INFO-VAX Sun, 07 Sep 2008 Volume 2008 : Issue 490 Contents: Re: bug in TCPIP Re: bug in TCPIP Re: bug in TCPIP eBook versions now coming out Re: huge USB disks and VMS Re: Loose Cannon-dian Re: OT: Carly speeks at convetion Re: OT: Carly speeks at convetion Re: Security alarm msg Re: Security alarm msg Re: Security alarm msg Re: Security alarm msg Re: Security alarm msg Was there an EV69 Alpha? RE: Was there an EV69 Alpha? Re: Was there an EV69 Alpha? Re: Was there an EV69 Alpha? Re: Was there an EV69 Alpha? Re: WGET Re: WGET Re: WGET Re: WGET Re: WGET Re: WGET Re: WGET ---------------------------------------------------------------------- Date: Sat, 6 Sep 2008 15:50:11 -0500 (CDT) From: sms@antinode.info (Steven M. Schweda) Subject: Re: bug in TCPIP Message-ID: <08090615501135_20201252@antinode.info> From: helbig@astro.multiCLOTHESvax.de (Phillip Helbig---remove CLOTHES to reply) One might argue that a more descriptive title would be helpful, as a bug in TCPIP is like a grain of sand on the beach. But I won't. > HP TCP/IP Services for OpenVMS Alpha Version V5.4 - ECO 6 > on a DEC 3000 - M600 running OpenVMS V7.3-2 > > I have my SMTP configured to drop connections trying to send email to > non-existent usernames. As is well known, they are only dropped if the > non-existent usernames are syntactically correct VMS usernames; in > particular, anything longer than 12 characters is not dropped (which > usually results in a bounce---which is bad, because this creates > backscatter spam). (By the way, is this fixed in a newer version of > TCPIP?) Old bug, previously discussed, where the word was that TCPUP V5.5 had the fix. http://groups.google.com/group/comp.os.vms/browse_thread/thread/4794fbf5d7c37026/2d3e6745a2cf4fa2 > I just discovered the following: send email to > USERfldlfkdndueldlf where USER is a valid username and the rest of the > letters makes the bogus username longer than 12 characters---it gets > delivered to USER. That's novel. ------------------------------------------------------------------------ Steven M. Schweda sms@antinode-info 382 South Warwick Street (+1) 651-699-9818 Saint Paul MN 55105-2547 ------------------------------ Date: Sat, 06 Sep 2008 17:06:30 -0400 From: JF Mezei Subject: Re: bug in TCPIP Message-ID: <48c2f12f$0$12358$c3e8da3@news.astraweb.com> Phillip Helbig---remove CLOTHES to reply wrote: > particular, anything longer than 12 characters is not dropped (which > usually results in a bounce---which is bad, because this creates Yes, this is fixed in 5.6. That was one hell of a large improvement. > TCPIP?) I just discovered the following: send email to > USERfldlfkdndueldlf where USER is a valid username and the rest of the > letters makes the bogus username longer than 12 characters---it gets > delivered to USER. Now, this is funny. I wonder how the weenie coder managed to get this one. If he did an indexed search of SYSUAF with KEY greater or equal to "USERhdhdhdhdhd" , he wouldn't get "USER ". Perhaps he did a "start" with only the first letter of the username and read sequentially, and only checked as many characters as were present in the SYSUAF key ? (at which point, USER = USER since you are only checking for 5 bytes), and then he would use the SYSUAF data to feed to the callable MAIL instead of using the bad username in the message. ------------------------------ Date: Sun, 7 Sep 2008 00:48:00 +0000 (UTC) From: helbig@astro.multiCLOTHESvax.de (Phillip Helbig---remove CLOTHES to reply) Subject: Re: bug in TCPIP Message-ID: In article <48c2f12f$0$12358$c3e8da3@news.astraweb.com>, JF Mezei writes: > Phillip Helbig---remove CLOTHES to reply wrote: > > > particular, anything longer than 12 characters is not dropped (which > > usually results in a bounce---which is bad, because this creates > > Yes, this is fixed in 5.6. That was one hell of a large improvement. OK. I WILL upgrade soon, I promise. Unless I will have problems keeping 7.3 VAX machines in the cluster. If there are problems, I'll have to decide whether to get rid of the VAXes (and stock up on tissues for all the tears) and replace them with ALPHAs or keep them and stay stuck at my current ALPHA software level. (I would probably go for the replacement, but it would be nice if the cluster would at least more or less mostly work so that I could first upgrade then do the replacement.) ------------------------------ Date: Sat, 6 Sep 2008 21:39:15 -0700 (PDT) From: yyyc186 Subject: eBook versions now coming out Message-ID: This is only slightly off topic and I