🚀 go-pugleaf

Mitsutaka Nakamura (minaka@comd.sharp.co.jp) wrote:
: $B$3$l$i$N$3$H$K$D$$$F!"2?$+>pJs$r$*;}$A$NJ}$*$i$l$^$;$s$G$7$g$&$+(B?

$B?t$+7nA0F1MM$N8=>]$rBN83$7$^$7$?!#(B

...$B$J$s$F$$$&$N$O!V>pJs!W$K$J$i$J$$$G$9$M(B :-)

  $B%X(B_$B%X(B   ----------------------------
$B%_!&!&(B $B%_(B  void@merope.pleiades.or.jp
 (     )$B!A(B          $B$/$5$+$Y$h$&$$$A(B
--------------------------------------

$@;yEg!w5~Bg>pJs!u5~Bg%^%$%3%s%/%i%V$G$9!#(J

<1827@shpask.comd.sharp.co.jp>$@$N5-;v$K$*$$$F(J
minaka@comd.sharp.co.jp$@$5$s$O=q$-$^$7$?!#(J

> $@@hF|!"(Jbindist $@$KF~$C$F$$$?(J rogue $@$r$7$F$$$?$H$3$m!"(J
> $@%W%l!<%d!<$,;`K4$7$F!"(Jscore list $@$rI=<($9$k<jA0$G!"(J
> $@FMA3%^%7%s$,(J reboot $@$7$F$7$^$$$^$7$?!#(J
      $@!E!E!E(J
> $@$3$l$i$N$3$H$K$D$$$F!"2?$+>pJs$r$*;}$A$NJ}$*$i$l$^$;$s$G$7$g$&$+!)(J

$@$3$l$O:G6a$O$"$^$jJ9$-$^$;$s$,!"(J
$@%j%j!<%9D>8e$O(J FAQ $@$@$C$?$+$b$7$l$^$;$s!#(J

fopen(file, "a+") $@$N=hM}$,$^$:$/$F!"(J
$@Mn$A$F$7$^$C$F$$$k$h$&$G$9!#(J

$@:,K\E*$J2r7h:v$G$O$"$j$^$;$s$,!"(J
$@0J2<$N$h$&$J(J patch $@$r(J rogue $@$N%=!<%9Ev$F$k$H!"(J
rogue $@$OLdBj$J$/F0:n$7$^$9!#(J

$@!t(JPC98 $@HG(J 386BSD $@$K$O%Q%C%A:Q$_$N%P%$%J%j$rF~$l$F$$$^$9!#(J

$@$"$H!"%$%s%9%H!<%k@h$N(J owner $@$J$I$d(J permission $@$bCm0U$7$F2<$5$$!#(J


*** score.c.orig	Sat Oct  3 11:41:44 1992
--- score.c	Sat Oct  3 11:43:06 1992
***************
*** 206,212 ****
  
  	md_lock(1);
  
! 	if ((fp = fopen(_PATH_SCOREFILE, "a+")) == NULL) {
  		message("cannot read/write/create score file", 0);
  		sf_error();
  	}
--- 206,213 ----
  
  	md_lock(1);
  
! 	if ((fp = fopen(_PATH_SCOREFILE, "r+")) == NULL
! 	 && (fp = fopen(_PATH_SCOREFILE, "w+")) == NULL) {
  		message("cannot read/write/create score file", 0);
  		sf_error();
  	}


--
$@;yEg(J $@>4!w5~ETBg3X9)3X8&5f2J>pJs9)3X@l96(J kojima@kuis.kyoto-u.ac.jp
         $@5~Bg%^%$%3%s%/%i%V(J             kojima@kmc.kyoto-u.ac.jp

$@!!8E>l$H?=$7$^$9!#(J

  $@3'CN$C$F$$$k$H;W$C$?$N$GCfB<$5$s$K$O%a!<%k$G%U%)%m!<$7$^$7$?$,!"(J
$@K:$l$F$$$kJ}$bB?$=$&$J$N$G!"0l1~(J fj.os.386bsd $@$NJ}$K$b=P$7$F$*$-(J
$@$^$9!#(J

