Skip site navigation (1)Skip section navigation (2)
Date:      Sat, 21 Oct 1995 00:30:26 -0700
From:      Dan Yergeau <yergeau@gloworm.Stanford.EDU>
To:        questions@freebsd.org
Cc:        mcw@hpato.aus.hp.com
Subject:   Re: 1.1.5.1, XFree86-3.1.2, Netscape 1.12 
Message-ID:  <199510210730.AAA02305@gloworm.Stanford.EDU>
In-Reply-To: Your message of "Tue, 10 Oct 1995 16:57:56 PDT." <199510102357.QAA11538@phaeton.artisoft.com> 

next in thread | previous in thread | raw e-mail | index | archive | help

Before anyone suggests installing the nls directory and the
XKeysymDB file, let me assure you that I have copies of those
installed (identical to those in the netscape-1.12 distribution).
ktrace indicates that they are being opened/read.

Apologies in advance if there was a *real* resolution to this that I
missed.  In addition to M C Wong, there seem to have been several
others who have posted about exactly this over the summer (as
archived on minnie).  The problem seemed to be dismissed as "user
error". 

I too am having this problem (i.e. complaints about a missing
XKeysymDB, then nearly endless "translation table syntax error"
messages).  Please note that I am running 1.1.5.1 with the
posted/archived stat/fstat/lstat patches (and XFree86 3.1.[12?]).

Proof that .../nls/C was found

  6883 netscape CALL  open(0xefbfc044,0,0x1b6)
  6883 netscape NAMI  "/usr/X11R6/lib/X11/nls/C"
  6883 netscape RET   open 3
  6883 netscape CALL  #189(0x3,0xefbfbb30)	<-- forwarded fstat
  6883 netscape RET   #189 0
  6883 netscape CALL  break(0x6a2ffc)
  6883 netscape RET   break 0
  6883 netscape CALL  read(0x3,0x2a2000,0x25bb67)
  6883 netscape GIO   fd 3 read 3098 bytes	<-- yep, its 3098 bytes
[contents of file deleted]
  6883 netscape RET   read 3098/0xc1a
  6883 netscape CALL  read(0x3,0x2a2000,0x25bb67)
  6883 netscape GIO   fd 3 read 0 bytes
       ""
  6883 netscape RET   read 0
  6883 netscape CALL  close(0x3)

The attempt for another read after the eof is a little strange, but
I suppose it could be written off as a side effect of the buffer
handling in stdio?

Proof that XKeysymDB was found

  6883 netscape CALL  open(0xefbfdbab,0,0xd0)
  6883 netscape NAMI  "/usr/X11R6/lib/X11/XKeysymDB"
  6883 netscape RET   open 4
  6883 netscape CALL  #188(0xefbfdbab,0xefbfadc8)  <-- forwarded stat
  6883 netscape NAMI  "/usr/X11R6/lib/X11/XKeysymDB"
  6883 netscape RET   #188 0
  6883 netscape CALL  break(0x8feffc)
  6883 netscape RET   break 0
  6883 netscape CALL  read(0x4,0x8fc000,0x2000)
  6883 netscape GIO   fd 4 read 4674 bytes	<-- yep, its 4674 bytes
[contents of file deleted]
  6883 netscape RET   read 4674/0x1242
  6883 netscape CALL  close(0x4)

At this point, the complaint about the Motif keysyms is printed.
Unless ktrace is lying, the files were found and read.  I'm guessing
that netscape didn't like some difference between the information it
got from the read and from the stat/fstat.

Perhaps the stat/ftat/lstat patch doesn't quite work.  Didn't struct
stat change in 4.4?  The move of stat/fsat/lstat to 188-190 would
make a lot more sense to me if it did.

I realize that 2.1-RELEASE will be out "any day now," and I will
jump onto 2.1 as soon as I know that there aren't any serious
release glitches.


Dan Yergeau
yergeau@gloworm.stanford.edu



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199510210730.AAA02305>