CONSLT32@UNOMA1.BITNET (Tim Russell) (07/18/89)
Bob Sloane <SLOANE@KUHUB.CC.UKANS.EDU> writes: > > Do you have any specific evidence that SLP diffs are less reliable than the > patch program or is this just your personal feelings? Other than the BITNET Certainly - SLP uses line numbers ONLY, while patch includes small sections around the changed parts and searches for them, moving the patch around if it needs to, because someone has made a local change. Patches still fail, but not nearly as often, and you're told about it and the failed patch is saved in a file with ".rej" on the end so you can apply it by hand. > problems of folding lines at 80 columns (which would kill the patch program) > what problems have you had with SLP diffs? Why should we depend on a program I've had the problem that patches have failed, mangled up my files, and NOT TOLD me that they did so. Patch lets you know, as I said. > that is unsupported for VMS when we have a program that works and IS > supported? How do we know that this patch program will be any better than SLP > diffs? I think we should stick with something that is a supported product and > that everyone has already. That's one advantage that SLP has, I agree: everyone has it. However, it's a small matter, I would think, to stick patch in with NEWS. SLP is supported? Sure, fill out an SPR and wait for three major releases of VMS to go by. At least we'd have the source to patch. -- ---------------------------------+-------------------------------------------- Tim Russell, Computer Operator | Internet: russell@zeus.unl.edu Campus Computing | Bitnet: russell@unoma1 University of Nebraska at Omaha | UUCP: uunet!zeus.unl.edu!russell ---------------------------------+--------------------------------------------
SLOANE@kuhub.cc.ukans.edu (Bob Sloane) (07/21/89)
Tim Russell <CONSLT32%UNOMA1.BITNET@VM1.NoDak.EDU> writes: >> what problems have you had with SLP diffs? Why should we depend on a program > > I've had the problem that patches have failed, mangled up my files, and NOT > TOLD me that they did so. Patch lets you know, as I said. This problem is addressed in the command file I posted to do updates. Did you look at it? It checksums the input file and makes sure it matches the original input file. It complains if there is a problem. As long as you apply the SLP patches to the same source, you will get the correct results. >> that is unsupported for VMS when we have a program that works and IS >> supported? How do we know that this patch program will be any better than SLP >> diffs? I think we should stick with something that is a supported product and >> that everyone has already. > > That's one advantage that SLP has, I agree: everyone has it. However, it's > a small matter, I would think, to stick patch in with NEWS. I think it is a MAJOR advantage, since I have been unable to locate a VMS version of PATCH. Would you like to convert/write one? I have better things to do than try to fix a problem that we aren't even sure IS a problem yet. Post a version of patch to the net that actually works on VMS, and I will be happy to debate the pro's and con's of using patch vs SLP. Until then, SLP is all we have. > SLP is supported? Sure, fill out an SPR and wait for three major releases > of VMS to go by. At least we'd have the source to patch. I will ask again. Do you know of any errors that should be fixed? All you have pointed out so far is that you have had troubles using the program, not that it doesn't work. As far as I know, EDIT/SUM works just fine is used properly. The problem of it not recoginizing that the input file it was patching was not the correct file is not an error, and is addressed in my UPDATE.COM command procedure by checksumming the input file. If you have something better, lets see it. +-------------------+-------------------------------------+------------------+ | Bob Sloane \Internet: SLOANE@KUHUB.CC.UKANS.EDU/Anything I said is | | Computer Center \ BITNET: SLOANE@UKANVAX.BITNET / my opinion, not my | | University of Kansas\ AT&T: (913) 864-0444 / employer's. | +-----------------------+-----------------------------+----------------------+
icsu6000@MTSUNIX1.BITNET (Mathisen) (07/21/89)
