[9fans] rfc / internet-draft viewer
I cleaned up my RFC/I-D viewer and mirroring tools and pushed them up to /n/9pio/contrib/lyndon/rfc.tar. They're a bit more functional than the existing /lib/rfc/grabrfc, and interface nicely with the plumber. Note that I have an /rc/bin/aux directory that I 'bind -a' to /bin/aux in my global namespace. If you don't want to do that binding you'll have to fiddle the mkfile a bit. Otherwise, add 'bind -a /rc/bin/aux /bin/aux' to /lib/namespace to make the mirror commands visible. RFC(1) RFC(1) NAME rfc, idmirror, rfcmirror - Display RFCs and Internet Drafts SYNOPSIS rfc [-p] docref rfc [-p] -dIis aux/idmirror [-v] aux/rfcmirror [-v] DESCRIPTION rfc displays IETF RFCs and Internet Drafts. docref is a plain integer in the case of RFCs, or a string of the form draft-* for Internet Drafts. rfc can also display indexes of drafts, RFCs, and STDs by specifying one of the following flags: -d display the Internet Drafts index -I display the index of recent RFCs -i display the unabridged RFC index -s display the list of STDs rfc invokes B to display the document; specifying -p instead prints the document on the standard output. idmirror and rfcmirror maintain the local mirrors of the IETF document repositories. The -v flag makes them print the names of documents added or removed from the local mir- ror. Only the text versions of the documents are mirrored. rfcmirror also puts a copy of the recent-RFCs index into /lib/news/latest_rfcs for use by news(1). Plumbing Adding the following to $home/lib/plumbing (ahead of the include basic line) makes RFC and Internet Draft references plumbable: type is text data matches '[Rr][Ff][Cc][ ]*([0-9]+)(.[Tt][Xx][Tt])?' plumb start rfc $1 type is text data matches '(draft-[a-z0-9-]+-[0-9]+)(.txt)?' plumb start rfc $1 FILES /lib/doc/ietf the local document mirror. BUGS The internet drafts FTP server (ftp.ietf.org) has an absurdly short command channel timeout. If you have a slow network connection this can cause idmirror to fail, report- ing an ftpfs rpc error. The problem is with the IETF's FTP server, not idmirror. -- 9fans: 9fans Permalink: https://9fans.topicbox.com/groups/9fans/Td6498de610f4eaa9-M47ba54071b9de1bbb3f787fa Delivery options: https://9fans.topicbox.com/groups/9fans/subscription
[9fans] Re: [9front] dropped emails
hiro writes: > only found out by accident, Not sure what's going on. I sent a couple of messages to 9front and 9front-bugs yesterday that vanished into a black hole ... --lyndon -- 9fans: 9fans Permalink: https://9fans.topicbox.com/groups/9fans/T75f937755279e2c1-Me40aee945910e9d382aed65f Delivery options: https://9fans.topicbox.com/groups/9fans/subscription
[9fans] news(1) patch: make -a and -n get along
[ Originally send to 9front-bugs, but this is applicable across the board ... ] This patch makes 'news -an' do the right thing. /n/dump/2021/0122/sys/src/cmd/news.c:44,65 - news.c:44,72 void main(int argc, char *argv[]) { - int i; + int i, aflag = 0, nflag = 0; + int doupdate = 1; + int printall = 0; + void (*printer)(char*) = print_item; Binit(&bout, 1, OWRITE); if(argc == 1) { - eachitem(print_item, 0, 1); + eachitem(print_item, printall, doupdate); exits(0); } ARGBEGIN{ case 'a': /* print all */ - eachitem(print_item, 1, 0); + doupdate = 0; + printall = 1; + // eachitem(print_item, 1, 0); break; case 'n': /* names only */ - eachitem(note, 0, 0); - if(n_items) - Bputc(&bout, '\n'); + doupdate = 0; + printer = note; + // eachitem(note, 0, 0); + // if(n_items) + // Bputc(&bout, '\n'); break; default: /n/dump/2021/0122/sys/src/cmd/news.c:66,73 - news.c:73,87 fprint(2, "news: bad option %c\n", ARGC()); exits("usage"); }ARGEND - for(i=0; ihttps://9fans.topicbox.com/groups/9fans/Td990983e87321183-M486341df8c5b1a9191d4c728 Delivery options: https://9fans.topicbox.com/groups/9fans/subscription
Re: [9fans] Re: [9front] dropped emails
Quoth Lyndon Nerenberg : > hiro writes: > > only found out by accident, > > Not sure what's going on. I sent a couple of messages to > 9front and 9front-bugs yesterday that vanished into a black > hole ... > > --lyndon As far as we can tell, they're currently in the queue, which is currently running at a crawl -- we're aware of the problem, and are poking at upas to figure out what's up. -- 9fans: 9fans Permalink: https://9fans.topicbox.com/groups/9fans/T75f937755279e2c1-M6fc5e69b078b82696be8dfc7 Delivery options: https://9fans.topicbox.com/groups/9fans/subscription
Re: [9fans] authoritative source for u9fs?