suppose I could be slammed for advertising, but ebook versions of "The Minimum You Need to Know" series are coming out in MobiPocket format. Probably best to search on my name since it seems like every eBook site taking the load from the MobiPockets database truncated "The Minimum You Need to Know" I don't really know how many of you use eBooks, but MobiPockets does have a free reader for MAC and Evil Empire operating systems, so if you were using it for reference and would like to do a keyword search, you would be able to do it. ------------------------------ Date: Sat, 06 Sep 2008 21:26:14 -0400 From: Howard S Shubs Subject: Re: huge USB disks and VMS Message-ID: In article , brooks@cuebid.ovms.usa.hp.nospam (Rob Brooks) wrote: > This will theoretically allow volumes up to 2TB in size. I'd be surprised if they stop there. Such disks may well be out by then. -- Don't bother with piddly crap like "gun control". Life is 100% fatal. Ban it. ------------------------------ Date: 6 Sep 2008 23:07:16 GMT From: billg999@cs.uofs.edu (Bill Gunshannon) Subject: Re: Loose Cannon-dian Message-ID: <6iggp4Fpp1j7U1@mid.individual.net> In article <48c09bf8$0$12400$c3e8da3@news.astraweb.com>, JF Mezei writes: > Bill Gunshannon wrote: > >> like is done here frequently, he once again refered to "anything ported >> from Unix" and described it as "riddled with bugs and buffer overflow >> risks because it is not really 'native' VMS software". > > There are Unix concepts which work well in Unix, but when recompiled on > VMS don't "translate" well. And that is Unix's fault how??? > > An example was the pop server. On Unix, there is an option to define the > filename where the log is to be written by the pop server (deamon in > Unix terms). This was recompiled on VMS and they didn't think of the impact. And that is Unix's fault how??? > > On VMS, the pop server image was installed with privileges. This meant > that anyone could invoke the program, specify a filename anywhere on the > system and the program would write a new file at that location (aka: > ability to overwrite any system files). And that is Unix's fault how??? Get the picture? None of this is Unix's fault but it sure lays waste the claim that VMS got all the best programmers and was soooooo much better engineered than Unix. Don't you think if security was such a big concern that someone might have looked at the consequences of such a port? I have ported a lot of software, both to and from Unix and other OSes (like Primos and DomainOS) and I can assure you I didn't ever try just blindly compiling code and throwing it into production. > > Another example is the IMAP/POP/XDM servers. They were ported from Unix > and they did not think of adding the code needed to handle the wrong > password logging and intrusion detection/evasion mechanisms. I doubt thinking had anything to do with it. Much more likely, they just didn't care. Or didn't see the added security as worth the effort. After all, it worked fine on Unix without it. :-) > So those > packages don't adhere to VMS standards and thus lower the security of > the VMS system compared to the good old days where all the software was > native to VMS and written specifically for VMS using VMS constructs and > services. More bullcrap. I doubt "VMS Standards" or "VMS Security" were even given a passing glance. There is nothing to show that "security" was the underlying principle in everything VMS did any more than that Unix didn't consider it at all. In both cases people had an idea of what they wanted to make their OSes do and the wrote code to do it. One was no more guaranteed to be secure than the other. How many years between the first release of Unix and the first reported case of one of these "weaknesses" being exploited? It is hardly a shortcoming that Unix didn't anticipate the change in the moral climate of the world that has brought us "hackers". > > It has become expected nowadasy to see such problems when software is > ported from Unix to VMS. Apache excpects text files in a format that is > not standard for VMS for instance. The Apache group does not write software for VMS. They have never claimed they did. Why would you expect them to program to a "standard" that isn't their standard? If the software gets ported to VMS it is VMS programmers doing it. If they do it badly you can't blame the original author because he chose not use your standard. > > > >> I was merely >> pointing out that SMG , while not "ported from Unix" and "really 'native' >> VMS