salzman%gaucho@RAND.ORG (Isaac Salzman) (03/03/90)
Well, once again I'm attempting to get the .amsalien thing to work but I'm not having any luck whatsoever. I've done some more purusing of various manuals, help files, etc.... I want to be able to read some of my MH folders using messages. In theory I should be able to: - add ~/Mail to my mspath in preferences - create .amsalien in ~/Mail and then go into CUI and and run scavenge folder-name. Well, here's what happens: CUI> list rush Getting folder list; please stand by... </i/salzman/Mail> (N) rush (N) rush.pend CUI> scav rush Recursively scavenge all subdirectories [No] ? Purge deletions after scavenging [No] ? There is no folder named 'rush'. CUI> So it sees the folder (when you list) but then it doesn't (when you scavenge). Btw, this is the version of Andrew that came with X11R4 (haven't had a chance to patch it up yet), on a Sparcstation-1/SunOS4.03. I ran trace on cui to see what it's actually doing. I noticed that it's looking for (via an access system call) .MS_MsgDir in the folder directory and prints the error just before it complains that there's no folder. Also, there were no calls to anything that tried to access .amsalien - so that check seems to be getting skipped altogether. Nathaniel, once upon a time you sent me a message describing what set debug options to use to try and track down the problems I was having (this was in the beta version), and I can't find the message. Can you send that info to me again? Thanks!! Is anyone else attempting to use this feature???!!! Ciao! * Isaac J. Salzman ---- * The RAND Corporation - Information Sciences Dept. /o o/ / * 1700 Main St., PO Box 2138, Santa Monica, CA 90406-2138 | v | | * AT&T : +1 213-393-0411 x6421 or x7923 (ISL lab) _| |_/ * Internet : salzman@rand.org / | | * UUCP : !uunet!rand.org!salzman | | | * CompuServe: 76167,1046
nsb@THUMPER.BELLCORE.COM (Nathaniel Borenstein) (03/05/90)
I think this one is actually fairly easy to diagnose. Somewhere deep in the bowels of the documentation, I'm pretty sure that it says that an AMS tree root must be a directory whose name starts with ".MESSAGES". That is, it can be ".MESSAGES", ".MESSAGES_foobar", etc. The reason for this restriction is that it is often necessary to find the root of an AMS tree, given a folder path name. The root is now well-defined as the first .MESSAGES* directory that it finds upon climbing up the tree. Given that your root does not follow this convention, I suspect that AMS is unable to figure out where to look for the .amsalien file, and then things break down further from there. My suggestion is that you create a symbolic link from ~/.MESSAGES_MH to ~/Mail, and make ~/.MESSAGES_MH the thing you add to your mspath preference. My guess is that if that does not solve the problem, it will at least get you considerably further along before the next problem rears its ugly head. BTW, I do recommend that you incorporate patches 1-3, especially since you're on a SPARC, for which several core dump bug fixes have been released. Also, in answer to your debugging question: the magic command you want for any messageserver-related debugging is "set debug 0 0 0 -1". You can control the quantity of output by replacing the -1 with some other number, but then you'll have to look at the code to figure out which number you want, which is probably not worth the trouble.... Hope that helps. -- Nathaniel
salzman@RAND.ORG (Isaac Salzman) (03/05/90)
>Nathaniel Borenstein <nsb@thumper.bellcore.com> writes: >Somewhere deep in the bowels of the documentation, I'm pretty sure that >it says that an AMS tree root must be a directory whose name starts with >".MESSAGES". That is, it can be ".MESSAGES", ".MESSAGES_foobar", etc. >The reason for this restriction is that it is often necessary to find >the root of an AMS tree, given a folder path name. The root is now >well-defined as the first .MESSAGES* directory that it finds upon >climbing up the tree. Well, I definitely think that was one of the problems. I actually took it one step further. I sym-linked ~/.MESSAGES to ~/Mail - so MH and messages just share the same tree - period. Bill Cattey gave me the