I've just accepted an important pull request to it that I missed, until I looked at it in response to this, so you should do another pull to get that. Generally, though it has been reasonably stable for some time. Changes are often just to cope with this year's #ifdefs or feature defines. On Fri, Jan 22, 2021 at 9:17 AM Dworkin Muller wrote: > On Fri, 22 Jan 2021 03:20:03 -0500, Fazlul Shahriar > wrote: > fshahriar> https://bitbucket.org/plan9-from-bell-labs/u9fs/src/master/ > > That looks much more believable than the others. Thank you very much > for the pointer. > > Dworkin -- 9fans: 9fans Permalink: https://9fans.topicbox.com/groups/9fans/Tee2220301f2a891c-M74684f2bf32f35f0b7c76a2d Delivery options: https://9fans.topicbox.com/groups/9fans/subscription
Re: [9fans] authoritative source for u9fs?
it's also on bitbucket not github mainly for historical reasons but I also can never decide which I dislike more. Originally, it was on Google Source and bitbucket had hg, so I used that. Now that they are all gits, perhaps there's no difference. At one time I thought that Atlassian did say they backed it up, which Github suggested wasn't needed because of the remote archives, but that might be outdated. On Sun, Jan 24, 2021 at 11:30 PM Charles Forsyth wrote: > I've just accepted an important pull request to it that I missed, until I > looked at it in response to this, so you should do another pull to get that. > Generally, though it has been reasonably stable for some time. Changes are > often just to cope with this year's #ifdefs or feature defines. > > > On Fri, Jan 22, 2021 at 9:17 AM Dworkin Muller > wrote: > >> On Fri, 22 Jan 2021 03:20:03 -0500, Fazlul Shahriar >> wrote: >> fshahriar> https://bitbucket.org/plan9-from-bell-labs/u9fs/src/master/ >> >> That looks much more believable than the others. Thank you very much >> for the pointer. >> >> Dworkin -- 9fans: 9fans Permalink: https://9fans.topicbox.com/groups/9fans/Tee2220301f2a891c-M0fd0a9e2ffb13d78e01af2c7 Delivery options: https://9fans.topicbox.com/groups/9fans/subscription
Re: [9fans] authoritative source for u9fs?
Charles Forsyth writes: > it's also on bitbucket not github mainly for historical reasons but I also > can never decide which I dislike more. :-) The nice thing about having it in hg is that mercurial is part of 9front, so there's no need to muck about getting git installed. --lyndon -- 9fans: 9fans Permalink: https://9fans.topicbox.com/groups/9fans/Tee2220301f2a891c-Mfba96c69e2d6c1b21c1a81d4 Delivery options: https://9fans.topicbox.com/groups/9fans/subscription
Re: [9fans] authoritative source for u9fs?
