From ocurtin@NOSPAM.usa.net Tue Mar 7 13:29:31 2000 Path: transfer.stratus.com!cambridge1-snf1.gtei.net!news.gtei.net!bos-service1.ext .raytheon.com!cyclone.swbell.net!newsfeed.berkeley.edu!newsfeed.nyu.edu!wesl ey.videotron.net!wagner.videotron.net.POSTED!ocurtin From: ocurtin@NOSPAM.usa.net (Curly++) Newsgroups: comp.terminals,comp.os.linux.admin Subject: Escape sequences for xterm? Organization: The Minetown Digger Reply-To: ocurtin@NOSPAM.usa.net Message-ID: User-Agent: slrn/0.9.5.4 (UNIX) Lines: 33 Date: Fri, 3 Mar 2000 15:08:59 -0500 NNTP-Posting-Host: 24.200.125.165 X-Complaints-To: abuse@videotron.net X-Trace: wagner.videotron.net 952113711 24.200.125.165 (Fri, 03 Mar 2000 15:01:51 EST) NNTP-Posting-Date: Fri, 03 Mar 2000 15:01:51 EST Xref: transfer.stratus.com comp.terminals:13804 I think what I need is a listing of xterm escape codes. All the listings I've found are horribly out of date. I've done a few extensive web crawls but haven't been able to find anything up-to-date except the source code... yeah, I know, but it's self-documenting! Maybe I'm going about this all wrong, so let me tell you what I'm doing, you can tell me if I'm stupidly ignoring an obvious and simple solution. I'm fiddling with the source for xterm. I need to monitor the user's interaction with some applications so as to divert some of the inputs to another app. Legacy stuff, this is a kludge to bridge the old and the new until the old can be fully phased out. I'm writing scripts with `expect` to wait for the user to fill in a form, then take an image of the screen and pass it to another app. I can't capture the data coming back, the screen is updated in tiny bits sort-of like what curses does. I've added an escape sequence to xterm so the expect script can be triggered by the few predictable strings and use `send_user` to trigger a screen dump to an external file. I'd like to use a "correct" sequence and submit this as a patch to the xterm maintainer. This seems to me a general solution to a set of problems that turn up occassionaly. -- Oisin "Curly++" Curtin ocurtin@SPAM.usa.net Surface Liaison, Minetown Digger Send no SPAM. http://pages.infinit.net/curlypp/ From m.ramsch@computer.org Tue Mar 7 13:29:35 2000 Path: transfer.stratus.com!cambridge1-snf1.gtei.net!news.gtei.net!bos-service1.ext .raytheon.com!cyclone.swbell.net!newsfeed.berkeley.edu!news-FFM2.ecrc.net!ne ws-fra.pop.de!news.csl-gmbh.net!informatik.tu-muenchen.de!lrz.de!news.rz.uni -passau.de!m.ramsch From: m.ramsch@computer.org (Martin Ramsch) Newsgroups: comp.terminals,comp.os.linux.admin Subject: Re: Escape sequences for xterm? Date: 3 Mar 2000 22:51:21 +0100 Organization: [posted via] Leibniz-Rechenzentrum, Muenchen (Germany) Lines: 19 Message-ID: References: NNTP-Posting-Host: news.rz.uni-passau.de X-Trace: wsc10.lrz-muenchen.de 952120277 12810 132.231.51.57 (3 Mar 2000 21:51:17 GMT) X-Complaints-To: news@lrz-muenchen.de NNTP-Posting-Date: 3 Mar 2000 21:51:17 GMT Mail-Copies-To: always User-Agent: slrn/0.9.6.2 (SunOS) X-Original-NNTP-Posting-Host: 132.231.20.19 Xref: transfer.stratus.com comp.terminals:13805 On Fri, 3 Mar 2000 15:08:59 -0500, Oisin "Curly++" Curtin wrote: > I think what I need is a listing of xterm escape codes. All the > listings I've found are horribly out of date. Did you also have a look at the file "ctlseqs.ms" which is part of every xterm packages? To convert this file into other formats use for example groff -ms -t -Tps ctlseqs.ms >ctlseqs.ps for Postscript output groff -ms -t -Tlj4 ctlseqs.ms >ctlseqs.lj4 for LaserJet 4 print file groff -ms -t -Tdvi ctlseqs.ms >ctlseqs.dvi for TeX dvi output groff -ms -t -Tlatin1 ctlseqs.ms >ctlseqs.txt for text output Regards, Martin -- Martin Ramsch PGP KeyID=0xE8EF4F75 FiPr=5244 5EF3 B0B1 3826 E4EC 8058 7B31 3AD7 From dickey@shell.clark.net Tue Mar 7 13:29:42 2000 Path: transfer.stratus.com!cambridge1-snf1.gtei.net!news.gtei.net!news.shore.net!n ewsfeed.mathworks.com!europa.netcrusader.net!205.252.116.205!howland.erols.n et!peer.news.verio.net.MISMATCH!iad-peer.news.verio.net!news.verio.net!iad-r ead.news.verio.net.POSTED!not-for-mail From: "T.E.Dickey" Subject: Re: Escape sequences for xterm? Newsgroups: comp.terminals,comp.os.linux.admin References: Organization: Clark Internet Services, Inc., Ellicott City, MD USA User-Agent: tin/1.5.2-20000119 ("Sumerland") (UNIX) (SunOS/5.6 (sun4u)) Lines: 24 Message-ID: Date: Sat, 04 Mar 2000 11:37:52 GMT NNTP-Posting-Host: 168.143.0.8 X-Complaints-To: abuse@verio.net X-Trace: iad-read.news.verio.net 952169872 168.143.0.8 (Sat, 04 Mar 2000 11:37:52 GMT) NNTP-Posting-Date: Sat, 04 Mar 2000 11:37:52 GMT Xref: transfer.stratus.com comp.terminals:13807 Martin Ramsch wrote: > On Fri, 3 Mar 2000 15:08:59 -0500, > Oisin "Curly++" Curtin wrote: >> I think what I need is a listing of xterm escape codes. All the >> listings I've found are horribly out of date. > Did you also have a look at the file "ctlseqs.ms" which is part of > every xterm packages? Technically, ctlseqs.ms was moved to the doc directory tree some time ago - depending on how you get xterm, you may not see it - I bundle it in my xterm tarballs for convenience. The XFree86 xterm supports ANSI color and VT220 emulation There's an faq at http://www.clark.net/pub/dickey/xterm/xterm.faq.html ftp://ftp.clark.net/pub/dickey/xterm (altavista usually shows my webpage - though I don't search for that ;-) -- Thomas E. Dickey dickey@clark.net http://www.clark.net/pub/dickey From dickey@shell.clark.net Tue Mar 7 13:29:46 2000 Path: transfer.stratus.com!cambridge1-snf1.gtei.net!news.gtei.net!bloom-beacon.mit .edu!howland.erols.net!peer.news.verio.net.MISMATCH!iad-peer.news.verio.net! news.verio.net!iad-read.news.verio.net.POSTED!not-for-mail From: "T.E.Dickey" Subject: Re: Escape sequences for xterm? Newsgroups: comp.terminals,comp.os.linux.admin References: Organization: Clark Internet Services, Inc., Ellicott City, MD USA User-Agent: tin/1.5.2-20000119 ("Sumerland") (UNIX) (SunOS/5.6 (sun4u)) Lines: 16 Message-ID: Date: Sat, 04 Mar 2000 14:39:17 GMT NNTP-Posting-Host: 168.143.0.8 X-Complaints-To: abuse@verio.net X-Trace: iad-read.news.verio.net 952180757 168.143.0.8 (Sat, 04 Mar 2000 14:39:17 GMT) NNTP-Posting-Date: Sat, 04 Mar 2000 14:39:17 GMT Xref: transfer.stratus.com comp.terminals:13809 Curly++ wrote: > I've added an escape sequence to xterm so the expect script can > be triggered by the few predictable strings and use `send_user` > to trigger a screen dump to an external file. I'd like to use a > "correct" sequence and submit this as a patch to the xterm > maintainer. This seems to me a general solution to a set of > problems that turn up occassionaly. While it's useful, this is generally frowned upon as a security problem. I don't think it'll get much general acceptance. -- Thomas E. Dickey dickey@clark.net http://www.clark.net/pub/dickey From celigne@tinyworld.co.uk Tue Mar 7 13:29:49 2000 Message-ID: <38C15701.7CF1A007@tinyworld.co.uk> Date: Sat, 04 Mar 2000 18:33:37 +0000 From: Paul Williams X-Mailer: Mozilla 4.6 [en] (Win95; I) X-Accept-Language: en MIME-Version: 1.0 Newsgroups: comp.terminals,comp.os.linux.admin Subject: Re: Escape sequences for xterm? References: Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit NNTP-Posting-Host: modem-88-4-60-62.vip.uk.com X-Trace: 4 Mar 2000 18:33:07 GMT, modem-88-4-60-62.vip.uk.com Lines: 16 X-Report: Report abuse to abuse@vip.uk.com Path: transfer.stratus.com!cambridge1-snf1.gtei.net!news.gtei.net!inmet!news3!logb ridge.uoregon.edu!newsfeed.icl.net!newsfeed.icl.net!iclnet!news1.vip.uk.com! modem-88-4-60-62.vip.uk.com Xref: transfer.stratus.com comp.terminals:13811 "T.E.Dickey" wrote: > > Curly++ wrote: > > > I've added an escape sequence to xterm so the expect script can > > be triggered by the few predictable strings and use `send_user` > > to trigger a screen dump to an external file. I'd like to use a > > "correct" sequence and submit this as a patch to the xterm > > maintainer. This seems to me a general solution to a set of > > problems that turn up occassionaly. > > While it's useful, this is generally frowned upon as a security problem. > I don't think it'll get much general acceptance. It is possible for a host computer to grab the contents of the screen of a Digital VT420. Howeve, xterm doesn't emulate a VT420 (yet?) From ocurtin@NOSPAM.usa.net Tue Mar 7 13:29:53 2000 Path: transfer.stratus.com!cambridge1-snf1.gtei.net!news.gtei.net!bloom-beacon.mit .edu!news-peer.gip.net!news.gsl.net!gip.net!feeder.qis.net!newsfeed.nyu.edu! wesley.videotron.net!wagner.videotron.net.POSTED!ocurtin From: ocurtin@NOSPAM.usa.net (Curly++) Newsgroups: comp.terminals,comp.os.linux.admin Subject: Re: Escape sequences for xterm? References: Organization: The Minetown Digger Reply-To: ocurtin@NOSPAM.usa.net Message-ID: User-Agent: slrn/0.9.5.4 (UNIX) Lines: 21 Date: Sun, 5 Mar 2000 15:32:03 -0500 NNTP-Posting-Host: 24.200.125.165 X-Complaints-To: abuse@videotron.net X-Trace: wagner.videotron.net 952288894 24.200.125.165 (Sun, 05 Mar 2000 15:41:34 EST) NNTP-Posting-Date: Sun, 05 Mar 2000 15:41:34 EST Xref: transfer.stratus.com comp.terminals:13817 On 3 Mar 2000 22:51:21 +0100, Martin Ramsch wrote: > On Fri, 3 Mar 2000 15:08:59 -0500, > Oisin "Curly++" Curtin wrote: > > I think what I need is a listing of xterm escape codes. All the > > listings I've found are horribly out of date. > > Did you also have a look at the file "ctlseqs.ms" which is part of > groff -ms -t -Tlatin1 ctlseqs.ms >ctlseqs.txt for text output Ah! Thanks, Martin. That's the answer all right! I'm not a gruff person, er, I mean I'm not a groff person so the .ms didn't mean anthing to me. I'll suggest that your answer be added to the FAQ for all us ignoramuses. :-) -- Oisin "Curly++" Curtin ocurtin@SPAM.usa.net Surface Liaison, Minetown Digger Send no SPAM. http://pages.infinit.net/curlypp/ From ocurtin@NOSPAM.usa.net Tue Mar 7 13:29:57 2000 Path: transfer.stratus.com!cambridge1-snf1.gtei.net!news.gtei.net!bloom-beacon.mit .edu!newsswitch.lcs.mit.edu!news-spur1.maxwell.syr.edu!news.maxwell.syr.edu! newsfeed.nyu.edu!wesley.videotron.net!wagner.videotron.net.POSTED!ocurtin From: ocurtin@NOSPAM.usa.net (Curly++) Newsgroups: comp.terminals,comp.os.linux.admin Subject: Re: Escape sequences for xterm? References: Organization: The Minetown Digger Reply-To: ocurtin@NOSPAM.usa.net Message-ID: User-Agent: slrn/0.9.5.4 (UNIX) Lines: 24 Date: Sun, 5 Mar 2000 15:39:42 -0500 NNTP-Posting-Host: 24.200.125.165 X-Complaints-To: abuse@videotron.net X-Trace: wagner.videotron.net 952288895 24.200.125.165 (Sun, 05 Mar 2000 15:41:35 EST) NNTP-Posting-Date: Sun, 05 Mar 2000 15:41:35 EST Xref: transfer.stratus.com comp.terminals:13818 On Sat, 04 Mar 2000 11:37:52 GMT, T.E.Dickey wrote: > Martin Ramsch wrote: > > On Fri, 3 Mar 2000 15:08:59 -0500, > > Oisin "Curly++" Curtin wrote: > >> I think what I need is a listing of xterm escape codes. All the > >> listings I've found are horribly out of date. > > > Did you also have a look at the file "ctlseqs.ms" which is part of > > every xterm packages? > > Technically, ctlseqs.ms was moved to the doc directory tree some time > ago - depending on how