .git/ directory?
(15:24:56) sea-gull: yes, it's created when you send "echo init > ctl".
(15:25:13) mjl-: ok
(15:25:15) sea-gull: the same with objects store(object store is objects dir).
(15:25:17) mjl-: all makes sense
(15:25:43) mjl-: did roman have ideas on how he would use gitfs?
(15:26:08) mjl-: the same for ron.  he's one of the people that would benefit & actually use a gitfs.  so if he has ideas on how he would like to use it, we can take that into account
(15:26:34) mjl-: i'm curious how he would want to do adds and commits
(15:26:46) mjl-: and checkout a branch to the current working dir
(15:26:52) sea-gull: As I said in the ml, it's possible to add virtual branches.
(15:27:34) mjl-: yes, how would that work?
(15:27:52) sea-gull: from his words he doesn't want to checkout branches.
(15:28:15) mjl-: checkout == make the current working directory that branch/version, right?
(15:28:39) sea-gull: aha
(15:28:43) mjl-: checkout to me sounds like "pull the changes from remote to local over e.g. http"
(15:29:22) sea-gull: pull = fetch + checkout
(15:29:33) mjl-: ah, ok
(15:29:54) mjl-: yeah.  this is confusing to me.  for example, hg pull == just fetch
(15:30:09) mjl-: and "hg update" == git's checkout
(15:30:12) sea-gull: I think Ron wants only read-only version of the git. It's much more easier.
(15:30:17) mjl-: yeah
(15:30:19) mjl-: exactly
(15:30:24) mjl-: and then he can use diff
(15:30:36) mjl-: which makes a lot of sense to want
(15:30:45) mjl-: you said in your mail:
(15:30:52) mjl-: I agree with you that viewing branches would make git more convenient.
(15:30:52) mjl-: But decision was made towards saving on extra space needed for each
(15:30:52) mjl-: extracted branch.
(15:30:53) sea-gull: he can use diff anyway. As I process read of the files as extracting them.
(15:31:20) sea-gull: Decision of git developers, not me.
(15:31:21) mjl-: yeah, but right now he can only access the right "current working directory version of the files", right?
(15:31:30) mjl-: ahhh
(15:31:34) mjl-: that was not clear to me
(15:32:02) mjl-: to me, your line read that you didn't plan on implementing that because it would require more space
(15:32:27) sea-gull: you can access old files(files under objects/)  as well as current working dir.
(15:33:08) mjl-: ok.  objects are of the sha1 strings, right?
(15:33:25) sea-gull: yes. I know that it's inconvenient.
(15:33:27) mjl-: "are in the form of sha1 strings'*
(15:33:31) mjl-: ok
(15:33:42) sea-gull: correct.
(15:33:44) mjl-: well, as you said, you can extract the branches on the fly
(15:34:42) mjl-: so:  1. read-only views of the branches (or any revision perhaps) would be good, but you already said that in your mail.
(15:34:54) sea-gull: the main question was: why they think that cp, rm can't be used with the 2 method?
(15:35:01) mjl-: 2. i'm curious how ron would want to do checkouts, adds, commits, etc
(15:35:45) mjl-: i think they have in their minds the idea that a cp into /mnt/git/files/some/where would automatically add the file to the git repo
(15:35:59) mjl-: but i don't know if ron wants to use it that way
(15:36:03) underspecified_ [n=eric@softbank220043052007.bbtec.net] entered the room.
(15:36:47) mjl-: perhaps ron wants to normally work in his current working directory (write files there), then sometimes view other branches (through /mnt/git/somebranch/*), and then do "something" to add/commit/checkout in the working directory
(15:37:47) mjl-: does he want a git/commit?  or an 'echo commit msg' >/mnt/git/ctl?  i don't know.  something to ask him
(15:38:58) yiyus: imo, in the root you should have a commit file and all the revisions (being only writable the last one)
(15:39:04) yiyus: then, when you want to commit
(15:39:13) sea-gull: git/commit isn't good solution cause of direct interacting with file server modules.
(15:39:31) yiyus: you do echo 'message' > commit and the last dir is renamed to the revision and a new last dir is created
(15:40:16) mjl-: yiyus: and then all writes that happen before a commit should go in a stash of some sort?
(15:40:24) mjl-: so the changes aren't lost if gitfs is killed?
(15:40:51) mjl-: i would thing the normal working directory, which is on fossil/kfs, would be just as easy?
