[457] in Kerberos-V5-bugs
Re: compiling beta3 on bsdi v1.1
daemon@ATHENA.MIT.EDU (Theodore Ts'o)
Fri Apr 8 16:49:42 1994
Date: Fri, 8 Apr 94 16:49:21 EDT
From: tytso@MIT.EDU (Theodore Ts'o)
To: "Kurt J. Lidl" <lidl@meketrex.pix.com>
Cc: krb5-bugs@MIT.EDU
In-Reply-To: Kurt J. Lidl's message of Fri, 8 Apr 1994 00:54:51 -0400 (EDT),
<199404080454.AAA18589@meketrex.pix.com>
From: "Kurt J. Lidl" <lidl@meketrex.pix.com>
Date: Fri, 8 Apr 1994 00:54:51 -0400 (EDT)
One thing that has shown up, that I can't explain is the
reference to the "tell" function in lib/krb5/ccache/file/fcc_nseq.c
and fcc_sseq.c files.
tell(fd) returns the offset of the current file pointer relative to the
beginning of the file. It should be equivalent to lseek(fd, 0,
SEEK_CUR), with the exception that lseek() probably causes the kernel to
stop doing read-ahead (because it assumes that you'll be doing
random-access read/writes), while tell() doesn't.
On the other hand, lseek() is POSIX compliant, and tell() isn't. I'll
probably just switch things to use lseek().
- Ted