software" was found to have "bugs and buffer overflow risks". > > Yes, and it is a rude awakening that even native software that dates > back many years isn't perfect. What this shows is that quality > assurance started to go down in the 1990s. Or maybe just that the quality of the code, the quality assurance of the product and, yes, the actual capabilities of the programmers were no different than anyone else in the industry. The desired result was different and they produced a product accordingly. And, no matter how big a piece of the market you would like to think it had, it never had enough of a footprint to attract the class of people who spent (and continue to spend) so much time looking for holes in other OSes. Like I ahve said here before. Primos is still in use commercially. Primos machines are on the Internet. There are even public machines for people (read: hackers) to play with. How many CERT vulnerabilities are there for Primos? > > >> That remains to be seen. Because they have never been reported or tracked >> by any outside source (look at the reluctance to trport any of these recent >> discoveries to CERT) > > Personally, I am not sure who is expected to report stuff to CERT. Can > *anyone* really report stuff, or do they expect to see reports only from > certain qualified people and/or the vendors ? I believe they will take any report. But the then try to verify it with the responsible parties. It is my understanding that in the case of VMS the norm is to ignore requests from CERT for these verifications. > > >> the idea that DEC's "checks and balances" and "code reviews, walk-throughs, >> programming standards" were any better than anyone elses. > > They were. But since the Palmer era, there has bneen so much slash and > burn done that VMS is no longer what it used to be. Sorry, but your claims that they were are no more verifiable than my claims that they weren't. Mine, however, are probably a lot more believable. > >> I think no one outside of DEC/Compaq/HP has any idea how many exploits >> equivalent to those found in Unix have or still exist in VMS. > > I suspect people outside of HP have a better grasp of the > vulnerabilities in VMS. HP doesn't care about VMS. Not caring and not knowing are nont synonomous. bill -- Bill Gunshannon | de-moc-ra-cy (di mok' ra see) n. Three wolves billg999@cs.scranton.edu | and a sheep voting on what's for dinner. University of Scranton | Scranton, Pennsylvania | #include ------------------------------ Date: Sat, 06 Sep 2008 16:51:47 -0400 From: JF Mezei Subject: Re: OT: Carly speeks at convetion Message-ID: <48c2edc2$0$1561$c3e8da3@news.astraweb.com> Bob Willard wrote: > Postmaster General ceased to be a cabinet position during Nixon's (first?) > term as president; since then, the USPS has been a private organization. Makes sense. I was misguided by a documentary I had seen about president Ford (*) during a highjacking of Air Force One where he promised the post of Postmaster General to one lady who found a way to send a message to the white house via fax. Would it be correct to state that the board of the USPS would then be political appointees ? (*) Harisson Ford ------------------------------ Date: Sat, 06 Sep 2008 17:16:44 -0400 From: Bob Willard Subject: Re: OT: Carly speeks at convetion Message-ID: JF Mezei wrote: > Would it be correct to state that the board of the USPS would then be > political appointees ? With the possible exceptions of the current CEO and COO (both of whom have years of USPS experience), the USPS Board of Gov's all seem to be political appointees. And, when the PM was a cabinet position, the PM was a very-political appointment. Many times, IIRC, POTUS appointed his campaing manager. Amazing that the mail goes mostly through, under a system that is managed by pure ;-) politicians. -- Cheers, Bob ------------------------------ Date: Sat, 6 Sep 2008 13:49:35 -0700 (PDT) From: Bob Gezelter Subject: Re: Security alarm msg Message-ID: <63eb20e4-f449-45fa-b553-66050320f657@79g2000hsk.googlegroups.com> On Sep 6, 9:56=A0am, "Tom Linden" wrote: > I noted following on opcon. =A0Why is the remote node id in decimal forma= t? > This is on 8.3 Itanium. > > Message from user AUDIT$SERVER on REX > Security alarm (SECURITY) and security audit (SECURITY) on REX, system id= : =A0 > 2060 > Auditable event: =A0 =A0 =A0 =A0 =A0Network breakin detection > Event time: =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A06-SEP-2008 06:49:14.22 > PID: =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A020F0B1A8 > Process name: =A0 =A0 =A0 =A0 =A0 =A0 TCPIP$FTPC00079 > Username: =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 newuser > Remote node id: =A0 =A0 =A0 =A0 =A0 998090410 > Remote node fullname: =A0 =A0 59-125-166-170.HINET-IP.hinet.net > Remote username: =A0 =A0 =A0 =A0 =A0FTP_3B7DA6AA > Status: =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 %LOGIN-F-NOSUCHUSER, no such = user > > -- > PL/I for OpenVMSwww.kednos.com Tom, I don't have access to an equivalent configuration at this instant. What is the inverse DNS for the address look like? - Bob Gezelter, http://www.rlgsc.com ------------------------------ Date: Sat, 06 Sep 2008 16:56:36 -0400 From: JF Mezei Subject: Re: Security alarm msg Message-ID: <48c2eedd$0$12415$c3e8da3@news.astraweb.com> Tom Linden wrote: > I noted following on opcon. Why is the remote node id in decimal format? > This is on 8.3 Itanium. > Remote node id: 998090410 It has been like that for a long while. You need to do a netstat -n to get the real IP if the connection is still active, or look at your router's syslog file to get the real IP address, or plug that decimal number in a symbol, do a show symbol, and extract individual hex bytes, convert them to decimal to get your usable IP address. I suspect some coder decided it was much easier to just feed an integer instead of doing an sprinff to convert 4 individual bytes into usable IP address. (dotted decimal notation). I wonder what they will do with IPv6 addresses. Will be a mighty long number :-) And we won't be able to feed it to a symbol because VMS symbols are limited to 32 bits, right ? ------------------------------ Date: Sat, 6 Sep 2008 22:28:29 +0000 (UTC) From: moroney@world.std.spaamtrap.com (Michael Moroney) Subject: Re: Security alarm msg Message-ID: "Tom Linden" writes: >I noted following on opcon. Why is the remote node id in decimal format? >This is on 8.3 Itanium. >Message from user AUDIT$SERVER on REX >Security alarm (SECURITY) and security audit (SECURITY) on REX, system id: >2060 >Auditable event: Network breakin detection >Event time: 6-SEP-2008 06:49:14.22 >PID: 20F0B1A8 >Process name: TCPIP$FTPC00079 >Username: newuser >Remote node id: 998090410 >Remote node fullname: 59-125-166-170.HINET-IP.hinet.net >Remote username: FTP_3B7DA6AA >Status: %LOGIN-F-NOSUCHUSER, no such user That is someone trying to break into your system via FTP. They use a brute force attack (being unaware of VMS's breakin evasion). Don't take it personally, it's likely a script kiddie at work, and the attacking system (59-125-166-170.HINET-IP.hinet.net) is as likely as not a trojaned zombie slave. Your IP address came out of a random number generator. Do an $ANALYZE/AUDIT/SINCE= and you may see hundreds, thousands or tens of thousands of attempts while a script hacks away. The "Remote node id" being a number corresponding to the IP address is one of many problems TCP/IP has with security auditing. ------------------------------ Date: Sun, 7 Sep 2008 00:43:23 +0000 (UTC) From: helbig@astro.multiCLOTHESvax.de (Phillip Helbig---remove CLOTHES to reply) Subject: Re: Security alarm msg Message-ID: In article , moroney@world.std.spaamtrap.com (Michael Moroney) writes: > That is someone trying to break into your system via FTP. They use a > brute force attack (being unaware of VMS's breakin evasion). > > Don't take it personally, it's likely a script kiddie at work, and the > attacking system (59-125-166-170.HINET-IP.hinet.net) is as likely as > not a trojaned zombie slave. Your IP address came out of a random number > generator. If only to keep the size of ACCOUNTNG.DAT under control, when I see this happening, I stop FTP and restart it immediately. I've done this perhaps 20 times over they years. I don't think there was a single time when this didn't immediately stop the attack. ------------------------------ Date: Sat, 06 Sep 2008 17:53:53 -0700 From: "Tom Linden" Subject: Re: Security alarm msg Message-ID: On Sat, 06 Sep 2008 17:43:23 -0700, Phillip Helbig---remove CLOTHES to reply wrote: > In article , moroney@world.std.spaamtrap.com > (Michael Moroney) writes: > >> That is someone trying to break into your system via FTP. They use a >> brute force attack (being unaware of VMS's breakin evasion). >> >> Don't take it personally, it's likely a script kiddie at work, and the >> attacking system (59-125-166-170.HINET-IP.hinet.net) is as likely as >> not a trojaned zombie slave. Your IP address came out of a random >> number >> generator. > > If