scw@cepu.UUCP (Stephen C. Woods) (10/11/84)
In article <273@wnuxb.UUCP> netnews@wnuxb.UUCP (Ron Heiby) writes: >I have installed the inews.c.2.pat in a test version of my 2.10.2 >inews (along with the corresponding patches for header.h and header.c) >on my Unix 5.0.5 system (AT&T 3B20S). I'd appreciate it if anyone >out there who has success[...]with digits. When I was doing beta testing >of rn, under 2.10, I had the same problem. I wasn't so careful that >time, and incoming news got botched in the same manner. Help!? Thanks. I have, and I can. I had the identical problem with the Beta test version and just didn't install the patch. But with rn 4.1 came a warning that the dependancies in the make file are wrong, I did some checking and .... Boy are they ever wrong. There is however a quick fix, add the following six(6) lines to your makefile: rparams.h: params.h;@touch rparams.h iparams.h: params.h;@touch iparams.h params.h: defs.h header.h;@touch params.h The reason that this patch is required can be deduced from a 'grep #include *.h' (here with the <...xyz.h> stuff removed). iparams.h:#include "params.h" params.h:#include "defs.h" params.h:#include "header.h" rparams.h:#include "params.h" Many *.o's depend on iparams.h or rparams.h but not on defs.h or params.h or header.h *EVEN THOUGH* they are dragged along. I have not tried to figure out just which module causes the DOXREFS to break, nor have I really attempted to determine the CORRECT dependancy for each *.o. You should put these lines into Makefile.{v7,usg,whatever} as well (just in case you need to run localize.sh). -- Stephen C. Woods (VA Wadsworth Med Ctr./UCLA Dept. of Neurology) uucp: { {ihnp4, uiucdcs}!bradley, hao, trwrb, sdcrdcf}!cepu!scw ARPA: cepu!scw@ucla-cs location: N 34 3' 9.1" W 118 27' 4.3"