you get xterm, you may not see it - I bundle it I must have an old copy then. AFAIK, I did check the lastest FAQ at clark. Can MR's reply be added so "us unwashed masses" can cut'n'paste the right command to see it? TIA, either way. -- Oisin "Curly++" Curtin ocurtin@SPAM.usa.net Surface Liaison, Minetown Digger Send no SPAM. http://pages.infinit.net/curlypp/ From ocurtin@NOSPAM.usa.net Tue Mar 7 13:30:01 2000 Path: transfer.stratus.com!cambridge1-snf1.gtei.net!news.gtei.net!inmet!news3!logb ridge.uoregon.edu!newsfeed.direct.ca!wesley.videotron.net!weber.videotron.ne t.POSTED!ocurtin From: ocurtin@NOSPAM.usa.net (Curly++) Newsgroups: comp.terminals,comp.os.linux.admin Subject: Re: Escape sequences for xterm? References: Organization: The Minetown Digger Reply-To: ocurtin@NOSPAM.usa.net Message-ID: User-Agent: slrn/0.9.5.4 (UNIX) Lines: 44 Date: Sun, 5 Mar 2000 16:15:02 -0500 NNTP-Posting-Host: 24.200.125.165 X-Complaints-To: abuse@videotron.net X-Trace: weber.videotron.net 952291293 24.200.125.165 (Sun, 05 Mar 2000 16:21:33 EST) NNTP-Posting-Date: Sun, 05 Mar 2000 16:21:33 EST Xref: transfer.stratus.com comp.terminals:13819 On Sat, 04 Mar 2000 14:39:17 GMT, T.E.Dickey wrote: > Curly++ wrote: > > > I've added an escape sequence to xterm so the expect script can > > be triggered by the few predictable strings and use `send_user` > > to trigger a screen dump to an external file. I'd like to use a > > "correct" sequence and submit this as a patch to the xterm > > maintainer. This seems to me a general solution to a set of > > problems that turn up occassionaly. > > While it's useful, this is generally frowned upon as a security problem. > I don't think it'll get much general acceptance. DISCLAIMER: I'm not a security expert and I don't play one on TV. You're referring to the log problem? I haven't read all about it, but I gather it's not logging per se, but the fact that the destination file is uncontrolled that causes a problem. I plan to make screen-scraping secure by allowing the user only limited control of the destination filename and ignoring the request if the file already exists. That means the screen-scrape will never, for example, add a line to ~/.profile. This forces the external script to delete files, but then the script needs much less logic to read the image that way. I'm not sure, but I think the solution to the logging problem is similar: restrict the name and location of the file to something less than anywhere/anything and use more elaborate fstat code in place of the access tests. Maybe even put a magic cookie at the top of new logs and never append to an existing log if it doesn't have a cookie on top. A cookie is a lot of extra code though, I wouldn't do it unless the real security experts say it's needed. BTW, I have to add this... My partner is not used to seeing me grinning from ear to ear and muttering under my breath "Wow! That's neat!" and similar things. This code is an absolute _pleasure_ to work with. -- Oisin "Curly++" Curtin ocurtin@SPAM.usa.net Surface Liaison, Minetown Digger Send no SPAM. http://pages.infinit.net/curlypp/ From ocurtin@NOSPAM.usa.net Tue Mar 7 13:30:05 2000 Path: transfer.stratus.com!cambridge1-snf1.gtei.net!news.gtei.net!inmet!news3!arcl ight.uoregon.edu!vixen.cso.uiuc.edu!howland.erols.net!news-peer.gip.net!news .gsl.net!gip.net!feeder.qis.net!newsfeed.nyu.edu!wesley.videotron.net!wagner .videotron.net.POSTED!ocurtin From: ocurtin@NOSPAM.usa.net (Curly++) Newsgroups: comp.terminals,comp.os.linux.admin Subject: Re: Escape sequences for xterm? References: <38C15701.7CF1A007@tinyworld.co.uk> Organization: The Minetown Digger Reply-To: ocurtin@NOSPAM.usa.net Message-ID: User-Agent: slrn/0.9.5.4 (UNIX) Lines: 15 Date: Sun, 5 Mar 2000 16:32:52 -0500 NNTP-Posting-Host: 24.200.125.165 X-Complaints-To: abuse@videotron.net X-Trace: wagner.videotron.net 952292493 24.200.125.165 (Sun, 05 Mar 2000 16:41:33 EST) NNTP-Posting-Date: Sun, 05 Mar 2000 16:41:33 EST Xref: transfer.stratus.com comp.terminals:13820 On 04 Mar 2000, Paul Williams wrote: > It is possible for a host computer to grab the contents of the screen > of a Digital VT420. Howeve, xterm doesn't emulate a VT420 (yet?) But doesn't that go back to the remote application? I think `expect` could divert the output, but having dealt with timing issues before, I feel more comfortable if the data *cannot* go out the wrong way no matter how hard it tries. -- Oisin "Curly++" Curtin ocurtin@SPAM.usa.net Surface Liaison, Minetown Digger Send no SPAM. http://pages.infinit.net/curlypp/ From dickey@shell.clark.net Tue Mar 7 13:30:09 2000 Path: transfer.stratus.com!cambridge1-snf1.gtei.net!news.gtei.net!homer.alpha.net! solaris.cc.vt.edu!news.vt.edu!newsfeed.berkeley.edu!newsfeed.usit.net!feed2. onemain.com!feed1.onemain.com!nuq-peer.news.verio.net!news.verio.net!iad-rea d.news.verio.net.POSTED!not-for-mail From: "T.E.Dickey" Subject: Re: Escape sequences for xterm? Newsgroups: comp.terminals,comp.os.linux.admin References: Organization: Clark Internet Services, Inc., Ellicott City, MD USA User-Agent: tin/1.5.2-20000119 ("Sumerland") (UNIX) (SunOS/5.6 (sun4u)) Lines: 17 Message-ID: Date: Sun, 05 Mar 2000 23:55:17 GMT NNTP-Posting-Host: 168.143.0.8 X-Complaints-To: abuse@verio.net X-Trace: iad-read.news.verio.net 952300517 168.143.0.8 (Sun, 05 Mar 2000 23:55:17 GMT) NNTP-Posting-Date: Sun, 05 Mar 2000 23:55:17 GMT Xref: transfer.stratus.com comp.terminals:13821 Curly++ wrote: >> Technically, ctlseqs.ms was moved to the doc directory tree some time >> ago - depending on how you get xterm, you may not see it - I bundle it > I must have an old copy then. > AFAIK, I did check the lastest FAQ at clark. Can MR's reply be > added so "us unwashed masses" can cut'n'paste the right command > to see it? TIA, either way. maybe (I've been more busy fixing bugs in various programs recently than updating webpages). -- Thomas E. Dickey dickey@clark.net http://www.clark.net/pub/dickey From ocurtin@NOSPAM.usa.net Tue Mar 7 13:30:12 2000 Path: transfer.stratus.com!cambridge1-snf1.gtei.net!cam-news-hub1.bbnplanet.com!ne ws.gtei.net!hermes.visi.com!news-out.visi.com!wesley.videotron.net!wagner.vi deotron.net.POSTED!ocurtin From: ocurtin@NOSPAM.usa.net (Curly++) Newsgroups: comp.terminals,comp.os.linux.admin Subject: Re: Escape sequences for xterm? References: Organization: The Minetown Digger Reply-To: ocurtin@NOSPAM.usa.net Message-ID: User-Agent: slrn/0.9.5.4 (UNIX) Lines: 23 Date: Mon, 6 Mar 2000 12:37:00 -0500 NNTP-Posting-Host: 24.200.125.165 X-Complaints-To: abuse@videotron.net X-Trace: wagner.videotron.net 952364485 24.200.125.165 (Mon, 06 Mar 2000 12:41:25 EST) NNTP-Posting-Date: Mon, 06 Mar 2000 12:41:25 EST Xref: transfer.stratus.com comp.terminals:13827 In comp.terminals, Curly++ proclaimed: > DISCLAIMER: I'm not a security expert and I don't play one on TV. Geez, I'll never get that TV job. I went back to look at Morten Welinder and Branden Robinson's postings to bugtraq. Ya now, I don't even have to think to see how *machines* could get confused splitting the data stream between the local screen-image and the remote host. But Morten Welinder clearly is describing the same kind of timing issue, except the failure is in human ethical subroutines. I didn't see it until I'd read it three times. Such is the life of an optimist. I'm off to carefully read (at least three times!) the proposed logging patch. -- Oisin "Curly++" Curtin ocurtin@SPAM.usa.net Surface Liaison, Minetown Digger Send no SPAM. http://pages.infinit.net/curlypp/ From jaltman@watsun.cc.columbia.edu Tue Mar 7 13:30:16 2000 Path: transfer.stratus.com!cambridge1-snf1.gtei.net!news.gtei.net!inmet!news3!logb ridge.uoregon.edu!europa.netcrusader.net!144.212.100.101!newsfeed.mathworks. com!news.new-york.net!newsmaster.cc.columbia.edu!watsun.cc.columbia.edu!jalt man From: jaltman@watsun.cc.columbia.edu (Jeffrey Altman) Newsgroups: comp.terminals,comp.os.linux.admin Subject: Re: Escape sequences for xterm? Date: 6 Mar 2000 22:29:06 GMT Organization: Columbia University Lines: 36 Message-ID: <8a1bfi$sn5$1@newsmaster.cc.columbia.edu> References: NNTP-Posting-Host: watsun.cc.columbia.edu X-Trace: newsmaster.cc.columbia.edu 952381746 29413 128.59.39.2 (6 Mar 2000 22:29:06 GMT) X-Complaints-To: postmaster@columbia.edu NNTP-Posting-Date: 6 Mar 2000 22:29:06 GMT Xref: transfer.stratus.com comp.terminals:13829 In article , Curly++ wrote: : In comp.terminals, Curly++ proclaimed: : : > DISCLAIMER: I'm not a security expert and I don't play one on TV. : : Geez, I'll never get that TV job. : : I went back to look at Morten Welinder and Branden Robinson's : postings to bugtraq. Ya now, I don't even have to think to see : how *machines* could get confused splitting the data stream : between the local screen-image and the remote host. But Morten : Welinder clearly is describing the same kind of timing issue, : except the failure is in human ethical subroutines. I didn't see : it until I'd read it three times. Such is the life of an : optimist. : : I'm off to carefully read (at least three times!) the proposed : logging patch. There is no safe way for a host to specify a file name that should be used for logging, or printing, or any other kind of output. If host automated logging is the be provided it MUST be: . authorized by the client . to a file specified by the client This should not be done via a standard control sequence. If it is going to be done it should be performed via a private sequence (preferably not a '?' sequence) or via an APC. Jeffrey Altman * Sr.Software Designer * Kermit-95 for Win32 and OS/2 The Kermit Project * Columbia University 612 West 115th St #716 * New York, NY * 10025 http://www.kermit-project.org/k95.html * kermit-support@kermit-project.org From ocurtin@NOSPAM.usa.net Tue Mar 7 13:30:21 2000 Path: transfer.stratus.com!cambridge1-snf1.gtei.net!news.gtei.net!inmet!news3!logb ridge.uoregon.edu!newsfeed.nyu.edu!wesley.videotron.net!wagner.videotron.net .POSTED!ocurtin From: ocurtin@NOSPAM.usa.net (Curly++) Newsgroups: comp.terminals,comp.os.linux.admin Subject: Re: Escape sequences for xterm? References: <8a1bfi$sn5$1@newsmaster.cc.columbia.edu> Organization: The Minetown Digger Reply-To: ocurtin@NOSPAM.usa.net Message-ID: User-Agent: slrn/0.9.5.4 (UNIX) Lines: 115 Date: Mon, 6 Mar 2000 22:10:52 -0500 NNTP-Posting-Host: 24.200.125.165 X-Complaints-To: abuse@videotron.net X-Trace: wagner.videotron.net 952399281 24.200.125.165 (Mon, 06 Mar 2000 22:21:21 EST) NNTP-Posting-Date: Mon, 06 Mar 2000 22:21:21 EST Xref: transfer.stratus.com comp.terminals:13831 In comp.os.linux.admin, Jeffrey Altman proclaimed: > In article , > Curly++ wrote: > : In comp.terminals, Curly++ proclaimed: > : > : > DISCLAIMER: I'm not a security expert and I don't play one on TV. > : > : Geez, I'll never get that TV job. > : > : I went back to look at Morten Welinder and Branden Robinson's > : postings to bugtraq. Ya now, I don't even have to think to see > : how *machines* could get confused splitting the data stream > : between the local screen-image and the remote host. But Morten > : Welinder clearly is describing the same kind of timing issue, > : except the