only to keep the size of ACCOUNTNG.DAT under control, when I see this > happening, I stop FTP and restart it immediately. I've done this > perhaps 20 times over they years. I don't think there was a single time > when this didn't immediately stop the attack. > I did that too, right away. I actually had no need for ftp on that node anyway, and if I did I could just enable it -- PL/I for OpenVMS www.kednos.com ------------------------------ Date: 6 Sep 08 14:42:23 EDT From: cook@wvnvms.wvnet.edu (George Cook) Subject: Was there an EV69 Alpha? Message-ID: I am doing some research for the Wikipedia DEC Alpha article. Was there an EV69 (was it only planned or actually manufactured?), and if so, is there any reliable online information about it? George Cook WVNET ------------------------------ Date: Sat, 6 Sep 2008 19:28:44 +0000 From: "Main, Kerry" Subject: RE: Was there an EV69 Alpha? Message-ID: <9D02E14BC0A2AE43A5D16A4CD8EC5A593ED5F5CF27@GVW1158EXB.americas.hpqcorp.net> > -----Original Message----- > From: George Cook [mailto:cook@wvnvms.wvnet.edu] > Sent: Saturday, September 06, 2008 2:42 PM > To: Info-VAX@Mvb.Saic.Com > Subject: Was there an EV69 Alpha? > > I am doing some research for the Wikipedia DEC Alpha article. Was > there an EV69 (was it only planned or actually manufactured?), and if > so, is there any reliable online information about it? > > > George Cook > WVNET Check these links: http://archives.neohapsis.com/archives/linux/axp/2002-q3/0276.html http://www.rubicad.net/news/alphamigration.htm Regards Kerry Main Senior Consultant HP Services Canada Voice: 613-254-8911 Fax: 613-591-4477 kerryDOTmainAThpDOTcom (remove the DOT's and AT) OpenVMS - the secure, multi-site OS that just works. ------------------------------ Date: Sat, 06 Sep 2008 15:33:00 -0400 From: =?ISO-8859-1?Q?Arne_Vajh=F8j?= Subject: Re: Was there an EV69 Alpha? Message-ID: <48c2daeb$0$90273$14726298@news.sunsite.dk> George Cook wrote: > I am doing some research for the Wikipedia DEC Alpha article. Was > there an EV69 (was it only planned or actually manufactured?), and if > so, is there any reliable online information about it? I have never heard about a system shipping with an EV69. Arne ------------------------------ Date: Sat, 06 Sep 2008 17:19:30 -0400 From: JF Mezei Subject: Re: Was there an EV69 Alpha? Message-ID: <48c2f41f$0$9639$c3e8da3@news.astraweb.com> George Cook wrote: > I am doing some research for the Wikipedia DEC Alpha article. Was > there an EV69 (was it only planned or actually manufactured?), and if > so, is there any reliable online information about it? Since the last batch of Alphaservers were either EV68 (like ES45) or EV7 (ES47 and higher), It is doubtful that EV69 was ever produced commercially. BUT, consider that they got suffiently good yields in their initial EV7 batches to start to commercialise an EV7z at higher clock rate. It is possible that the same happened with EV68, but that they decided against commercialising an EV68 with faster clock rate which might have been called EV69. This is pure speculation on my part. ------------------------------ Date: Sat, 6 Sep 2008 21:34:44 -0700 (PDT) From: yyyc186 Subject: Re: Was there an EV69 Alpha? Message-ID: On Sep 6, 4:19=A0pm, JF Mezei wrote: > possible that the same happened with EV68, but that they decided against > =A0commercialising an EV68 with faster clock rate which might have been > called EV69. > Some how I don't think they would have ever release a model called "EV69". It sounds too much like a porn title. ------------------------------ Date: Sat, 06 Sep 2008 14:18:03 -0400 From: =?ISO-8859-15?Q?Arne_Vajh=F8j?= Subject: Re: WGET Message-ID: <48c2c95b$0$90274$14726298@news.sunsite.dk> Tom Linden wrote: > GNU Wget 1.10.1a > > wget -r -k http://some_web_site/A/B/C > > downloads everythhing from some_web_site and not just C > as I would have expected. Is this expected behaviour? Yes. -r, --recursive specify recursive download. Arne ------------------------------ Date: Sat, 06 Sep 2008 12:00:22 -0700 From: "Tom Linden" Subject: Re: WGET Message-ID: On Sat, 06 Sep 2008 11:18:03 -0700, Arne Vajhøj wrote: > Tom Linden wrote: >> GNU Wget 1.10.1a >> wget -r -k http://some_web_site/A/B/C >> downloads everythhing from some_web_site and not just C >> as I would have expected. Is this expected behaviour? > > Yes. > > -r, --recursive specify recursive download. > > Arne I had assumed it would be recursive from the point specified as in the above example, not from the top level. BTW, when running