(15:41:31) mjl-: sea-gull: git/commit could just be a shell script that writes "echo commit msg >$dotgit/fs/ctl" or something
(15:41:32) yiyus: yep, in the working directory
(15:41:33) sea-gull: yiyus: what about adding and removing files from the index file?
(15:42:23) sea-gull: mjl-: what about updating files under the mount point?
(15:42:37) mjl-: sea-gull: couldn't you bind the gitfs on .git/fs/, and make some handy shell scripts like git/add, git/checkout, git/commit that look for a .git/fs/ dirctory in ./, then ../, then ../../, then ../../../ etc?
(15:43:05) yiyus: sea-gull: I guess the easier solution is a ctl file, though you could have an index dir and cp/rm from there too
(15:43:59) yiyus: it would feel right doing something like cp workingdir/*.c index/
(15:44:08) mjl-: sea-gull: hmm, what do you mean by "updating files under the mount point"?
(15:44:43) sea-gull: mjl-: forget. what you told(1 step ago) is the way it's done in git.
(15:45:06) mjl-: yiyus: but wouldn't you then also want to edit in index/*, and wouldn't it then be just another working directory?
(15:45:10) sea-gull: mjl-: It's a good idea.
(15:45:48) mjl-: sea-gull: yeah i thought so, about looking in ./ then ../ then ../../ etc.  mercurial does it too.  the difference will be that the scripts then don't use the git repo accessing modules, but just talk to the gitfs.
(15:46:21) mjl-: so the scripts would really just be simple scripts for convience, so you don't have to echo some string to some path like ../../../.git/fs/ctl or /mnt/git/ctl
(15:46:29) yiyus: mjl-: yes, it would be like a working directory, what is the problem with that?
(15:46:43) sea-gull: guys, would you mind if I paste somewhere this log and point to that in the mail?
(15:46:54) yiyus: btw, I have not really thought about this, but it is how I understand using common file operations
(15:47:11) yiyus: (and I don't know too much about git...)
(15:48:16) sea-gull: indeed, it's a good idea presenting index file as a dir.
(15:48:30) mjl-: sea-gull: yeah, that's fine by me
(15:48:30) sea-gull: cp,rm could be used instead of adding and removing.
(15:48:50) sea-gull: but what about stages?
(15:49:11) sea-gull: each file has a stage in the index.
(15:49:12) mjl-: yiyus: i mean:  what would it add to have such an index/, if it's no different from the working dir that you already have?
(15:49:27) sea-gull: mjl-: it's different.
(15:49:30) yiyus: do you know the equivalent of stages in hg? I don't really know what they are
(15:49:45) mjl-: hmm, i have to read up on stages i think
(15:50:03) sea-gull: stages are flags that indicate file state: 0 for ordinary files
(15:50:05) yiyus: mjl-: in the working directory you would have object files, etc, which would never go to the index
(15:51:02) sea-gull: 1 - for files from the src commit that's to be merged, 2 - for files from dst commit.
(15:52:10) mjl-: yiyus: i see.  so a write to a file in index/*, that would update the working directory too?  if not, how are you going to compile your new code?  inside index/*?  if so, that would create new files, putting them in the index... right?
(15:53:20) yiyus: mjl-: what you write to the index file goes to the working dir too, but not the other way around
(15:54:03) yiyus: in fact, I think it would be better to forget about cp, don't allow creation in the index dir and just use bind
(15:56:10) mjl-: yiyus: i can see the point about the initial idea.  but compile & edit cycles wouldn't work:  the error messages named the working dir file.  and writes (after fixing the syntax error) wouldn't go to the index dir
(15:56:25) mjl-: what is the idea on using bind?
(16:00:28) mjl-: i have no ideas on recording the stage info anywhere
(16:01:58) sea-gull: it can be done via concatenation stage to the name of the file.
(16:06:06) sea-gull: mjl-: more about stages: git help rev-parse
(16:06:42) mjl-: sea-gull: btw, could you add the modified tables.[bm] to the gitfs repo?  that would make it easier for others to try it (get it to compile)
(16:06:57) mjl-: a mkfile that has the files that are supposed to compile is nice too
(16:07:28) mjl-: are you using compiling now by calling limbo explicitly?
(16:07:40) sea-gull: mjl-: I'm using mk.
(16:08:23) sea-gull: mjl-: should new tables.m modify /modules/tables.m ?
(16:08:47) mjl-: i think so.  you're using the all() function in Strhash, which isn't there in stock inferno, is it?
(16:09:14) sea-gull: no, it isn't.
