Date: Wed, 8 Mar 95 21:46:41 EST From: Snuffles@kew.com Subject: UUPC-Info-Request Digest 1995 #7 To: uupc-info-digest@kew.com Message-ID: Reply-To: UUPC-Info-Request@kew.com UUPC-Info-Request Digest Wed, 8 Mar 95 Volume 1995: Issue 7 Today's Topics: Bug in expire? emergency! fix for newgroup bug (what a cute bug!) HELP -Reply -Reply looking for voicemail/answering-machine (2 msgs) news, etc. news under 1.12n DOS Outbound question Protocols (2 msgs) sendbats discrepancy UUCP 1.12m/n Dos UUPC 1.12n bug report To subscribe to UUPC-Info-Digest, send the command in the body of a message to listserv@kew.com: subscribe uupc-info-Digest To signoff from UUPC-Info-Digest, use "signoff" instead of "subscribe". You can also send an "index" to the listserv to get an index of back issues and other files available for retrieval. ---------------------------------------------------------------------- Date: Tue, 7 Mar 95 22:01:12 (null) From: eric@terra.xs4all.nl Subject: Bug in expire? To: UUPC/Extended mailing list -----BEGIN PGP SIGNED MESSAGE----- Hi, I just wanted to see the header of a file, and so I looked at the article directly by going into my news directories. When looking around there I noticed that there were some really old files in it. Apparently expire didn't remove them. My news normally gets expired after 7 days (the default of expire), so I run expire without parameters. I just removed all the files that were dated between 11-1-94 and 3-1-94 (4OS2: del /s /[d11-1-94,3-1-95]) and I noticed that that gave me about 20Mb more diskspace free. That while the total news feed was about 36Mb.... I think that was quite a lot of wasted diskspace. Therefore, I have the strong feeling that expire has a bug. Anyone with me who thinks the same? I use uupc 1.12k 32bit OS/2 version. I have my news feed in the UNIX style (not in SNEWS). - -- Eric Veldhuyzen TEAM OS/2 Eric.Veldhuyzen@si.hhs.nl CIS: [100010,3051] Eric@terra.xs4all.nl PGP-KeyID: 0xFB64FCB3 ************ FIGHT to keep your right to PRIVACY. Use PGP! ************ -----BEGIN PGP SIGNATURE----- Version: 2.6 iQCVAgUBL1zJzSRI0DH7ZPyzAQE9iwQAn5uvgZeeEfumTjIJD8BFs51jTLO0F7m3 7Zw0LZc6RWn1PdrYhB2rRjllccJN1QabnmeqVUP+NCr11yn0FuGI6mkuaShPpqA2 PjoWblMqujy8miZSm/iXL+PJT0429xp/C5HHqFyFrM8QzjNyqJYW1qTOyZnTc8sF maw30FVrWeM= =u0UC -----END PGP SIGNATURE----- ------------------------------ Date: Tue, 7 Mar 1995 20:10:57 +0100 From: rommel@ars.muc.de Subject: emergency! To: UUPC/Extended mailing list There seems to be a fatal bug in "newgroup" processing. When I got a "newgroup comp.mail.uucp" control message today, that section in my active file until this time looked like: .... comp.windows.interviews 2343 2311 y comp.mail.elm 10155 10044 y comp.mail.uucp 3051 2996 y comp.mail.mime 3627 3593 y comp.mail.sendmail 8867 8659 y comp.dcom.modems 41265 40288 y comp.dcom.isdn 10584 10260 y comp.dcom.fax 5759 5642 y comp.dcom.lans.ethernet 3501 3479 y .... After that message was processed, that same section looked like: .... comp.windows.interviews 2343 2311 y comp.mail.elm 10155 10044 y comp.mail.uucp 3051 2996 y comp.mail.uucp 0 0 y That means, the rest of my active file was lost! Fortunately, I could restore it. So, until we can fix that, everybody using 1.12n should turn of the honorcontrol option ASAP and check his/her active file! I looked in my newsrun.log and found: 3/07-19:38 newsrun: Retained 138 articles, of which 0 were duplicates an 3/07-19:38 Batched: Skipped 1260 bytes in 31 lines after article 63 3/07-19:38 Control Message: newgroup de.alt.fan.zipp 3/07-19:38 newsgroup added: de.alt.fan.zipp 3/07-19:38 Batched: Skipped 2208 bytes in 52 lines after article 125 3/07-19:38 newsrun: Retained 146 articles, of which 0 were duplicates an 3/07-19:38 Control Message: newgroup comp.mail.uucp 3/07-19:38 newsgroup added: comp.mail.uucp 3/07-19:39 newsrun: Retained 116 articles, of which 0 were duplicates an 3/07-19:39 Control Message: cancel <95030566434@killer.bfoot.itc.com> It may have to do something with the previous newgroup de.alt.fan.zipp or with the fact that the comp.mail.uucp group already existed. By the way, I see several of these Skipped messages per day. Apparently (I checked one batch) for some of my received article batches, the number after the "#! rnews " is a few bytes larger than the actual article size of one or a few articles in the batch. Strange. Is it possible that the rnews->newsrun step (decompression) perhaps strips some trailing blanks or such stuff? Kai Uwe Rommel -- /* Kai Uwe Rommel Muenchen, Germany * * rommel@ars.muc.de CompuServe 100265,2651 * * rommel@informatik.tu-muenchen.de Fax +49 89 324 4524 */ DOS ... is still a real mode only non-reentrant interrupt handler, and always will be. -Russell Williams -- Drew Derbyshire UUPC/extended e-mail: software@kew.com Telephone: 617-641-3452 "You Belong to the City, You Belong to the Night . . ." ------------------------------ Date: Tue, 07 Mar 1995 19:24:29 -0500 From: Software@kew.com Subject: fix for newgroup bug (what a cute bug!) To: UUPC/Extended mailing list On Tue, 7 Mar 1995 20:31:17 +0100, "Kai Uwe Rommel" wrote: > Sorry, couldn't resist. :-) > > Now, looking at the new active.c, the search loop in ad_newsgroup() > looks like: > > struct grp *cur = group_list; > > while ((strcmp(grp, cur->grp_name) != 0)) > { > if (cur->grp_next != NULL) > { > if ( equali( cur->grp_name, grp ) ) /* Group exists? */ > return KWFalse; /* Yes --> Cannot add it again */ > > cur = cur->grp_next; > } > else /* Add group at end */ > break; > > } /* while ((strcmp(grp, cur->grp_name) != 0)) */ > > cur->grp_next = (struct grp *) malloc(sizeof(struct grp)); > cur = cur->grp_next; > > Now try to find out what happens when comp.mail.uucp is searched when > it already exists! Found it? Right! The outer while() already aborts > and execution does not get to the equali() comparison. What happens in > this case? The comp.mail.uucp is added a second time and the rest of > the list is cut off! > > Context diff for the fix: > > *** active.c~ Sun Jan 15 19:48:34 1995 > --- active.c Tue Mar 07 21:26:10 1995 > *************** > *** 327,347 **** > > KWBoolean add_newsgroup(const char *grp, const KWBoolean moderated) > { > ! struct grp *cur = group_list; > > ! while ((strcmp(grp, cur->grp_name) != 0)) > ! { > ! if (cur->grp_next != NULL) > { > if ( equali( cur->grp_name, grp ) ) /* Group exists? */ > return KWFalse; /* Yes --> Cannot add it again */ > > ! cur = cur->grp_next; > ! } > ! else /* Add group at end */ > break; > > ! } /* while ((strcmp(grp, cur->grp_name) != 0)) */ > > /*--------------------------------------------------------------------*/ > /* We're at the end of the chain with no group found, add it */ > --- 327,344 ---- > > KWBoolean add_newsgroup(const char *grp, const KWBoolean moderated) > { > ! struct grp *cur; > > ! for (cur = group_list; KWTrue; cur = cur->grp_next) > { > + > if ( equali( cur->grp_name, grp ) ) /* Group exists? */ > return KWFalse; /* Yes --> Cannot add it again */ > > ! if (cur->grp_next == NULL) > break; > > ! } /* for */ > > /*--------------------------------------------------------------------*/ > /* We're at the end of the chain with no group found, add it */ > > I have triple checked, this seems to be correct. Why did you change the > 1.12k code at all, BTW? Planning for the future hack of keeping the active file sorted? Mostly, because I hate code like that in the middle of a loop. You're to a man who routinely whacks braces. I'm a control freak! -ahd- -- Drew