aboutsummaryrefslogtreecommitdiffstats
path: root/wiretap/wtap-int.h
diff options
context:
space:
mode:
authorGuy Harris <guy@alum.mit.edu>2001-08-25 02:56:31 +0000
committerGuy Harris <guy@alum.mit.edu>2001-08-25 02:56:31 +0000
commit837e848329ac0b9ce00438a914d13f9910cc90f3 (patch)
tree7a332f2df44d0595b7a9cac1f4f5c5b08526a3a8 /wiretap/wtap-int.h
parentecccb6e0618fd88921ad9dfe7281b4e7f5ad1179 (diff)
Sigh. Shomiti apparently didn't know that the RFC 1761 data link types
were just DLPI data link types, and didn't know that the list had expanded at some point and that Sun *used* some of the new types (e.g., in atmsnoop), or decided on their own to go beyond those types to encode an Oh-So-Useful link speed indication, or just didn't *care* that they were just DLPI data link types. Therefore, we have to map Shomiti link types to wiretap types using a different mapping table. For now, we assume files with a version number of 2 are snoop files, and version numbers of 3, 4, and 5 are Shomiti files; Shomiti claims to use a version number of 2 as well, but to determine whether a file with a version number of 2 is a snoop file or a Shomiti file requires that we look at the header of the first packet and assume that if there's more than 3 bytes of padding it's a Shomiti file. The return value from "fwrite()" is a "size_t"; make the variable into which we store it a "size_t", and then fix up the bugs that were revealed by the compiler warnings that produced - "fwrite()" returns 0, not a negative number, on an I/O error. svn path=/trunk/; revision=3866
Diffstat (limited to 'wiretap/wtap-int.h')
0 files changed, 0 insertions, 0 deletions