(16:09:24) mjl-: if you copy tables.[bm], add Strhash.all(), and make the PATH in tables.m just "tables.dis", it'll work if you just run myfs from the gitfs checkout dir
(16:58:21) mennis [n=mennis@adsl-068-016-104-079.sip.asm.bellsouth.net] entered the room.
(17:03:05) wardub [n=inferno@sel158.154.und.nodak.edu] entered the room.
(17:09:55) mennis left the room (quit: Client Quit).
(17:14:19) mennis [n=mennis@adsl-068-016-104-079.sip.asm.bellsouth.net] entered the room.
(17:17:24) yiyus: mjl-: sorry, I'm at work, about binds... I was only thinking loudly, it is not a good idea
(17:20:39) wardub left the room (quit: Remote closed the connection).
(17:33:43) wardub [n=inferno@sel158.154.und.nodak.edu] entered the room.
(17:34:32) wardub: I'm still not sure if anyone is around but I'll leave this here.  I decided to try and compile inferno, but I am getting a ton of errors.  the first few look like this gaux.c:1:18: error: lib9.h: No such file or directory
(17:34:32) wardub: In file included from regaux.c:2:
(17:34:32) wardub: /home/inferno/inferno_root/inferno-os/utils/include/regexp.h:16: error: expected specifier-qualifier-list before 'Rune'
(17:34:32) wardub: /home/inferno/inferno_root/inferno-os/utils/include/regexp.h:21: error: expected specifier-qualifier-list before 'Rune'
(17:34:34) wardub: /home/inferno/inferno_root/inferno-os/utils/include/regexp.h:29: error: expected specifier-qualifier-list before 'Rune'
(17:34:37) wardub: /home/inferno/inferno_root/inferno-os/utils/include/regexp.h:40: error: expected specifier-qualifier-list before 'Rune'
(17:34:40) wardub: /home/inferno/inferno_root/inferno-os/utils/include/regexp.h:65: error: expected declaration specifiers or '...' before 'Rune'
(17:34:43) wardub: /home/inferno/inferno_root/inferno-os/utils/include/regexp.h:66: error: expected ')' before '*' token.
(17:34:45) wardub: any ideas what I'm doning wrong?
(17:35:22) sea-gull: wardub: did you change mkconfig?
(17:35:29) wardub: I need to go to class so I'll check back in a few hours
(17:35:36) wardub: Yes I did sea-gull
(17:35:50) wardub: Changed root, host and object
(17:37:27) sea-gull: wardub: where did you get the source?
(17:37:56) wardub: http://www.ueber.net/inferno/tgz/
(17:39:41) soul9: ☺
(17:39:48) sea-gull: soul9: what?
(17:39:53) soul9: wtf? i didn't even see that...
(17:39:59) soul9: no, no problem
(17:40:13) sea-gull: I think it's issue with mkconfig
(17:40:27) soul9: just i didn't see, now it's hosted on github, ueber.net and gcode
(17:40:45) soul9: oh, noes, it's plan9 hosted on github, nvm
(17:41:25) sea-gull: ueber contains tgz version, so no need to use svn.
(17:42:56) soul9: ok
(17:43:10) mjl-: google svn checkouts used to be very slow (i think they're a bit faster nowadays).  so i put up the .tgz's to get started quickly.  it is probably still lots faster than google svn checkouts.
(17:43:21) soul9: yeah, it can't find lib9.h, i bet that's an mkconfig problem
(17:43:26) mjl-: yeah
(17:43:33) mjl-: wardub: could you paste your mkconfig somewhere?
(17:44:34) wardub: http://pastebin.com/d3f7d6d76
(17:44:49) wardub: k sorry guys I gotta go, I'll try what you say later. Thanks
(17:45:02) sea-gull: $
(17:45:08) sea-gull: remove it
(17:45:28) sea-gull: OBJTYPE=$386
(17:45:55) soul9: ☺
(17:45:56) mjl-: exactly, what sea-gull says ☺
(17:46:04) soul9: OBJTYPE=386 no?
(17:46:17) soul9: i have that...
(17:46:35) sea-gull: I mean that he should remove $ from $386
(17:46:37) wardub: Ohh that was dumb ty
(18:07:37) soul9: mjl-: i made ircfs completely disconnect again by sending a too long line...dunno wether you just don't want to handle that or what...
(18:08:59) mjl-: that's not the intended behaviour no
(18:09:13) mjl-: it's ircfs that disconnects after a long line?