this how do you keep it running after logging out, throw into the background as independent process -- PL/I for OpenVMS www.kednos.com ------------------------------ Date: Sat, 06 Sep 2008 14:18:41 -0500 From: BRAD@rabbit.turquoisewitch.com (Brad Hamilton) Subject: Re: WGET Message-ID: In article , Tom Linden wrote: >On Sat, 06 Sep 2008 11:18:03 -0700, Arne Vajhøj wrote: [...] >> -r, --recursive specify recursive download. >> >> Arne > >I had assumed it would be recursive from the point specified >as in the above example, not from the top level. > >BTW, when running this how do you keep it running after logging out, >throw into the background as independent process Create a command procedure and submit it to batch? ------------------------------ Date: Sat, 06 Sep 2008 12:22:43 -0700 From: "Tom Linden" Subject: Re: WGET Message-ID: On Sat, 06 Sep 2008 12:18:41 -0700, Brad Hamilton wrote: > In article , Tom Linden > wrote: >> On Sat, 06 Sep 2008 11:18:03 -0700, Arne Vajhøj wrote: > [...] >>> -r, --recursive specify recursive download. >>> >>> Arne >> >> I had assumed it would be recursive from the point specified >> as in the above example, not from the top level. >> >> BTW, when running this how do you keep it running after logging out, >> throw into the background as independent process > > Create a command procedure and submit it to batch? It is running interactively, I want to let it continue running after logging out, don't want to restart it. -- PL/I for OpenVMS www.kednos.com ------------------------------ Date: Sat, 06 Sep 2008 15:31:44 -0400 From: =?ISO-8859-15?Q?Arne_Vajh=F8j?= Subject: Re: WGET Message-ID: <48c2da9f$0$90273$14726298@news.sunsite.dk> Tom Linden wrote: > On Sat, 06 Sep 2008 11:18:03 -0700, Arne Vajhøj wrote: >> Tom Linden wrote: >>> GNU Wget 1.10.1a >>> wget -r -k http://some_web_site/A/B/C >>> downloads everythhing from some_web_site and not just C >>> as I would have expected. Is this expected behaviour? >> >> Yes. >> >> -r, --recursive specify recursive download. > > I had assumed it would be recursive from the point specified > as in the above example, not from the top level. I would too. Any back links on the page ? > BTW, when running this how do you keep it running after logging out, > throw into the background as independent process Batch job or (other) detached process. Arne ------------------------------ Date: Sat, 6 Sep 2008 16:02:57 -0500 (CDT) From: sms@antinode.info (Steven M. Schweda) Subject: Re: WGET Message-ID: <08090616025772_20201252@antinode.info> From: "Tom Linden" > GNU Wget 1.10.1a Not the latest. > wget -r -k http://some_web_site/A/B/C > > downloads everythhing from some_web_site and not just C > as I would have expected. Is this expected behaviour? It depends. Wget with "-r" follows links. If the links lead up, then wget follows them up. If the links lead down, then wget follows them down. There is this option, however ("wget -h"): [...] Recursive accept/reject: [...] -np, --no-parent don't ascend to the parent directory. [...] which sounds as if it might be helpful here. From: =?ISO-8859-15?Q?Arne_Vajh=F8j?= > Yes. Trust no one, I always say. From: "Tom Linden" > I had assumed it would be recursive from the point specified > as in the above example, not from the top level. It does recursive link following from the page you specify, but by default, it doesn't know up from down. > It is running interactively, I want to let it continue running after > logging out, don't want to restart it. Wrong OS. Try something UNIX-like if you need that feature. ------------------------------------------------------------------------ Steven M. Schweda sms@antinode-info 382 South Warwick Street (+1) 651-699-9818 Saint Paul MN 55105-2547 ------------------------------ Date: Sun, 07 Sep 2008 00:24:15 +0200 From: Michael Kraemer Subject: Re: WGET Message-ID: Tom Linden schrieb: > I had assumed it would be recursive from the point specified > as in the above example, not from the top level. wget -r is recursive in either direction. The depth of recursion can - in theory - be controlled with the "-l" switch. "-l 0" specifies "unlimited" AFAIR, the default is 5. But don't hold your breath, wget's recursion not always does what it is supposed to do. (but it's a valuable tool though) > BTW, when running this how do you keep it running after logging out, > throw into the background as independent process On Unix, I do sth like nohup wget -r ... & ------------------------------ End of INFO-VAX 2008.490 ************************