> > Tim Russell <CONSLT32%UNOMA1.BITNET@VM1.NoDak.EDU> writes: > >> what problems have you had with SLP diffs? Why should we depend on a progr > > > > I've had the problem that patches have failed, mangled up my files, and NOT > > TOLD me that they did so. Patch lets you know, as I said. > > This problem is addressed in the command file I posted to do updates. Did you > look at it? It checksums the input file and makes sure it matches the original Actually, this happened to me also... I FTP'd the original distribution from Kuhub, and applied the update procedure, and I couldn't compile the result... This on a MVAX, running VMS 4.7 > > >> that is unsupported for VMS when we have a program that works and IS > >> supported? How do we know that this patch program will be any better than S > >> diffs? I think we should stick with something that is a supported product > and > >> that everyone has already. This is odd, we insist only on using DEC supplied supported tools to manage a non-DEC supplied non-Supported package... Patch could easily be supplied with the News software, and then everybody would have it too. > > > > That's one advantage that SLP has, I agree: everyone has it. However, it's > > a small matter, I would think, to stick patch in with NEWS. I agree... Patch is a far superior method of applying changes to code, with Fuzz factors and other options, including the context-sensitivity, it's a big win... My experience with SLP isn't great, but it distinctly reminds me of receiving changes to Unix source as ed scripts... There are better ways Now somebody just needs to port patch over to VMS. On to other things... 1) Maybe what I'm looking for is in the NEWS software, but how about some statistics tools for the NEWS software, including things like arbitron, volume measurements, who is reading what groups, etc etc ad nauseum... 2) I would like to see a Unix rn-style interface... I've tried several news readers under Unix, and for mucking through a volume of news, nothing beats it. (I haven't worked with NN yet.) The interface with ANU-NEWS, doesn't appear conducive to cranking through 1000+ articles a day. Startup time appears long, and moving around is kind of a pain waiting for news to redraw the screen for everything... Using TPU as a pager seems to be real slow. (I should mention I only maintain the VMS news system here, I read all mine from Unix, so I haven't delved into all the nooks and crannies that might make this easier). Plus you only get appx 18 lines of text, although maybe this is reconfigurable. 3) Maybe a little more documentation on installing the NNTP interfaces could be provided... I never did get it working between my Ultrix 2.3 system and my VMS machine, I had to drop back to Batching the news up on the ultrix side, setting up a news proxy, and then copying the newsbatch files over to the news directories. I'm sure it's probably just some stupid little thing that I'm doing wrong, but I have a lot less VMS experience than Unix, and my Unix stuff works great... Plus I have no VMS manuals, only online documentation, so that hurts too. (Campus is too cheap to buy manuals). 4) Has there been any thought into breaking up the news stuff into different parts, such as $NEWS being the news reader, and maybe a different package to do the management stuff? This is just a thought. Maybe it would speed things up a bit, I'm not sure. -- +-----------------------------------------------------------------------------+ | Jaye Mathisen,systems manager Internet: icsu6000@caesar.cs.montana.edu| | 410 Roberts Hall BITNET: icsu6000@mtsunix1.bitnet | | Dept. of Computer Science UUCP: utah-gr!mts-cs!icsu6000 | | Montana State University PHONE: (406) 994-{4780,3931} | | Bozeman MT 59717 "Digital doesn't have cost-competitive technology. -- | | When it is desparate to keep a position in a market it sues -- Dal Allan | +-----------------------------------------------------------------------------+
SLOANE@kuhub.cc.ukans.edu (Bob Sloane) (07/21/89)
In article <8907202351.AA15388@caesar.cs.montana.edu>, Mathisen <icsu6000@MTSUNIX1.BITNET> writes: > Actually, this happened to me also... I FTP'd the original distribution from > Kuhub, and applied the update procedure, and I couldn't compile the result... > This on a MVAX, running VMS 4.7 It would have been nice if you had mentioned this to me. I actually do try to make the update procedures work, despite all the flak I have been taking. Did you use the original source from kuhub? Several other people have managed to use this procedure without problems. It is difficult to fix problems that aren't reported. > This is odd, we insist only on using DEC supplied supported tools to manage > a non-DEC supplied non-Supported package... Patch could easily be > supplied with the News software, and then everybody would have it too. I don't "insist on using only DEC supplied supported tools," I just insist on using tools that exist. Patch doesn't exist for VMS as far as I can tell, and I have looked on all of the DECUS tapes from 1981 to 1988, as well as several FTP sites of VMS software. > I agree... Patch is a far superior method of applying changes to code, with > Fuzz factors and other options, including the context-sensitivity, it's a > big win... My experience with SLP isn't great, but it distinctly reminds > me of receiving changes to Unix source as ed scripts... There are better ways > Now somebody just needs to port patch over to VMS. Exactly. I have been working on getting the mailing list/newsgroup gateway going, counting votes for the newsgroup creation, recovering from a trashed disk, and doing my regular work, too. I am getting a little tired of people saying patch is god's gift to computers. IT DOESN'T EXIST ON VMS. IF YOU WANT IT, YOU WILL HAVE TO PORT IT YOURSELF, BECAUSE I DON'T HAVE TIME. Is that clear enough??? > On to other things... > > 1) Maybe what I'm looking for is in the NEWS software, but how about > some statistics tools for the NEWS software, including things like arbitron, > volume measurements, who is reading what groups, etc etc ad nauseum... This is something I have been wanting myself. > 2) I would like to see a Unix rn-style interface... I've tried several > news readers under Unix, and for mucking through a volume of news, nothing > beats it. (I haven't worked with NN yet.) The interface with ANU-NEWS, doesn't > appear conducive to cranking through 1000+ articles a day. Startup time appears > long, and moving around is kind of a pain waiting for news to redraw the > screen for everything... Using TPU as a pager seems to be real slow. > (I should mention I only maintain the VMS news system here, I read all mine > from Unix, so I haven't delved into all the nooks and crannies that might > make this easier). Plus you only get appx 18 lines of text, although maybe > this is reconfigurable. There is a certain philosophical difference between unix and VMS. Unix program tend to be small, fast, and obscure, while VMS programs tend to be bigger, slower, and easier to use. You apparently belong to the unix school of thought, while I belong to the VMS school, and apparently so does Geoff. If you want to use rn, write it, or use unix. Please do try to change NEWS, because I happen to think it is already a LOT better than rn in many ways. > 3) Maybe a little more documentation on installing the NNTP interfaces > could be provided... I never did get it working between my Ultrix 2.3 system > and my VMS machine, I had to drop back to Batching the news up on the > ultrix side, setting up a news proxy, and then copying the newsbatch files > over to the news directories. I'm sure it's probably just some stupid little > thing that I'm doing wrong, but I have a lot less VMS experience than Unix, > and my Unix stuff works great... Plus I have no VMS manuals, only online > documentation, so that hurts too. (Campus is too cheap to buy manuals). I agree that the documentation needs more work. I have it down on my list of projects, but it isn't very high on the list right now. > 4) Has there been any thought into breaking up the news stuff into different > parts, such as $NEWS being the news reader, and maybe a different package to > do the management stuff? This is just a thought. Maybe it would speed > things up a bit, I'm not sure. I doubt that this would make the news reader any faster. From what I can tell, NEWS is I/O bound, and reducing memory size wouldn't help that much. +-------------------+-------------------------------------+------------------+ | Bob Sloane \Internet: SLOANE@KUHUB.CC.UKANS.EDU/Anything I said is | | Computer Center \ BITNET: SLOANE@UKANVAX.BITNET / my opinion, not my | | University of Kansas\ AT&T: (913) 864-0444 / employer's. | +-----------------------+-----------------------------+----------------------+
CONSLT32@UNOMA1.BITNET (Tim Russell) (07/23/89)
