[Bug] tcprewrite provides incorrect checksum for certain ipv4 packets about tcpreplay

[ChuckCottrill]

After looking deeper into the cause of this, the assumption which produced the problem, was that a user would always want to change (fix) the header length.
Users may only want to change (fix) the header length when desired. Thus a flag to turn on (default off) this behavior, would avoid surprises for those using 4.4.0, and moving to 4.4.1

[ChuckCottrill]

This issue is related to issue #845
The issue resolves the problems encountered with modifying packet length.
This issue does not address the underlying quesion, what should be the packet length calculation result.

[fklassen]

This issue does not address the underlying quesion, what should be the packet length calculation result.

I would like to mark this issue resolved by PR #843. I'm a bit concerned about this statement. Can you elaborate? Can a new ticket be opened if the packet length calculation is incorrect (other than issue #845)?

[fklassen]

Verified PR #846 fixes this:

~/git/tcpreplay/build Bug_703_844_PR_846_optionally_fix_pkt_hdr_len* ⇡
❯ src/tcprewrite --cachefile=pcaps/pcap.cache \
    --infile=pcaps/cap-original-packet-3.pcap \
    --outfile=pcaps/cap-4.5.0-packet-out.pcap \
    --endpoints=10.200.1.1:10.200.1.2
Warning in ../../src/tcprewrite.c:post_args() line 231:
pcaps/cap-original-packet-3.pcap was captured using a snaplen of 1514 bytes.  This may mean you have truncated packets.
~/git/tcpreplay/build Bug_703_844_PR_846_optionally_fix_pkt_hdr_len* ⇡
❯ diff pcaps/cap-4.4.0-packet-3.pcap pcaps/cap-4.5.0-packet-out.pcap
~/git/tcpreplay/build Bug_703_844_PR_846_optionally_fix_pkt_hdr_len* ⇡
❯

[fklassen]

Fixed in PR #846