In article <1993Mar29.180032.27777@kuis.kyoto-u.ac.jp>
	kojima@kutsuda.kuis.kyoto-u.ac.jp (& Akira) writes:

> <1827@shpask.comd.sharp.co.jp>$@$N5-;v$K$*$$$F(J
> minaka@comd.sharp.co.jp$@$5$s$O=q$-$^$7$?!#(J
> 
> > $@@hF|!"(Jbindist $@$KF~$C$F$$$?(J rogue $@$r$7$F$$$?$H$3$m!"(J
> > $@%W%l!<%d!<$,;`K4$7$F!"(Jscore list $@$rI=<($9$k<jA0$G!"(J
> > $@FMA3%^%7%s$,(J reboot $@$7$F$7$^$$$^$7$?!#(J
>       $@!E!E!E(J
> > $@$3$l$i$N$3$H$K$D$$$F!"2?$+>pJs$r$*;}$A$NJ}$*$i$l$^$;$s$G$7$g$&$+!)(J
> 
> $@$3$l$O:G6a$O$"$^$jJ9$-$^$;$s$,!"(J
> $@%j%j!<%9D>8e$O(J FAQ $@$@$C$?$+$b$7$l$^$;$s!#(J

  $@$=$NDL$j$G$9!#(JRogue $@$GM7$V$H%9%3%"%U%!%$%k=q$-49$($GMn$A$k$H$$$&$N$O!"(J
FAQ $@$G$7$?!#$3$N8=>]$,6+$P$l$?Ev=i$O!"!V(JRogue $@$J$s$+$GM7$P$:$K(J NetHack
$@$GM7$S$^$7$g$&!W$H$$$&2a7c$J0U8+$b=P$^$7$?$h$M!#(J

  $@860x$O#2$D$"$j$^$9!#(J

(1) pccons.c $@$N5U%9%/%m!<%k$N(J BUG$@!#(J($@$3$l$O(J patchkit $@$r$"$F$k$HD>$k$O$:(J
    $@$G$9!#(J)

(2) open() $@Ey$N(J "a+" $@$N%H%i%V%k!#(J


> fopen(file, "a+") $@$N=hM}$,$^$:$/$F!"(J
> $@Mn$A$F$7$^$C$F$$$k$h$&$G$9!#(J

$@!!:#2s$N%1!<%9$O!"62$i$/$3$l$,860x$G$7$g$&!#CfB<$5$s$K$*J9$-$7$?$H$3$m!"(J
	$@!V(Jpatchkit $@$N(J kernel $@%Q%C%A$O$"$F$F$"$j$^$9!#$3$N8=>]$O(J com2
	$@$+$i(J login $@$7$FM7$s$G$$$?$iH/@8$7$^$7$?!W(J
$@$H$$$&$3$H$@$=$&$G$9$+$i!#(J

$@!!$A$J$_$K!"(Jpatchkit-0.2.2 $@$N(J patch00047 $@$K(J

PATCH FILE:  patch00047
PATCH:          FFLUSH STDIO LIBRARY FIX
AUTHOR:         Dave Rivers (rivers@ponds.uucp)
DESCRIPTION:

	I've discovered the following discrepency in the stdio library
w/ regards to opening a stream using  fopen("file","a+")  and later
calling fflush() on that stream. The fopen() call resets the flags to
__SRW because of the '+' in the mode string.  Later on, fflush() checks
to ensure that the stream being flushed is actually writable.  To do
this, it looks for the __SWR value in the flags, and *not* __SRW.  There
is a comment in stdio.h that implies that __SWR and __SRW will never be
set at the same time.