In article <6116@kuhub.cc.ukans.edu>, Bob Sloane <SLOANE@KUHUB.CC.UKANS.EDU> wri tes: > It would have been nice if you had mentioned this to me. I actually do try to > make the update procedures work, despite all the flak I have been taking. Did > you use the original source from kuhub? Several other people have managed to > use this procedure without problems. It is difficult to fix problems that > aren't reported. Bob, you're taking things too personally! You're not taking flak, SLP is. I don't think there's anyone here who could possibly have anything but the most extreme praise for all the hard work you've done: archiving the NEWS source, answering tons of questions, creating the newsgroup, gatewaying the list and the newsgroup (which is wonderful, I might add!), and working on patches. I've already told you in private mail, but I think it needs to be said here: without your work NEWS would be nowhere near what it is today, and for that we all thank you. > Exactly. I have been working on getting the mailing list/newsgroup gateway > going, counting votes for the newsgroup creation, recovering from a trashed > disk, and doing my regular work, too. I am getting a little tired of people > saying patch is god's gift to computers. IT DOESN'T EXIST ON VMS. IF YOU > WANT IT, YOU WILL HAVE TO PORT IT YOURSELF, BECAUSE I DON'T HAVE TIME. Is that > clear enough??? Crystal, sir! :-) Obviously I wasn't clear enough in my original post: > Also, as far as patches go, I think the easiest thing is to do what Unix >people do: keep track of the "patchlevel" of the software, and require >anyone who wants to be able to automatically apply patches to be up to >the current patchlevel. This will require that there be a patch archive >somewhere, hopefully Kuhub? I think I'm going to look at the sources to >"diff" and "patch" from Unix and see how hard they would be to port. >As it is now, EDIT/SUM will totally mangle your source file and not say >a word. At least patch will fail. Notice: "I think /I'm/ going to look..." I meant me. And sometime this week I'm going to try to get them both working, and then we'll see. I'm sorry for all the confusion and hurt feelings. -- ---------------------------------+-------------------------------------------- Tim Russell, Computer Operator | Internet: russell@zeus.unl.edu Campus Computing | Bitnet: russell@unoma1 University of Nebraska at Omaha | UUCP: uunet!zeus.unl.edu!russell ---------------------------------+--------------------------------------------
SLOANE@kuhub.cc.ukans.edu (Bob Sloane) (07/24/89)
In article <2791@zeus.unl.edu>, Tim Russell <CONSLT32@UNOMA1.BITNET> writes: > Bob, you're taking things too personally! You're not taking flak, SLP is. Thanks for the words of encouragement. Most of the flak has been in Email, people flaming me about being a VMSoid, and how dare I compare a unix product to a VMS product, and that I must be brain damaged to use VMS, anyway. There were some votes against the newsgroup simply because people out there hate VMS and anything connected with it. Personally, I think all the people talking about patch are right, it would be better. It just doen't exist. Of course the proper solution is for Geoff to go to each site, individually, and update the software personally, making any local modification the site wanted while he is there. If Geoff can't do it, I would be happy to, for a small fee. :-) +-------------------+-------------------------------------+------------------+ | Bob Sloane \Internet: SLOANE@KUHUB.CC.UKANS.EDU/Anything I said is | | Computer Center \ BITNET: SLOANE@UKANVAX.BITNET / my opinion, not my | | University of Kansas\ AT&T: (913) 864-0444 / employer's. | +-----------------------+-----------------------------+----------------------+
gih900@UUNET.UU.NET (Geoff Huston) (08/02/89)
Bob Sloane writes: >Of course the >proper solution is for Geoff to go to each site, individually, and update the >software personally, making any local modification the site wanted while he is >there. If Geoff can't do it, I would be happy to, for a small fee. :-) I CAN do that - but, like Bob, it would be for a modest fee :-) Geoff
gih900@UUNET.UU.NET (Geoff Huston) (08/02/89)
Tim Russell writes :-- > I like your suggestion! :-) I imagine Geoff wouldn't mind visiting the >US.. (Or more likely moving here..) But I have a better one! Guess what I >got in the mail today after our router came back up from a power failure. No - the original thought was just fine :-) . I tried to get US DECUS to arrange a Pre-symposium seminar on NEWS for Spring 88 and get them to get me over the Pacific, but it seemed not to get very far. But if you want a personalized NEWS config/debug/customization service let me know :-) Geoff