[news.software.b] problems with rn and rrn

rusty@weyl.Berkeley.EDU (Rusty Wright) (10/10/87)

Where does one get a good nntp distribution from?  I've copied the rn
sources from uunet.uu.net, copied all of the patches from there, and
applied all of the patches.  I then copied the nntp.tar.Z file from
uunet.uu.net, uncompressed and extracted it, copied everything from
the rn directory into the nntp/rrn directory and ran the
MAKE_RRN_CANGES shell script (it runs patch) but not all of the
patches are correct; i.e. patch generates the following .rej files:

Configure.rej
Makefile.SH.rrn.rej
bits.c.rej
init.c.rej
ngdata.c.rej
rcstuff.c.rej

Should I just ignore this and blithely proceed?

And while I'm pointing out problems with distributions, one (or more)
of the patch files for rn generates a .rej file.

--------------------------------------
	rusty c. wright
	rusty@weyl.berkeley.edu ucbvax!weyl!rusty

evan@ndcheg.UUCP (Evan Bauman) (10/11/87)

In article <1316@cartan.Berkeley.EDU>, rusty@weyl.Berkeley.EDU (Rusty Wright) writes:
> Where does one get a good nntp distribution from?  I've copied the rn
> sources from uunet.uu.net, copied all of the patches from there, and
> applied all of the patches.  I then copied the nntp.tar.Z file from
> uunet.uu.net, uncompressed and extracted it, copied everything from
> the rn directory into the nntp/rrn directory and ran the
> MAKE_RRN_CANGES shell script (it runs patch) but not all of the
> patches are correct; i.e. patch generates the following .rej files:
> 
> Configure.rej
> Makefile.SH.rrn.rej
> bits.c.rej
> init.c.rej
> ngdata.c.rej
> rcstuff.c.rej
> 
> Should I just ignore this and blithely proceed?


I copied the same files from uunet.uu.net and was unable to build rrn.
It seems that the patch files do not exactly coincide with the source
available on uunet (although I have the highest regard for L. Wall's
Patch program).

I needed to do some detective work to apply some of the patches.  It wasn't
too difficult as I used the textedit tool on the Sun.  Be very sure that 
the Makefile is patched accordingly.

In the end, I was able to build rrn after manually deleting references
to /usr/include/sys/jioctl.h which is not in my version of Unix.
Users can use rrn to read news over the ethernet, but the posting
method still does not work.  After the user tells rrn to send the article,
I get a message that the server rejected the posting.  I think that my
nntp_access file is correct.  Anyone have any ideas??

	Evan Bauman
	Univ. of Notre Dame
	..!iuvax!ndmath!ndcheg!evan

hyc@umix.cc.umich.edu (Howard Chu) (10/14/87)

The basic problem is just that the released nntp sources (1.3, I think?)
were written for rn at patchlevel #29 or thereabouts, and rn is up to 39 or
40 now. I built our rrn based on a level 39 rn, and just went thru the .rej
files and applied patches by hand. (Thank god for multiple windows! Love
'dem Apollos...) It wasn't too difficult, although there are occasional
sections of code that the patches try to tweak which are completely
discarded in patchlevel 39 of rn...

I've also discovered, thru various user complaints, that the patches
provided don't change everything that can be affected by the remote server
setup. For instance, the save-article routine always invokes the saver
script with a filename of the form spooldir/news/group/article. When
using a remote rn, however, an article is always saved with the name
/tmp/rrn{pid}.{article}. Simple enough to fix, but also a rather annoying
oversight... (The saver scripts complain about the article not existing,
and then return saying the article was saved anyway. All that's saved
is 2 blank lines...) (munge up the definition of '%A' in intrp.c to take
care of this...)

Aside from these minor stumbling blocks though, I found it rather a
convenient thing to use, and have installed it successfully on Vax,
Apollo, and Sun systems...
-- 
  -- Howard Chu
	"Of *course* it's portable. It's written in C, isn't it?"