(18:10:52) mjl-: http://www.ueber.net/code/log/full/ircfs/72#r70
(18:10:56) mjl-: that was the intended fix
(18:11:54) soul9: or maybe it lets a long lines go through
(18:12:03) soul9: and the the server says fuck you ☺
(18:12:42) soul9: ah, yeah, i see
(18:12:57) soul9: i pasted something really weird in, i admit
(18:13:54) soul9: like, i just generated an 512 chars long line, i'm going to paste it in...
(18:14:01) soul9: splat
(18:14:09) soul9: yeah, it's fine
(18:14:19) soul9: it tells me writing: line too long
(18:14:42) soul9: so it was just the fact that it was a weird command output, that wasn't one line, it was several lines formatted
(18:14:47) mjl-: ah
(18:15:05) mjl-: i bet you wrote stuff like newlines
(18:15:12) mjl-: they will be passed on
(18:15:21) mjl-: so if you write text\nquit
(18:15:23) mjl-: you'll quit irc
(18:15:26) soul9: :)
(18:15:30) mjl-: perhaps i should filter that :P
(18:15:43) soul9: but it was because of a too long line, same error as i had some time ago
(18:15:51) mjl-: hmm, or maybe it doesn't
(18:16:02) mjl-: let such raw commands through.  but it will let other data through
(18:16:10) mjl-: perhaps the server didn't like that
(18:16:23) soul9: say, a \n?
(18:16:34) soul9: lemme try a \n
(18:16:40) soul9: 17:16 johnny@wxcv  ~  $ 
(18:16:40) soul9:  17:16 johnny@wxcv  ~  $ 
(18:16:45) soul9: nope, that's fine too
(18:16:48) soul9: weird... :-/
(18:17:19) soul9: i wish i remembered what it was i pasted in, but it was at least 20 or so lines, in a private message
(18:17:34) mjl-: it wasn't a flood?
(18:17:42) soul9: yes, it was
(18:17:58) mjl-: ok, ircfs doesn't stall writes
(18:17:59) soul9: ok, so that's supposed to happen if i talk too much? even in priv.?
(18:18:03) soul9: ok
(18:18:03) mjl-: yeah
(18:18:08) mjl-: servers will kill you
(18:18:11) mjl-: kind of strange
(18:18:19) soul9: so i should pay attention how much i send in one message
(18:18:22) soul9: :)
(18:18:23) mjl-: they can also just slow down their reading of commands
(18:18:32) soul9: they don't, they killed me :D
(18:18:47) mjl-: typical irc op bofh attitude ;)
(18:19:03) soul9: :D
(18:19:30) mjl-: # 17:21 soul9 (soul9!n=none@unaffiliated/johnnybuoy) has quit: Excess Flood
(18:19:34) soul9: yep :D
(18:19:34) mjl-: that's in my logs.  i bet that was it
(18:20:20) mjl-: the easiest solution is if you just don't paste too much/quickly at a time :P
(18:20:37) mjl-: perhaps the server limits are documented (per server0
(18:21:31) mjl-: time to cook some dinner, bbl
(18:21:38) soul9: bon appetit
(18:21:41) soul9: and thanks
(20:07:54) wardub left the room (quit: Remote closed the connection).
(20:09:21) wardub [n=inferno@sel158.154.und.nodak.edu] entered the room.
(20:12:12) j1m [n=jim@unas-228.rsity.ru] entered the room.
(20:45:11) dagle left the room (quit: Read error: 104 (Connection reset by peer)).
(20:45:14) j1m1 [n=jim@unas-228.rsity.ru] entered the room.
(20:51:09) dagle [n=weechat@host162-104.bornet.net] entered the room.
(20:58:41) j1m left the room (quit: Read error: 110 (Connection timed out)).
(23:50:27) wardub left the room (quit: Remote closed the connection).
(00:00:32) sea-gull left the room (quit: "Lost terminal").
(00:18:31) mennis left the room (quit: Client Quit).
(02:11:55) dagle1 [n=weechat@host162-104.bornet.net] entered the room.
(02:12:23) dagle left the room (quit: Connection reset by peer).
(06:00:49) eno_ [n=eno@adsl-70-137-131-139.dsl.snfc21.sbcglobal.net] entered the room.
(06:02:45) eno left the room (quit: Nick collision from services.).
(06:02:53) eno_ is now known as eno
(06:12:49) maht left the room (quit: Read error: 110 (Connection timed out)).
(07:42:22) maht [n=maht__@85.189.31.174.proweb.managedbroadband.co.uk] entered the room.
(08:27:01) eno_ [n=eno@adsl-70-137-139-149.dsl.snfc21.sbcglobal.net] entered the room.