other piece of the answer. In order for .amsalien and scavenge to work you need to first "CREATE" the folder with CUI. I tried this on existing MH folders and it works just fine. If the directory doesn't exist it creates it, if it already exists it just creates .MS_MsgDir and leaves everything else alone. >BTW, I do recommend that you incorporate patches 1-3, especially since >you're on a SPARC, for which several core dump bug fixes have been >released. I've been intending to, but someone stole my disk space!! :-) So I gather with patch level 3 everything works, even when compiling with -O? >Also, in answer to your debugging question: the magic command you want >for any messageserver-related debugging is "set debug 0 0 0 -1". You >can control the quantity of output by replacing the -1 with some other >number, but then you'll have to look at the code to figure out which >number you want, which is probably not worth the trouble.... Thanks, I won't loose the message this time! >Hope that helps. -- Nathaniel Sure does! Thanks again - to you and Bill.... To anyone listening at CMU: you may want to add these little tidbit's of info to the doc's on how to use the .amsalien stuff. Just in case there is one more person out there interested in using it - it'll probably save them a couple hours. :-) -- * Isaac J. Salzman ---- * The RAND Corporation - Information Sciences Dept. /o o/ / * 1700 Main St., PO Box 2138, Santa Monica, CA 90406-2138 | v | | * AT&T : +1 213-393-0411 x6421 or x7923 (ISL lab) _| |_/ * Internet : salzman@rand.org / | | * UUCP : !uunet!rand.org!salzman | | | * CompuServe: 76167,1046
nsb@THUMPER.BELLCORE.COM (Nathaniel Borenstein) (03/06/90)
Excerpts from info-andrew: 5-Mar-90 Re: .amsalien Isaac Salzman@rand.org (2501)
> So I gather with patch level 3 everything works, even when compiling with -O
No, that's a Sun compiler bug, not an Andrew bug. But I think that at
least -O is no longer the default!
salzman@RAND.ORG (Isaac Salzman) (03/06/90)
>Nathaniel Borenstein <nsb@thumper.bellcore.com> writes: >No, that's a Sun compiler bug, not an Andrew bug. But I think that at >least -O is no longer the default! Well yeah, that makes sense. I recall detecting a Sun compiler bug with the beta version in the whitepages code. I don't know if I'd *totally* blame it on Sun's compiler (though there's no excuse for different behavior with or w/o -O). These bugs tend to crop up in some "not so kosher" (and not so portable) code (e.g. passing structs or unions and expecting pointers - that sort of thing), where assumptions are made about what the compiler is generating. I don't think I've seen any mention of this on the list. Has anyone tried building Andrew on a SPARC machine using GCC (GNU C Compiler, where using -O tends not to break your code)? Thanks! -- * Isaac J. Salzman ---- * The RAND Corporation - Information Sciences Dept. /o o/ / * 1700 Main St., PO Box 2138, Santa Monica, CA 90406-2138 | v | | * AT&T : +1 213-393-0411 x6421 or x7923 (ISL lab) _| |_/ * Internet : salzman@rand.org / | | * UUCP : !uunet!rand.org!salzman | | |
wollman@NEWTON.UVM.EDU (Garrett Wollman) (02/01/91)
Does anybody out there use .amsalien for a netnews database? What kind of daemons, etc., do you run for it? [We can't afford the disk space to duplicate our news spool specifically for AMS...] Has anyone worked out some documentation for it (more than the gragmentary note at the beginning of Bboards.ins)? Is there a way to make such daemons work without write permission on the news spool itself? [In just testing it, today, I tried creating symbolic links from ~/.MESSAGES_public/nn (the root is also $LOCAL) under my UID, but it didn't like that idea... -GAWollman
nsb@THUMPER.BELLCORE.COM (Nathaniel Borenstein) (02/01/91)
The bottom line is that it works, but the daemons, as currently set up, are mind-bogglingly slow. Curt Galloway (at SCO) was looking into speeding them up last I talked to him, but I don't know if he actually got anywhere. I'd be glad to provide anyone else who is interested with the same advice I offered him on how to speed up the daemons -- there are some fairly obvious things that could be tried. -- Nathaniel