$@$H$$$&%Q%C%A$,$"$j$^$9$N$G!"$3$l$r$"$F$F(J libc.a $@$r:n$jD>$7!"(Jrogue $@$r(J
$@:F%3%s%Q%$%k$7$?$iD>$k$+$b$7$l$^$;$s!#(J

::::::::

  Rogue $@$N%H%i%V%k4XO"5-;v$rN.$7$^$9!#(J

($@$=$N#1(J)

From: dec@lazarus.nrtc.northrop.com (Dwight E. Cass)
Newsgroups: comp.unix.bsd
Subject: STDIO Mode "a+" screwed up
Message-ID: <39897@gremlin.nrtc.northrop.com>
Date: 27 Jul 92 17:35:48 GMT

Hi folks,

   I have just spent the weekend trying to find out why rogue crashes
when it attempts to update its score file and have found the problem:

   In: /usr/src/lib/libc/stdio/flags.c - Mode "a+" translates to an
       open flag with O_APPEND (in __sflags()).  This means that if
       you seek the file to the beginning and start writting, all the
       written will be APPENDED to the end of the file - rather than
       updating the top of the file as specified by the plus sign.

   What's the cure - ouch.  It seems that __sflags() is called by all of
the open routines and that there is not a clean way of passing back
the fact that append is desired (why pass it back - you have to seek
to the bottom of the file [lseek (fd, 0, SEEK_END)] after the file
is opened (i.e., after you have returned from __sflags()).  So - my
proposal (which I will implement tonight and send patches for
tommorrow), add one more flag to the ->flags work in the FILE *
structure.  This flag, __SUPD 0x4000, will indicate that the file
has been opened in an append mode (i.e., I will always set this flag
for "a", "a+", "ab+", and look at it rather than O_APPEND to make the
decision as to whether to lseek or not.)  We will see what
side-effects this has.

Any other ideas?
/dec

Dwight E. Cass                 dec@nrtc.northrop.com | But there's a hole
Automation Sciences Laboratory, Northrop Corporation |   in the bucket
One Research Park, Palos Verdes Peninsula, CA  90274 | Dear Liza,
(310) 544-5393                                       |   Dear Liza...


($@$=$N#2(J)

From: osynw@warp.mhd.montana.edu (Nate Williams)
Newsgroups: comp.unix.bsd
Subject: First posting of the 386BSD 0.1 unofficial bug-list 1/1
Keywords: info and stuff
Message-ID: <1992Aug7.235255.3116@coe.montana.edu>
Date: 7 Aug 92 23:52:55 GMT


U N O F F I C I A L    3 8 6 B S D     B U G     L I S T
--------------------------------------------------------
Date:		Aug.  7, 1992
Revision:	0.1

    :
========================================
Problem:
	running rogue will reboot the system.
Explanation:
	when rogue updates it's score file, you get a automatic reboot. 
	No panic, no crash dump, nothing but a reboot.
Solution:
	don't play rogue :-(

Submitted by:
	bs@germany.eu.net (Bernard Steiner)
Solved by:

========================================
    :


($@$=$N#3(J)

From: terry@cs.weber.edu (A Wizard of Earth C)
Newsgroups: comp.unix.bsd
Subject: Re: PROG004 bug (Rogue)
Message-ID: <1992Sep21.042416.23382@fcom.cc.utah.edu>
Date: 21 Sep 92 04:24:16 GMT

In article <NwkHr*F00@channelz.GUN.de> swildner@channelz.GUN.de (Sascha Wildner) writes:
>I just read about Rogue rebooting the machine in the bug-list. Here it didn't
>reboot, but would print the message "Can't lock score file -more-" every time
>I tried to quit the game. It appeared that Rogue couldn't find the
>/var/games directory. When I created one, Rogue did no longer complain about
>the score file.
>
>P.S.: I still can't save my games. :(

1)	The reboot is a result of non-reentrancy in the reverse-scrolling
	code.  There was a patch posted here to pccons.c, and it will be
	included in the Alpha-3 code.

2)	modify the open() to "a+" instead of "a"... the file is being truncated.


