Nestor -- 1.0 PL0
=================
This version was announced in the  newsgroups  "news.software.nntp"  and
"news.software.b"  and released to the Usenet community by anonymous FTP
and  mail-server  on   May   23,   1991.    Announcement   appeared   in
"comp.archives" on May 25, 1991.


Nestor -- 1.0 PL1
=================
Patch #1 was posted in "news.software.nntp" and  "comp.sources.bugs"  on
June 3, 1991.

Topic:	added the ``Changes'' file. You looking at it now...
From:	EHK
Added:	Patch #1 [Changes Makefile]

Topic:	New sections in the README file about bug-reports and archiving
From:	EHK
Added:	Patch #1 [README nestor.1]

Topic:	Documented ``LOG_LOCAL7'' is somewhat surprising for 4.3BSD users
From:	Gordon Moffett <netnews@netcomsv.netcom.COM>
Added:	Patch #1 [README]

Topic:	SunOS 4.1.1 warned about `illegal combination of pointer and integer'
From:	Gordon Moffett <netnews@netcomsv.netcom.COM>
Added:	Patch #1 [nestor.c]

Topic:	Use <stdin> as input if the filename argument is a "-"
From:	Ian Dickinson <vato@warwick.ac.uk>
Added:	Patch #1 [nestor.c nestor.1]

Topic:	Gets confused about some not-covered entries in the syslog
From:	Ian Dickinson <vato@warwick.ac.uk>
	Gordon Moffett <netnews@netcomsv.netcom.COM>
	Emmett Hogan <hogan@csl.sri.com>
Added:	Patch #1 [scan.c]

	Nestor gets confused about some messages in the syslog that were
	redirected from stderr i.e.:
	   nntpd[82]: /usr/lib/newsbin/inject/inews: defhdrs.awk: no Subject
	and also mis-interpreted other syslog info i.e.:
	   nntpd[82]: loadav at 6, exiting

	Fixed this to change the algorithm in ScanLine(). After retrieving
	the hostname and the keyword, *first* lookup the keyword; if that
	is successful, the info in front of it is most likely a hostname,
	so it is it reasonable to proceed.

Topic:	Also report remote-postings if there a no remote-readings
From:	Ian Dickinson <vato@warwick.ac.uk>
Added:	Patch #1 [report.c]

	If you have a couple of news clients posting via the nntp mini-inews
	program, but reading articles over NFS, the remote posting statistics
	were not shown.

Topic:	Now report the number of logfile lines successfully parsed
From:	EHK
Added:	Patch #1 [nestor.c scan.c]


Nestor -- 1.0 PL2
=================
Patch #2 was posted in "news.software.nntp" and  "comp.sources.bugs"  on
June 12, 1991.

Topic:	Typo in the manual page and change in presentation of usage information
From:	Bart Muyzer <bartm@cv.ruu.nl>
Added:	Patch #2 [nestor.1 nestor.c]

Topic:	Also install the manual page during a `make install'
From:	Bart Muyzer <bartm@cv.ruu.nl>
Added:	Patch #2 [Makefile README]

Topic:	nestor memory/segmentation faults on certain input lines
From:	Nike Horton <horton@cse.ogi.edu>
Added:	Patch #2 [nestor.h scan.c]

Topic:	current time was always reported as 'statistics end date'
From:	EHK   [on request from Ian Dickinson <vato@warwick.ac.uk>]
Added:	Patch #2 [nestor.h nestor.c]

	Even on old logfiles, the current time was reported as  end  time
	of  the  statistics  --  this is not what you want. So, at EOF we
	want to remember the last nntp/nntpxmit syslog  entry.   Since  I
	don't  want  to  strcpy()  every  entry,  I  decided  to  use two
	alternating buffers while reading the syslog  file.   This  makes
	sure that we always still have the last entry of interest at EOF.

Topic:	'nntplink' not fully supported -- now documented
From:	EHK   [on report by John Palkovic <palkovic@linac.fnal.gov>]
Added:	Patch #2 [nestor.1]

	Besides 'nntpxmit', there's  another  program  to  take  care  of
	Usenet  news  transmissions: NNTPLINK.  Although nestor correctly
	picks up the statistics that are logged  by  'nntplink',  verbose
	mode ('-v') will show that it doesn't like many of the additional
	'nntplink' log entries. This is now documented in the manual page
	in the "CAVEATS" section.
	  I  hope  to better support 'nntplink' in future versions, but I
	have to dig into 'nntplink'  first  in  order  to  find  out  its
	perculiarities.