Yes but bitbucket ditched hg so its now all git On Mon, Jan 25, 2021 at 12:10 AM Lyndon Nerenberg wrote: > Charles Forsyth writes: > > > it's also on bitbucket not github mainly for historical reasons but I > also > > can never decide which I dislike more. > :-) > > The nice thing about having it in hg is that mercurial is part of > 9front, so there's no need to muck about getting git installed. > > --lyndon > -- 9fans: 9fans Permalink: https://9fans.topicbox.com/groups/9fans/Tee2220301f2a891c-Mf163251a751866fcf764cc32 Delivery options: https://9fans.topicbox.com/groups/9fans/subscription
Re: [9fans] Re: [9front] dropped emails
Quoth o...@eigenstate.org: > Quoth Lyndon Nerenberg : > > hiro writes: > > > only found out by accident, > > > > Not sure what's going on. I sent a couple of messages to > > 9front and 9front-bugs yesterday that vanished into a black > > hole ... > > > > --lyndon > > As far as we can tell, they're currently in > the queue, which is currently running at a > crawl -- we're aware of the problem, and are > poking at upas to figure out what's up. As far as we know, the mails directed at working addresses should be sent now. We're now running a hacked version of runq that should be processing messages in parallel, but without clobbering the system with load. In case of mailing list flakiness, please complain -- and put me in the cc list. If you do not get this message, please report it at once. -- 9fans: 9fans Permalink: https://9fans.topicbox.com/groups/9fans/T75f937755279e2c1-M0895c0835dc165995d228528 Delivery options: https://9fans.topicbox.com/groups/9fans/subscription
Re: [9fans] authoritative source for u9fs?
On 1/25/21, Charles Forsyth wrote: > I've just accepted an important pull request to it that I missed, until I > looked at it in response to this, so you should do another pull to get > that. > Generally, though it has been reasonably stable for some time. Changes are > often just to cope with this year's #ifdefs or feature defines. > Under NetBSD 9.1, I needed to add -D_NETBSD_SOURCE to get rid of warnings for undefined initgroups() and ruserok(). The compiler seem to recommend getgroups() and cuserid() as alternatives. I'm not sure how best to deal with this for other platforms. Lucio. PS: The new executable seems noticeably bigger than whatever I used previously. Which happens to be the NetBSD "pkg" version (/usr/pkgsrc/filesysytems/u9fs). Hmm, the differences are quite significant... -- 9fans: 9fans Permalink: https://9fans.topicbox.com/groups/9fans/Tee2220301f2a891c-M5d92402e11109bf407c47b81 Delivery options: https://9fans.topicbox.com/groups/9fans/subscription
Re: [9fans] authoritative source for u9fs?
On Mon, 25 Jan 2021 07:10:18 +0200, Lucio De Re wrote: lucio.dere> On 1/25/21, Charles Forsyth wrote: lucio.dere> > I've just accepted an important pull request to it that I missed, until I lucio.dere> > [...] lucio.dere> Under NetBSD 9.1, I needed to add -D_NETBSD_SOURCE to get rid of lucio.dere> warnings for undefined initgroups() and ruserok(). The compiler seem lucio.dere> to recommend getgroups() and cuserid() as alternatives. I'm not sure lucio.dere> how best to deal with this for other platforms. Charles, with the version I pulled down on Friday, I had to change the start of plan9.h to get it to build cleanly on FreeBSD 12.2: #ifdef __APPLE__ #define _DARWIN_C_SOURCE #elif !defined(__FreeBSD__) /* !defined added by Dworkin */ /* magic to get SUSV2 standard, including pread, pwrite*/ #define _XOPEN_SOURCE 500 #endif Defining _XOPEN_SOURCE on FreeBSD 12 suppresses the visibility of a goodly amount of the BSD-related calls, including ruserok(). Dworkin -- 9fans: 9fans Permalink: https://9fans.topicbox.com/groups/9fans/Tee2220301f2a891c-Md9caea9503919914ff24231f Delivery options: https://9fans.topicbox.com/groups/9fans/subscription
[9fans] getting 9front ssh to use RSA key?