failure is in human ethical subroutines. I didn't see > : it until I'd read it three times. Such is the life of an > : optimist. > : > : I'm off to carefully read (at least three times!) the proposed > : logging patch. > > There is no safe way for a host to specify a file name that should > be used for logging, or printing, or any other kind of output. If It's a connundrum, all right. I started this thinking it would be a breeze. Making the part I needed was simple and straightforward. Fitting it into the configurator was a bit more work. But making it secure enough to distribute... this is tough! > host automated logging is the be provided it MUST be: > > . authorized by the client Yes. Authorization may be implicit, just as you "authorize" someone to modify your log files by connectiong to your anonymous FTP server, so too you might provide a very limited authorization to write to a file within a carefully proscribed namespace. > . to a file specified by the client I've only really considered *nix for building a name. But I suspect this can be easily addapted to other OSs. I had hoped I could use the code that opens the log in the patch posted to bugtraq, but the method seems broken to me. Also, I can't see how it deals with the symlink problem, but I haven't had time to play with it yet. It is broken in terms of the forced naming. First, it allows the file to be anywhere because it opens it in the current directory. I think it should be restricted to a selection between the user's home directory and the system-wide tmp directory. Second, using the time to create a "unique" name is just wrong. A user starting two sessions from the same shell script will have a problem, but can deal with it by adding "sleep" to slow down that expensive, fast computer. However, if it's two different users clicking at the same time, the user cannot predict or prevent the collision. It would have been less work and more reliable to use UID and PID instead of time. Still not perfect, but at least the user can clean out old files to prevent collisions. I'm planning to put two restrictions. First, a screen dump must go to a new file. This is harsh, but I think I need it to deal with the symlink race problem. Second, the pathname of the file is severely restricted. As long as the script/host has just a little control, it can do nice things for the user. It doesn't need full control of the pathname. The pathname will be [dir]/screen[.UID.PID].txt where [dir] is required and must be either "/tmp" or "/home/me". Home would be determined from /etc/passwd, not from $HOME. I'd like to make ".UID.PID" replaceable within a narrow range. With these restrictions, an external script which started the xterm (see /bin/expect) can have just enough control to detect some trigger and grab a screen image. A script should not be forced to do extra work between keystrokes to clear the file if a name change will be sufficient. Requiring that a new file be created for each screen dump partly protects existing files from modification, but doesn't fully deal with the symlink race condition. I expect that can be solved using the low-level I/O library, (i.e. open instead of fopen) but I haven't set up the experiment yet. Now, I *don't* think like a cracker, so if I've missed anything, please point it out to me! > This should not be done via a standard control sequence. If it is > going to be done it should be performed via a private sequence (preferably > not a '?' sequence) or via an APC. Hmm. I was thinking... OSC Ps ; P1 ; P2 ; Pt ST where Ps is the unique value to select this service, p1 is 0 for "use home", 1 for "use /tmp" p2 is 0 for mode 0666, 1 for mode 0600 Pt is: empty to use UID.PID " " (blank) to eliminate UID.PID or 1-8 digits, letters and period. A malicious host or a broken program could fill a partition with files named screen.1.txt, screen.2.txt and so on. But I think that's all it can do. Is filling the disk a serious enough threat to limit the Pt to one character? That wouldn't be harsh, but I think it's unnecessary. Why is this a bad idea? -- Oisin "Curly++" Curtin ocurtin@SPAM.usa.net Surface Liaison, Minetown Digger Send no SPAM. http://pages.infinit.net/curlypp/ From flo@rdel.co.uk Tue Mar 7 13:30:26 2000 Path: transfer.stratus.com!cambridge1-snf1.gtei.net!news.gtei.net!inmet!news3!arcl ight.uoregon.edu!newsfeed.mathworks.com!news-out.cwix.com!newsfeed.cwix.com! howland.erols.net!news-out-b.news.pipex.net.MISMATCH!tank.news.pipex.net!pip ex!warm.news.pipex.net!pipex!mailsvr!not-for-mail From: Paul Williams Newsgroups: comp.terminals,comp.os.linux.admin Subject: Re: Escape sequences for xterm? Date: Tue, 07 Mar 2000 08:41:52 +0000 Organization: It's in one of the piles Lines: 37 Message-ID: <38C4C0D0.D5AD4FF@rdel.co.uk> References: <8a1bfi$sn5$1@newsmaster.cc.columbia.edu> NNTP-Posting-Host: sswc019.rrds.co.uk Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Trace: rdel.co.uk 952418542 20743 172.16.10.30 (7 Mar 2000 08:42:22 GMT) X-Complaints-To: postmaster@rdel.co.uk NNTP-Posting-Date: 7 Mar 2000 08:42:22 GMT X-Mailer: Mozilla 4.7 [en] (X11; I; SunOS 5.6 sun4m) X-Accept-Language: en Xref: transfer.stratus.com comp.terminals:13832 Curly++ wrote: > > In comp.os.linux.admin, Jeffrey Altman proclaimed: > > > > This should not be done via a standard control sequence. If it is > > going to be done it should be performed via a private sequence > > (preferably not a '?' sequence) or via an APC. > > Hmm. I was thinking... > > OSC Ps ; P1 ; P2 ; Pt ST > where Ps is the unique value to select this service, > p1 is 0 