Derbyshire UUPC/extended e-mail: software@kew.com Telephone: 617-641-3452 "You Belong to the City, You Belong to the Night . . ." ------------------------------ Date: Mon, 06 Mar 1995 16:05:24 -0500 From: Software@kew.com Subject: HELP -Reply -Reply To: UUPC/Extended mailing list On Mon, 06 Mar 1995 12:54:14 MST, "David L. West" wrote: > > I hope you're not invoking MAIL, you should be running RMAIL > > directly if there is another front-end involved. > > Actually, I *was* calling mail, but I have since switched to > calling rmail instead. For inbound mail, I am simply going to > parse out the MIME info from the file that uupc creates in /mail > and do the appropriate things to decode them and create > 'attachments' in GroupWise. No problem. > > Reading the RFC for MIME, it looks like I will have to modify my > outbound headers in order to achieve MIME compliance. The kludge I > spoke of earlier would be something like so: > > 1) Before calling rmail, capture the contents of the /spool > subdirectories to an array of filenames. See below. > 2) Call rmail > > 3) Re-examine the directory to dope out what file(s) rmail created, > then parse through them and modify the headers as appropriate. HUH? WHOA. Why aren't you just writing the headers before you call RMAIL? You should never have to look in the UUCP spool, but rather have RMAIL deliver to you on outbound (forward files, system alias file, VMS queue support, whatever), and RMAIL should leave what you hand it pretty much alone. > I'm having enough success with this whole GroupWise->uupc thing that > I'm going to go ahead and register uupc, since it looks like it'll > become part of my day-to-day life. Let me know if you've ever run > into people looking for a cheap way to gateway GroupWise to the net. Duly cc'ed to the UUPC-Info list for interested parties. -ahd- -- Drew Derbyshire UUPC/extended e-mail: software@kew.com Telephone: 617-641-3452 Phasers locked on target, Captain. ------------------------------ Date: Tue, 7 Mar 95 11:23:14 +0100 From: fae@iitb.fhg.de Subject: looking for voicemail/answering-machine To: UUPC/Extended mailing list I got your email address from some NEWS in comp.mail. We will build a mail-gateway from MIME-audio/basic to voicemail. This can be run on SUN or on MSDOS or MS-windows .. Searching via archie etc we found some answering-machines: We found a software called "am" for NeXT. We found some ZyXEL programm "voicem" for PC. Do you know about some more products ? Do you have some peace of software handling the voice-mode of modems? My interest is buildung gateways between different services: There is a MIME-Gateway email -> fax There is a Gateway email -> call of pagers There is a MIME-Convert-Gateway to . etc. If you are interested in these things, please let me know. I'll send a summary after response to this mail. Regards Hans-Juergen Faerber Fraunhofer Gesellschaft Germany -- Drew Derbyshire UUPC/extended e-mail: software@kew.com Telephone: 617-641-3452 "It's not good for an old memory to mean so much today . . ." - Night Ranger ------------------------------ Date: Tue, 07 Mar 1995 14:20:10 EST From: bacharach!andya@vax.cs.pitt.edu Subject: looking for voicemail/answering-machine To: UUPC/Extended mailing list >From: pitt!iitb.fhg.de!fae >Subject: looking for voicemail/answering-machine >To: UUPC/Extended mailing list [snip] >Do you have some peace of software handling the voice-mode of modems? I just did a hack to UUCICO (1.12n) to allow it to pass non-data calls to another piece of software. (I found a fax program that could be run this way and wanted to try it.) From what little I know about the Zyxel modems, they report something about a voice connect also. Would this hack be of interest to anyone