If I have PasswordAuthentication enabled on a remote host (tested on MacOS and FreeBSD so far), I can log in to them without any problem. However, if I have passwords disabled, but have an RSA key on the Plan 9 host and the corresponding pub key in authorized_keys on those remote hosts, I'm failing to log in with the error message: ssh: auth: no key matches proto=rsa service=ssh role=client Presumably, this means I haven't set something up somewhere. Currently, I do not have an auth server - I'm doing everything from a terminal, slowly working my way up to a full world. Is there a way to make this work in such an environment, without jumping through more hoops than getting an auth server going would take? Thanks. Dworkin -- 9fans: 9fans Permalink: https://9fans.topicbox.com/groups/9fans/Td072863a97c9d3e9-M41815fdb9ca0d8fb85abfa9a Delivery options: https://9fans.topicbox.com/groups/9fans/subscription
Re: [9fans] authoritative source for u9fs?
Quoth Lyndon Nerenberg : > Charles Forsyth writes: > > > it's also on bitbucket not github mainly for historical reasons but I also > > can never decide which I dislike more. > :-) > > The nice thing about having it in hg is that mercurial is part of > 9front, so there's no need to muck about getting git installed. > > --lyndon Don't worry, git's on the way too. there's interest in migrating 9front, so we'll need to ship with it. -- 9fans: 9fans Permalink: https://9fans.topicbox.com/groups/9fans/Tee2220301f2a891c-Mf7cb93d2725b2ec85e9107f4 Delivery options: https://9fans.topicbox.com/groups/9fans/subscription
Re: [9fans] getting 9front ssh to use RSA key?
Quoth Dworkin Muller : > If I have PasswordAuthentication enabled on a remote host (tested on > MacOS and FreeBSD so far), I can log in to them without any problem. > However, if I have passwords disabled, but have an RSA key on the Plan 9 > host and the corresponding pub key in authorized_keys on those remote > hosts, I'm failing to log in with the error message: > >ssh: auth: no key matches proto=rsa service=ssh role=client > > Presumably, this means I haven't set something up somewhere. > Currently, I do not have an auth server - I'm doing everything from a > terminal, slowly working my way up to a full world. Is there a way to > make this work in such an environment, without jumping through more > hoops than getting an auth server going would take? > > Thanks. > > Dworkin there's an example in the rsa(8) manpage. -- 9fans: 9fans Permalink: https://9fans.topicbox.com/groups/9fans/Td072863a97c9d3e9-M07ee0b71cb7faf20cff1fe30 Delivery options: https://9fans.topicbox.com/groups/9fans/subscription
Re: [9fans] getting 9front ssh to use RSA key?
On Sun, 24 Jan 2021 22:33:59 -0800, o...@eigenstate.org wrote: ori> Quoth Dworkin Muller : ori> > If I have PasswordAuthentication enabled on a remote host (tested on ori> > MacOS and FreeBSD so far), I can log in to them without any problem. ori> > However, if I have passwords disabled, but have an RSA key on the Plan 9 ori> > host and the corresponding pub key in authorized_keys on those remote ori> > hosts, I'm failing to log in with the error message: ori> > ori> >ssh: auth: no key matches proto=rsa service=ssh role=client ori> > ori> > Presumably, this means I haven't set something up somewhere. ori> > Currently, I do not have an auth server - I'm doing everything from a ori> > terminal, slowly working my way up to a full world. Is there a way to ori> > make this work in such an environment, without jumping through more ori> > hoops than getting an auth server going would take? ori> > ori> > Thanks. ori> > ori> > Dworkin ori> ori> there's an example in the rsa(8) manpage. That's what I thought I'd been doing, and doing it again just now gives the same results. ``cat /mnt/factotum/ctl'' gives two lines, one starting ``key proto=pass server=lethe service=ssh ...'' and the other ``key proto=rsa service=ssh ...''. I've triple checked that what's in .authorized_keys on the remote host (lethe) matches the output of auth/rsa2ssh. I've also verified that the proto=rsa line in factotum matches my lib/ssh/rsa file. Thus my confusion. Thanks. Dworkin -- 9fans: 9fans Permalink: https://9fans.topicbox.com/groups/9fans/Td072863a97c9d3e9-M877d35325ba8c613afebac3c Delivery options: https://9fans.topicbox.com/groups/9fans/subscription