for "use home", 1 for "use /tmp" > p2 is 0 for mode 0666, 1 for mode 0600 > Pt is: empty to use UID.PID > " " (blank) to eliminate UID.PID > or 1-8 digits, letters and period. It's a shame that the writers of xterm didn't structure the code space for OSC in the same way that DEC did with DCS, ie. use the same Param ;...; Param Intermediate Final structure as for CSI, with extra, textual parameters appearing after Final and before ST. It would've made the parser a lot simpler, because you'd know unambiguously where the text starts, even with a variable number of parameters. At the moment, it seems that the existing codes are of the form OSC Ps ; Pt ST which is OK until you add more Ps (or Pn) parameters before Pt. Then you have to start doing something special in the parser based on the value of the first Ps. (Of course, this decision is historical -- no reflection on the current maintainers. Never let history get in the way of a good whinge, I say!) From jaltman@watsun.cc.columbia.edu Tue Mar 7 13:30:30 2000 Path: transfer.stratus.com!cambridge1-snf1.gtei.net!news.gtei.net!bloom-beacon.mit .edu!newsswitch.lcs.mit.edu!news-spur1.maxwell.syr.edu!news.maxwell.syr.edu! newspeer1.nac.net!news.new-york.net!newsmaster.cc.columbia.edu!watsun.cc.col umbia.edu!jaltman From: jaltman@watsun.cc.columbia.edu (Jeffrey Altman) Newsgroups: comp.terminals,comp.os.linux.admin Subject: Re: Escape sequences for xterm? Date: 7 Mar 2000 14:23:45 GMT Organization: Columbia University Lines: 41 Message-ID: <8a33dh$bko$1@newsmaster.cc.columbia.edu> References: <8a1bfi$sn5$1@newsmaster.cc.columbia.edu> <38C4C0D0.D5AD4FF@rdel.co.uk> NNTP-Posting-Host: watsun.cc.columbia.edu X-Trace: newsmaster.cc.columbia.edu 952439025 11928 128.59.39.2 (7 Mar 2000 14:23:45 GMT) X-Complaints-To: postmaster@columbia.edu NNTP-Posting-Date: 7 Mar 2000 14:23:45 GMT Xref: transfer.stratus.com comp.terminals:13834 In article <38C4C0D0.D5AD4FF@rdel.co.uk>, Paul Williams wrote: : > OSC Ps ; P1 ; P2 ; Pt ST : > where Ps is the unique value to select this service, : > p1 is 0 for "use home", 1 for "use /tmp" : > p2 is 0 for mode 0666, 1 for mode 0600 : > Pt is: empty to use UID.PID : > " " (blank) to eliminate UID.PID : > or 1-8 digits, letters and period. : : It's a shame that the writers of xterm didn't structure the code space : for OSC in the same way that DEC did with DCS, ie. use the same : : Param ;...; Param Intermediate Final : : structure as for CSI, with extra, textual parameters appearing after : Final and before ST. It would've made the parser a lot simpler, because : you'd know unambiguously where the text starts, even with a variable : number of parameters. At the moment, it seems that the existing codes : are of the form : : OSC Ps ; Pt ST : : which is OK until you add more Ps (or Pn) parameters before Pt. Then you : have to start doing something special in the parser based on the value : of the first Ps. The whole purpose of the control sequences OSC, APC, DCS, PM, PU1, PU2 are so that something special will be done with the contents. If you can represent the command using X3.64 notation then you should not be using an OSC but should instead be using a private CSI sequence. Just to set the record straight, the OSC sequences for Set Window Title and Set Icon Name were developed by DEC for the DECterm and were later supported on windowing versions of the VT terminals. Jeffrey Altman * Sr.Software Designer * Kermit-95 for Win32 and OS/2 The Kermit Project * Columbia University 612 West 115th St #716 * New York, NY * 10025 http://www.kermit-project.org/k95.html * kermit-support@kermit-project.org From flo@rdel.co.uk Tue Mar 7 13:30:32 2000 Path: transfer.stratus.com!cambridge1-snf1.gtei.net!news.gtei.net!bloom-beacon.mit .edu!newsswitch.lcs.mit.edu!news-spur1.maxwell.syr.edu!news.maxwell.syr.edu! tank.news.pipex.net!pipex!warm.news.pipex.net!pipex!mailsvr!not-for-mail From: Paul Williams Newsgroups: comp.terminals,comp.os.linux.admin Subject: Re: Escape sequences for xterm? Date: Tue, 07 Mar 2000 15:27:28 +0000 Organization: It's in one of the piles Lines: 57 Message-ID: <38C51FE0.11DD3325@rdel.co.uk> References: <8a1bfi$sn5$1@newsmaster.cc.columbia.edu> <38C4C0D0.D5AD4FF@rdel.co.uk> <8a33dh$bko$1@newsmaster.cc.columbia.edu> NNTP-Posting-Host: sswc019.rrds.co.uk Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Trace: rdel.co.uk 952442879 27522 172.16.10.30 (7 Mar 2000 15:27:59 GMT) X-Complaints-To: postmaster@rdel.co.uk NNTP-Posting-Date: 7 Mar 2000 15:27:59 GMT X-Mailer: Mozilla 4.7 [en] (X11; I; SunOS 5.6 sun4m) X-Accept-Language: en Xref: transfer.stratus.com comp.terminals:13835 Jeffrey Altman wrote: > > In article <38C4C0D0.D5AD4FF@rdel.co.uk>, > Paul Williams wrote: > : > OSC Ps ; P1 ; P2 ; Pt ST > : > where Ps is the unique value to select this service, > : > p1 is 0 for "use home", 1 for "use /tmp" > : > p2 is 0 for mode 0666, 1 for mode 0600 > : > Pt is: empty to use UID.PID > : > " " (blank) to eliminate UID.PID > : > or 1-8 digits, letters and period. > : > : It's a shame that the writers of xterm didn't structure the code space > : for OSC in the same way that DEC did with DCS, ie. use the same > : > : Param ;...; Param Intermediate Final > : > : structure as for CSI, with extra, textual parameters appearing after > : Final and before ST. It would've made the parser a lot simpler, because > : you'd know unambiguously where the text starts, even with a variable > : number of parameters. At the moment, it seems that the existing codes > : are of the form > : > : OSC Ps ; Pt ST > : > : which is OK until you add more