else? Drew, would you be interested in adding this to the regular build of uucico? Andy -- Andy Alexander Bacharach Inc. -- Gas Detection Group 625 Alpha Drive, Pittsburgh, PA 15238 +1 412 963 2102 bacharach!andya@vax.cs.pitt.edu ------------------------------ Date: Tue, 07 Mar 1995 19:27:40 -0500 From: uupcinfo@kew.com Subject: news, etc. To: UUPC/Extended mailing list On Tue, 07 Mar 1995 18:28:50 -0500, "Drew Derbyshire - UUPC/extended Software Support" wrote: > By the way, I see several of these Skipped messages per day. > Apparently (I checked one batch) for some of my received article > batches, the number after the "#! rnews " is a few bytes larger than > the actual article size of one or a few articles in the > batch. Strange. Is it possible that the rnews->newsrun step > (decompression) perhaps strips some trailing blanks or such stuff? Dunno. But it adds a clue for a corrupted gzip message we keep getting on athena. I need to look into it. -- Drew Derbyshire UUPC/extended e-mail: software@kew.com Telephone: 617-641-3452 To sign off from uupc-info, send the command "signoff uupc-info" in the body of a message to listserv@kew.com. DO NOT send this request to the list itself! For human assistance with the list itself, send mail to snuffles@kew.com. "That which thy fathers have bequeathed to thee, earn it anew if thou wouldst possess it." - Goethe "Faust" ------------------------------ Date: Mon, 06 Mar 1995 20:44:25 -0500 From: Software@kew.com Subject: news under 1.12n DOS To: UUPC/Extended mailing list Has anyone actually tried receiving news under 1.12n for DOS? What happened? -ahd- -- Drew Derbyshire UUPC/extended e-mail: software@kew.com Telephone: 617-641-3452 "An important, but often overlooked, feature of running a computer system is maintaining its security from outside intrusion, internal sabotage, and just plain user stupidity." -- Unix System Administration, Fiedler and Hunter ------------------------------ Date: Tue, 7 Mar 1995 22:27:04 -0600 (CST) From: cel@tenet.edu Subject: Outbound question To: UUPC/Extended mailing list I have finally gotten my outbound to work... almost. My local novel login name is 2lce therefore outbound mail is being given the address of 2lce@cbs-engineering.com. I would prefer to have it use an alias. I'm a bit confused on which alias file to edit and exactly what to [ut in it. I would prefer my outbound mail to look like cliff.lee@cbs-engineering.com. What should my Alias entry be and which alias file should I be editing; the one in uupc-in or uupc-out? I'm guesing uupc-out, but attempts to modify that have not yeilded anything of consequence. Cliff Lee cel@tenet.edu ------------------------------ Date: Mon, 06 Mar 1995 11:35:54 ARG From: adt@dia.edu.ar Subject: Protocols To: UUPC/Extended mailing list Maybe it's a naive question, but, which is the full list of protocols supported by uucico, in the last version? Thanks. Best regards. --- Alberto Daniel Teszkiewicz adt@dia.edu.ar ------------------------------ Date: Mon, 06 Mar 1995 12:42:23 -0500 From: uupcinfo@kew.com Subject: Protocols To: UUPC/Extended mailing list On Mon, 06 Mar 1995 11:35:54 ARG, adt@dia.edu.ar wrote: > Maybe it's a naive question, but, which is the full list > of protocols supported by uucico, in the last version? Over serial lines: v, G, g Over network connections, the above plus 't' and 'e'. Note that the 't' protocol module was defective in releases previous to 1.12n. -- Drew Derbyshire UUPC/extended e-mail: software@kew.com Telephone: 617-641-3452 To sign off from uupc-info, send the command "signoff uupc-info" in the body of a message to listserv@kew.com. DO NOT send this request to the list itself! For human assistance with the list itself, send mail to snuffles@kew.com. General