del@thrush.mlb.semi.harris.com (Don Lewis) (08/08/90)
I just got trn and I'm pretty impressed. I have a couple of questions though. 1) The soft pointer file seems to be incompatible between rrn pl 47 and trn. When you switch between them they have to rebuild the soft pointers. 2) How hard would it be to hack up trn and nntp so that the active2 and thread files could be retrieved from the news server via nntp? -- Don "Truck" Lewis Harris Semiconductor Internet: del@mlb.semi.harris.com PO Box 883 MS 62A-028 Phone: (407) 729-5205 Melbourne, FL 32901
davison%drivax@uunet.uu.net (Wayne Davison) (08/09/90)
Don Lewis (del@thrush.mlb.semi.harris.com) wrote: > 1) The soft pointer file seems to be incompatible between rrn pl 47 and > trn. When you switch between them they have to rebuild the soft pointers. The problem here is that the active and active2 files are not identical in many instances (the active2 file keeps the newsgroup counts that mthreads has processed). The differences are inconsequential to operation, but they do make a difference in the .rnsoft pointers. The obvious solution is to stop using rrn :-). Another solution would be to make the active2 file a more exact "copy" of the active file (which is something that I am likely to do in a future patch). Some other solutions include using a .trnsoft file instead of sharing the .rnsoft file, or totally revamping the new-newsgroup-finding code. > 2) How hard would it be to hack up trn and nntp so that the active2 and > thread files could be retrieved from the news server via nntp? It would be really nice if the nntp support would be enhanced to provide access to non-article files in some manner. In looking at the nntp code, it didn't appear to be too difficult (especially if you're accessing the .thread file in the group's spool directory). Trn's code would be very easy to change. However, don't look to me to twiddle the nntp code. The last thing I need right now is to generate and maintain another off-shoot development project. If someone does something in this area, please get in touch with me. Another possibility that I would like to see explored, which might save transfer time (but not disk space) on the client machines: "clone" the server's thread database as it updates. This would involve sending the completed thread files after they are updated via rsh, uucp, ftp, etc. -- Wayne Davison \ /| / /|\/ /| /(_) davison%drivax@uunet.uu.net davison@drivax.UUCP (_)/ |/ /\|/ / |/ \ ...!uunet!drivax!davison (W A Y N e)
del@thrush.mlb.semi.harris.com (Don Lewis) (08/09/90)
In article <QYQN3ZN@drivax.UUCP> davison%drivax@uunet.uu.net (Wayne Davison) writes: >Don Lewis (del@thrush.mlb.semi.harris.com) wrote: >> 1) The soft pointer file seems to be incompatible between rrn pl 47 and >> trn. When you switch between them they have to rebuild the soft pointers. > >The problem here is that the active and active2 files are not identical >in many instances (the active2 file keeps the newsgroup counts that mthreads >has processed). The differences are inconsequential to operation, but they >do make a difference in the .rnsoft pointers. > >The obvious solution is to stop using rrn :-). Another solution would be >to make the active2 file a more exact "copy" of the active file (which is >something that I am likely to do in a future patch). Some other solutions >include using a .trnsoft file instead of sharing the .rnsoft file, or totally >revamping the new-newsgroup-finding code. Problem solved. I changed mthreads so that it writes C News compatible ten digit article numbers in the active2 file. Now trn and rrn happily coexist. -- Don "Truck" Lewis Harris Semiconductor Internet: del@mlb.semi.harris.com PO Box 883 MS 62A-028 Phone: (407) 729-5205 Melbourne, FL 32901