Ps (or Pn) parameters before Pt. Then you > : have to start doing something special in the parser based on the value > : of the first Ps. > > The whole purpose of the control sequences OSC, APC, DCS, PM, PU1, PU2 > are so that something special will be done with the contents. If you > can represent the command using X3.64 notation then you should not be > using an OSC but should instead be using a private CSI sequence. You can't represent a command containing a sixel string or filename as a CSI sequence, so yes, you must use one of the extension mechanisms you listed above. However, that doesn't stop you structuring these in the same way as CSI _up to a point_. It makes sense to use as much of the existing parser for your new sequences, and then just differ at the point where you need parameters that don't fit the CSI structure. Take a look at all the DCS sequences defined by DEC, and you'll see that that's exactly what they did. > Just to set the record straight, the OSC sequences for Set Window Title > and Set Icon Name were developed by DEC for the DECterm and were later > supported on windowing versions of the VT terminals. Then I'd say that DEC only ever intended their OSCs to have one selective parameter. Any more than that, and your sequence starts getting ambiguous, because ';' doesn't have any standard meaning in OSC. Hmm...are you saying that DEC defined the sequence "OSC Ps ; title BEL"? Because that's the only version of Set Window Title and Set Icon Name that xterm used to accept. Or was that miscopied across from the DECterm implementation? Which "windowing versions of VT terminals" supported these sequences? From bjh21@cus.cam.ac.uk Tue Mar 7 13:30:35 2000 Path: transfer.stratus.com!cambridge1-snf1.gtei.net!news.gtei.net!bloom-beacon.mit .edu!howland.erols.net!news-out-b.news.pipex.net.MISMATCH!tank.news.pipex.ne t!pipex!server1.netnews.ja.net!pegasus.csx.cam.ac.uk!bjh21 From: bjh21@cus.cam.ac.uk (Ben Harris) Newsgroups: comp.terminals,comp.os.linux.admin Subject: Re: Escape sequences for xterm? Date: 7 Mar 2000 16:14:47 GMT Organization: University of Cambridge, England Lines: 15 Message-ID: <8a39tn$et8$1@pegasus.csx.cam.ac.uk> References: <38C4C0D0.D5AD4FF@rdel.co.uk> <8a33dh$bko$1@newsmaster.cc.columbia.edu> <38C51FE0.11DD3325@rdel.co.uk> NNTP-Posting-Host: ursa.cus.cam.ac.uk Xref: transfer.stratus.com comp.terminals:13837 In article <38C51FE0.11DD3325@rdel.co.uk>, Paul Williams wrote: >Hmm...are you saying that DEC defined the sequence "OSC Ps ; title BEL"? >Because that's the only version of Set Window Title and Set Icon Name >that xterm used to accept. Or was that miscopied across from the DECterm >implementation? My rather vague understanding, from reading the PuTTY sources (PuTTY supports both DECterm and xterm title sequences) is that DECterm uses "OSC Ps L title ST" or somesuch. -- Ben Harris Unix Support, University of Cambridge Computing Service. If I wanted to speak for the University, I'd be in ucam.comp-serv.announce. From jaltman@watsun.cc.columbia.edu Tue Mar 7 13:30:39 2000 Path: transfer.stratus.com!cambridge1-snf1.gtei.net!news.gtei.net!bloom-beacon.mit .edu!news-peer.gip.net!news.gsl.net!gip.net!news.maxwell.syr.edu!newsfeed.ny u.edu!newsmaster.cc.columbia.edu!watsun.cc.columbia.edu!jaltman From: jaltman@watsun.cc.columbia.edu (Jeffrey Altman) Newsgroups: comp.terminals,comp.os.linux.admin Subject: Re: Escape sequences for xterm? Date: 7 Mar 2000 16:27:46 GMT Organization: Columbia University Lines: 46 Message-ID: <8a3am2$hs5$1@newsmaster.cc.columbia.edu> References: <38C4C0D0.D5AD4FF@rdel.co.uk> <8a33dh$bko$1@newsmaster.cc.columbia.edu> <38C51FE0.11DD3325@rdel.co.uk> NNTP-Posting-Host: watsun.cc.columbia.edu X-Trace: newsmaster.cc.columbia.edu 952446466 18309 128.59.39.2 (7 Mar 2000 16:27:46 GMT) X-Complaints-To: postmaster@columbia.edu NNTP-Posting-Date: 7 Mar 2000 16:27:46 GMT Xref: transfer.stratus.com comp.terminals:13838 In article <38C51FE0.11DD3325@rdel.co.uk>, Paul Williams wrote: : : You can't represent a command containing a sixel string or filename as a : CSI sequence, so yes, you must use one of the extension mechanisms you : listed above. However, that doesn't stop you structuring these in the : same way as CSI _up to a point_. It makes sense to use as much of the : existing parser for your new sequences, and then just differ at the : point where you need parameters that don't fit the CSI structure. Take a : look at all the DCS sequences defined by DEC, and you'll see that that's : exactly what they did. And when you can do that you use a DCS instead of an OSC or APC. DEC was very consistent about this. All DCS sequences follow X3.64 and the others do not. : > Just to set the record straight, the OSC sequences for Set Window Title : > and Set Icon Name were developed by DEC for the DECterm and were later : > supported on windowing versions of the VT terminals. : : Then I'd say that DEC only ever intended their OSCs to have one : selective parameter. Any more than that, and your sequence starts : getting ambiguous, because ';' doesn't have any standard meaning in OSC. : : Hmm...are you saying that DEC defined the sequence "OSC Ps ; title BEL"? : Because that's the only version of Set Window Title and Set Icon Name : that xterm used to accept. Or was that miscopied across from the DECterm : implementation? DEC defined it. DECSWT (Set Window Title) DECSIN (Set Icon Name) : Which "windowing versions of VT terminals" supported these sequences? VT5xx has it. And I think the VT420 had it but I don't have a manual to check against at the