Nestor -- 1.0 PL3
=================
Patch #3 was posted in "news.software.nntp" and  "comp.sources.bugs"  on
July 8, 1991.

Topic:	first field should allow machine names of 23+ characters
From:	Steven Goodman <manager@spool.mu.edu>
Added:	Patch #3 [report.c]

	Steven  sent me a  modified  "report.c"  that  would  allow  more
	character  positions  for  the  hostname. The report indeed looks
	a lot nicer now. Thanks Steven!


Nestor -- 1.0 PL4
=================
Patch #4 was posted in "news.software.nntp" and  "comp.sources.bugs"  on
January 5, 1992

Phew,  it's been a long time...  sorry folks, I simply couldn't find the
time and energy  to hack it.  Anyway,  now that I bought a NeXT at home,
times are changing <:-)     [what? a NextStep front-end? well, maybe...]

Topic:	performance stats should give a percentage, not a ratio failed/succesful
From:	Ian Dickinson <vato@warwick.ac.uk>
Added:	Patch #4 [report.c]

Topic:	first attempt to add support for NNTPXFER statistics
From:	Ian Dickinson <vato@warwick.ac.uk>
Added:	Patch #4 [nestor.h nestor.c report.c scan.c nestor.1]

	Ian loved  to have some  NNTPXFER statistics support,  so he sent
	me the code to handle it -- thanks Ian.

Topic:	a different compiler told me about some sloppy code -- fixed
From:	EHK
Added:	Patch #4 [nestor.h scan.c]

Topic:	nntplink kept the connection open for more than 70 years :-)
From:	EHK [on report by Bill Trost <trost@reed.edu>]
Added:	Patch #4 [scan.c]

	I got a logfile that showed  NNTPLINK to keep the connection open
	open  for quite a long time,  thereby causing an  elapsed time of
	of more than 70 years... So, now we treat an elapsed time of more
	than twenty-four hours bogus (is that a reasonable value, folks?)
	and we watch out for zeroes as well :-(

Topic:	added a missing sanity check around an sscanf()
From:	EHK
Added:	Patch #4 [nestor.h scan.c]

Topic:	assertions in some of the report routines to prevent header-only reports
From:	EHK
Added:	Patch #4 [report.c nestor.1]

	Non-existent  NNTP  transfer statistics  (e.g. NNTPXFER not being
	used at all)  caused a "header-only" report to be printed.  Fixed
	this by first checking if there's any work to do at all.

Nestor -- 1.0 PL5
=================
Patch #5 was posted in "news.software.nntp" and  "comp.sources.bugs"  on
January 11, 1993

Topic:	if all offered articles were rejected, the connection was not shown
From:	Ian Dickinson <vato@csv.warwick.ac.uk>
Added:	Patch #5 [report.c]

Topic:	The report now contains totals for each section
From:	Ian Dickinson <vato@csv.warwick.ac.uk>
Added:	Patch #5 [report.c]

        Ian  really  wanted the totals,  so he sent me the code to handle
        this -- once again, thanks Ian.

Topic:	"msgid" syslog messages on slow feeds confuse nestor
From:	Ian Dickinson <vato@csv.warwick.ac.uk>
Added:	Patch #5 [scan.c]

Topic:	"inews" pathnames slip in occasionally; nestor thinks it's a hostname
From:	EHK
Added:	Patch #5 [scan.c]

Topic:	Add extra test to prevent useless work if no statistics found at all
From:	EHK 
Added:	Patch #5 [nestor.c report.c]

	To be honest with you... "useless work" in the topic line  above
	may be read as "core dump" on some architectures :-)

Topic:	'make spotless' should also throw away the local "defs.h" file
From:	EHK
Added:	Patch #5 [Makefile]

Topic:	text following #endif violates ANSI standard
From:	EHK
Added:	Patch #5 [support.c]

Topic:	changed PATCHLEVEL into a string, to simplify users making local changes
From:	EHK [on request by various people]
Added:	Patch #5 [patchlevel.h nestor.c]