(08:39:00) eno left the room (quit: Read error: 110 (Connection timed out)).
(10:03:50) sea-gull [n=sea-gull@95-28-30-183.broadband.corbina.ru] entered the room.
(11:00:45) sea-gull left the room (quit: Read error: 110 (Connection timed out)).
(11:12:39) sea-gull [n=sea-gull@95-28-6-56.broadband.corbina.ru] entered the room.
(11:14:43) sea-gull_ [n=sea-gull@95-28-6-56.broadband.corbina.ru] entered the room.
(11:37:28) j1m1 left the room (quit: Read error: 60 (Operation timed out)).
(12:35:51) j1m1 [n=jim@unas-228.rsity.ru] entered the room.
(13:19:24) sea-gull_ left the room (quit: Read error: 60 (Operation timed out)).
(13:33:56) sea-gull left the room (quit: Read error: 110 (Connection timed out)).
(14:13:34) sea-gull [n=sea-gull@95-28-118-124.broadband.corbina.ru] entered the room.
(14:36:53) j1m1 left the room (quit: Read error: 110 (Connection timed out)).
(15:31:07) eno_ is now known as eno
(15:58:10) mennis [n=mennis@adsl-158-220-153.asm.bellsouth.net] entered the room.
(16:53:18) sea-gull left the room (quit: Read error: 113 (No route to host)).
(17:23:39) sea-gull [n=sea-gull@95-28-97-190.broadband.corbina.ru] entered the room.
(17:41:26) sea-gull_ [n=sea-gull@95-28-94-56.broadband.corbina.ru] entered the room.
(17:55:01) sea-gull left the room (quit: Connection timed out).
(18:11:02) eno_ [n=eno@adsl-70-137-139-149.dsl.snfc21.sbcglobal.net] entered the room.
(18:11:54) eno left the room (quit: Read error: 104 (Connection reset by peer)).
(18:16:03) gualteri left the room (quit: "leaving").
(18:25:10) sea-gull_ left the room (quit: Read error: 60 (Operation timed out)).
(18:53:00) eno_ is now known as eno
(21:22:11) paigeadele [i=erratic@75-146-55-252-Washington.hfc.comcastbusiness.net] entered the room.
(21:37:25) sea-gull [n=sea-gull@95-28-130-0.broadband.corbina.ru] entered the room.
(22:11:06) sea-gull: mjl-: hi. How can /module/test.m can include /usr/inferno/tables.m (this path can change)?
(22:13:54) sea-gull: I'm asking this I don't how to be with tables.[bm]. All other source files are located at /dis and /modules.
(22:51:53) sea-gull left the room (quit: Client Quit).
(23:41:24) j1m [n=jim@unas-228.rsity.ru] entered the room.
(23:48:59) j1m left the room.
(23:49:03) j1m [n=jim@unas-228.rsity.ru] entered the room.
(00:05:55) mennis left the room (quit: Client Quit).
(04:43:40) arcfide [n=arcfide@h-69-3-100-36.chcgilgm.dynamic.covad.net] entered the room.
(05:23:22) mennis [n=mennis@adsl-158-220-153.asm.bellsouth.net] entered the room.
(05:26:31) mennis left the room (quit: Client Quit).
(05:27:27) mennis [n=mennis@adsl-158-220-153.asm.bellsouth.net] entered the room.
(05:30:31) mennis left the room (quit: Client Quit).
(05:40:00) mennis [n=mennis@adsl-158-220-153.asm.bellsouth.net] entered the room.
(06:15:56) mennis left the room (quit: Client Quit).
(09:25:59) Fish [n=SPARCman@AVelizy-152-1-20-75.w82-120.abo.wanadoo.fr] entered the room.
(09:48:20) arcfide left the room.
(10:37:07) yiyus left the room (quit: "leaving").
(10:37:40) j1m left the room.
(11:57:34) Fish left the room (quit: Read error: 110 (Connection timed out)).
(12:01:57) gualteri [n=unknown@crespins.disca.upv.es] entered the room.
(12:04:11) Fish [n=SPARCman@AVelizy-152-1-20-75.w82-120.abo.wanadoo.fr] entered the room.
(12:36:28) sea-gull [n=sea-gull@95-28-27-46.broadband.corbina.ru] entered the room.
(13:13:53) yiyus [n=12427124@je.je.je] entered the room.
(13:34:20) Fish left the room (quit: Remote closed the connection).
(13:55:10) sea-gull left the room (quit: Read error: 60 (Operation timed out)).