moment. Jeffrey Altman * Sr.Software Designer * Kermit-95 for Win32 and OS/2 The Kermit Project * Columbia University 612 West 115th St #716 * New York, NY * 10025 http://www.kermit-project.org/k95.html * kermit-support@kermit-project.org From flo@rdel.co.uk Tue Mar 7 13:30:42 2000 Path: transfer.stratus.com!cambridge1-snf1.gtei.net!news.gtei.net!bloom-beacon.mit .edu!newsswitch.lcs.mit.edu!news-spur1.maxwell.syr.edu!news.maxwell.syr.edu! tank.news.pipex.net!pipex!warm.news.pipex.net!pipex!mailsvr!not-for-mail From: Paul Williams Newsgroups: comp.terminals,comp.os.linux.admin Subject: Re: Escape sequences for xterm? Date: Tue, 07 Mar 2000 16:32:49 +0000 Organization: It's in one of the piles Lines: 20 Message-ID: <38C52F31.57255BB4@rdel.co.uk> References: <38C4C0D0.D5AD4FF@rdel.co.uk> <8a33dh$bko$1@newsmaster.cc.columbia.edu> <38C51FE0.11DD3325@rdel.co.uk> <8a3am2$hs5$1@newsmaster.cc.columbia.edu> NNTP-Posting-Host: sswc019.rrds.co.uk Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Trace: rdel.co.uk 952446800 7533 172.16.10.30 (7 Mar 2000 16:33:20 GMT) X-Complaints-To: postmaster@rdel.co.uk NNTP-Posting-Date: 7 Mar 2000 16:33:20 GMT X-Mailer: Mozilla 4.7 [en] (X11; I; SunOS 5.6 sun4m) X-Accept-Language: en Xref: transfer.stratus.com comp.terminals:13839 Jeffrey Altman wrote: > > In article <38C51FE0.11DD3325@rdel.co.uk>, > Paul Williams wrote: > : > : You can't represent a command containing a sixel string or filename as a > : CSI sequence, so yes, you must use one of the extension mechanisms you > : listed above. However, that doesn't stop you structuring these in the > : same way as CSI _up to a point_. It makes sense to use as much of the > : existing parser for your new sequences, and then just differ at the > : point where you need parameters that don't fit the CSI structure. Take a > : look at all the DCS sequences defined by DEC, and you'll see that that's > : exactly what they did. > > And when you can do that you use a DCS instead of an OSC or APC. > DEC was very consistent about this. All DCS sequences follow X3.64 and > the others do not. But this was _only_ a DEC decision. IIRC, DCS is not defined by X3.64 to have any more structure than OSC or APC. From dickey@shell.clark.net Tue Mar 7 17:04:08 2000 Path: transfer.stratus.com!cambridge1-snf1.gtei.net!news.gtei.net!homer.alpha.net! solaris.cc.vt.edu!news.vt.edu!news-out.cwix.com!newsfeed.cwix.com!newshub.no rtheast.verio.net!verio!iad-peer.news.verio.net!news.verio.net!iad-read.news .verio.net.POSTED!not-for-mail From: "T.E.Dickey" Subject: Re: Escape sequences for xterm? Newsgroups: comp.terminals,comp.os.linux.admin References: <8a1bfi$sn5$1@newsmaster.cc.columbia.edu> Organization: Clark Internet Services, Inc., Ellicott City, MD USA User-Agent: tin/1.5.2-20000119 ("Sumerland") (UNIX) (SunOS/5.6 (sun4u)) Lines: 15 Message-ID: Date: Tue, 07 Mar 2000 20:34:01 GMT NNTP-Posting-Host: 168.143.0.8 X-Complaints-To: abuse@verio.net X-Trace: iad-read.news.verio.net 952461241 168.143.0.8 (Tue, 07 Mar 2000 20:34:01 GMT) NNTP-Posting-Date: Tue, 07 Mar 2000 20:34:01 GMT Xref: transfer.stratus.com comp.terminals:13841 Curly++ wrote: > I had hoped I could use the code that opens the log in the patch > posted to bugtraq, but the method seems broken to me. Also, I briefly, the point of the patch is to ensure that the effective user owns the file - and no one else. if the file is in a directory that prevents non-owners from deleting the file, then that covers most of the remaining issues. (I added a bit to that though - it's in the most recent patch I made for xterm). -- Thomas E. Dickey dickey@clark.net http://www.clark.net/pub/dickey From dickey@shell.clark.net Tue Mar 7 17:04:19 2000 Path: transfer.stratus.com!cambridge1-snf1.gtei.net!news.gtei.net!bloom-beacon.mit .edu!micro-heart-of-gold.mit.edu!netnews.com!newspeer1.nac.net!news.new-york .net!iad-peer.news.verio.net!news.verio.net!iad-read.news.verio.net.POSTED!n ot-for-mail From: "T.E.Dickey" Subject: Re: Escape sequences for xterm? Newsgroups: comp.terminals,comp.os.linux.admin References: <38C4C0D0.D5AD4FF@rdel.co.uk> <8a33dh$bko$1@newsmaster.cc.columbia.edu> <38C51FE0.11DD3325@rdel.co.uk> <8a39tn$et8$1@pegasus.csx.cam.ac.uk> Organization: Clark Internet Services, Inc., Ellicott City, MD USA User-Agent: tin/1.5.2-20000119 ("Sumerland") (UNIX) (SunOS/5.6 (sun4u)) Lines: 24 Message-ID: <4Rdx4.1096$G15.28286@iad-read.news.verio.net> Date: Tue, 07 Mar 2000 20:42:40 GMT NNTP-Posting-Host: 168.143.0.8 X-Complaints-To: abuse@verio.net X-Trace: iad-read.news.verio.net 952461760 168.143.0.8 (Tue, 07 Mar 2000 20:42:40 GMT) NNTP-Posting-Date: Tue, 07 Mar 2000 20:42:40 GMT Xref: transfer.stratus.com comp.terminals:13842 Ben Harris wrote: > In article <38C51FE0.11DD3325@rdel.co.uk>, > Paul Williams wrote: >>Hmm...are you saying that DEC defined the sequence "OSC Ps ; title BEL"? >>Because that's the only version of Set Window Title and Set Icon Name >>that xterm used to accept. Or was that miscopied across from the DECterm >>implementation? > My rather vague understanding, from reading the PuTTY sources (PuTTY > supports both DECterm and xterm title sequences) is that DECterm uses > "OSC Ps L title ST" or somesuch. yes (OSC is supposed to end with ST, DECterm does the Right Thing). I had the impression that Sun's use of this sort of escape sequence preceded DEC's - but since both were implemented ~12 years ago, I'm not sure (I was using Apollo's at the time, and didn't much like X or xterm ;-). -- XFree86 xterm recognizes both... -- Thomas E. Dickey dickey@clark.net http://www.clark.net/pub/dickey