					Terry Lambert
					terry_lambert@gateway.novell.com
					terry@icarus.weber.edu
---
Any opinions in this posting are my own and not those of my present
or previous employers.
-- 
-------------------------------------------------------------------------------
                                        "I have an 8 user poetic license" - me
 Get the 386bsd FAQ from agate.berkeley.edu:/pub/386BSD/386bsd-0.1/unofficial
-------------------------------------------------------------------------------


P.S. $@$=$&$=$&(J NetHack $@$G;W$$=P$7$?!#;d$O<+Bp$K(J GATEWAY 2000 4DX2-66V $@$r(J
     $@;}$C$F$^$9$,!"$3$$$D$K(J 386BSD $@$H(J XFree86-1.2 $@$rF~$l$F!";n$7$K(J X11
     $@BP1~$N(J NetHack $@$r%$%s%9%H!<%k$7$FAv$i$;$F$_$?$H$3$m!"$=$N$+$C$H$S(J
     $@%9%T!<%I$K6C$$$F$^$9!#(J($@#X$,B.$$$C$9!#(J)

     # 4DX2-66V $@$N(J bytebench $@7k2L$r?M$K8+$;$?$i!V(JSPARCclassic $@AjEv$@!W(J
     # $@$H8@$o$l$?$7!#(J

-- 
$@8E>l(J $@@59T(J ($@$3$P(J $@$^$5$f$-(J)              (((((((
  E-mail: masa-k@sag.hitachi-sk.co.jp   )$@"_(J $@"_(J)
                                       (  v  __
                                        $@!5(J-$@!R(J/

In article <1993Mar30.030737.12084@sag.hitachi-sk.co.jp>
	 masa-k@sag.hitachi-sk.co.jp (Masayuki Koba) writes:

>> > $B@hF|!"(Jbindist $B$KF~$C$F$$$?(J rogue $B$r$7$F$$$?$H$3$m!"(J
>> > $B%W%l!<%d!<$,;`K4$7$F!"(Jscore list $B$rI=<($9$k<jA0$G!"(J
>> > $BFMA3%^%7%s$,(J reboot $B$7$F$7$^$$$^$7$?!#(J
>>       $B!E!E!E(J
>> > $B$3$l$i$N$3$H$K$D$$$F!"2?$+>pJs$r$*;}$A$NJ}$*$i$l$^$;$s$G$7$g$&$+!)(J

$B;yEg$5$s!"8E>l$5$s!"%U%)%m!<$"$j$,$H$&$4$6$$$^$7$?!#(J

>> $B$3$l$O:G6a$O$"$^$jJ9$-$^$;$s$,!"(J
>> $B%j%j!<%9D>8e$O(J FAQ $B$@$C$?$+$b$7$l$^$;$s!#(J

FAQ version 2 (09 Aug 92) $B$OC5$7$?$s$G$9$1$I$M(J(^_^;)

>> fopen(file, "a+") $B$N=hM}$,$^$:$/$F!"(J
>> $BMn$A$F$7$^$C$F$$$k$h$&$G$9!#(J
>
>$B!!$A$J$_$K!"(Jpatchkit-0.2.2 $B$N(J patch00047 $B$K(J
>
>PATCH:          FFLUSH STDIO LIBRARY FIX
>
>$B$H$$$&%Q%C%A$,$"$j$^$9$N$G!"$3$l$r$"$F$F(J libc.a $B$r:n$jD>$7!"(Jrogue $B$r(J
>$B:F%3%s%Q%$%k$7$?$iD>$k$+$b$7$l$^$;$s!#(J

$B$H$$$&$3$H$G!"(Jlibc.a $B$O4{$K:n$jD>$7$F$k$N$G!"$d$C$F$_$k$D$b$j$G$9!#(J
$B$G$b(J reboot $B$NLdBj$,(J kernel $B$K;D$C$?$^$^$K$J$k$N$b$d$G$9$,!#(J

$B!t(J386BSD $B$N(J cpio $B$C$F(J SYSV $B7O$G$OE83+$G$-$^$;$s$M!#(J(;_;)
$B!t(JSUN $B$N(J cpio $B$GE83+$G$-$^$7$?$1$I(J$B!#(J


/* $B!J%+%s%,%k!<$O$*$3$C$F$$$^$9!K(J
 * $BCfB<=<N4(J <minaka@comd.sharp.co.jp>
 */