Custer plotted strategy on his Series-I ------------------------------ Date: Mon, 06 Mar 1995 14:17:16 -0500 From: Software@kew.com Subject: sendbats discrepancy To: UUPC/Extended mailing list On Sat, 04 Mar 1995 21:50:07 -0500, "Mustafa Soysal" wrote: > It is a bit undesirable for sendbats to batch and compress everything in > sight when invoked. It should accept a system name and only batch that > system - in my opinion. The current default behavior is reasonable, but there is nothing to say it can't also accept one or more system names. However, that's a feature, and I'm still back on the bugs. *sigh* > This keeps the spool uncluttered, and also > makes batching more efficient, allowing more of it to accumulate... If you add the 'B' flag to the third field (this is a UUPC/extended specific flag), sendbats will only send full size batches, as defined by the configuration batch size. As for cluttering the spool, you're gonna either clutter the news directories or the spool, so that part doesn't matter much. By the way, a second undocumented flag is 'c', which suppresses the compression of the batch. -- Drew Derbyshire UUPC/extended e-mail: software@kew.com Telephone: 617-641-3452 Ed Meese for President, Samantha Fox for Vice! ------------------------------ Date: Wed, 08 Mar 1995 07:35:52 -0500 From: Help@kew.com Subject: UUCP 1.12m/n Dos To: UUPC/Extended mailing list On Mon, 27 Feb 1995 16:54:06 +0100, "Johannes Beekhuizen" wrote: > Johannes Beekhuizen schreef aan jou: > > JB> available. Since UUPC 1.12m RNEWS and UUX abort with runtime > JB> error R6800 - stack overflow. This will be corrected in the next release, we're seeing the problem as well. > Just another question: is it intentional that there's a message sent to user > 'news' at the sending site? That's under control of UUXQT, as directed by the originating system. If the sending site does not want to be notified, it can add an option to the generating UUX command to NOT return a status on error. -- Drew Derbyshire UUPC/extended e-mail: software@kew.com Telephone: 617-641-3452 "You can come back baby because Rock'N'Roll never forgets." - Bob Seger ------------------------------ Date: Tue, 07 Mar 1995 21:45:52 -0500 From: Help@kew.com Subject: UUPC 1.12n bug report To: UUPC/Extended mailing list On Tue, 7 Mar 95 7:49:18 +0100, "Jochen Friedrich" wrote: > Hi Drew, > > during my first poll last night, newsrun was hanging in an endless loop. > I checked this out using the debugger and found the following code section > responsible for the endless loop: > > static KWBoolean deliver_remote(const struct sys *node, > IMFILE *imf, > const char *msgID, > const char *path) > { > > /*--------------------------------------------------------------------*/ > /* Verify we make the maximum number of hops limit */ > /*--------------------------------------------------------------------*/ > > if ( strlen( path ) > (node->maximumHops/2) ) /* Near limit? */ > { > size_t hops = 0; > char *p = (char *) path; > > >>>>>>>>>>>> > while( (p = strchr( p++, '!' )) != NULL ) > hops++; > <<<<<<<<<<<< > > The last expression is undefined for the variable p. I assume, the > expression should be: > > while( (p = strchr( ++p, '!' )) != NULL ) > hops++; > > (First increment the pointer, then check for another "!"...) Aw, heck. I KNEW I put that code in. I just didn't know where -- I just rewrote it from scratch in sys.c, this time without the bug. This explains an error someone else had with a loop with L flag. -ahd- -- Drew Derbyshire UUPC/extended e-mail: software@kew.com Telephone: 617-641-3452 This place is more fun than people should be legally be allowed to have. ------------------------------ End of UUPC-Info-Request Digest ******************************