Announce plans for changing default packet manipulation module to bf_pktpy

[jafingerhut]

Feel free to suggest alternative steps, if these are lacking:

• Announcement:
  • Add near top of README.md the plans to make a backwards-incompatible change within X months from the announcement, which is: changing the default packet manipulation module from scapy to bf_pktpy
  • include a link to instructions on how to select scapy or bf_pktpy as the desired packet manipulation module.
  • If one of the following two lines of code are reached, consider printing a warning that the default will change in release 1.0:
    • https://github.com/p4lang/ptf/blob/main/ptf#L450
    • https://github.com/p4lang/ptf/blob/main/src/ptf/packet.py#L35
• Soon before the next step, make a release to PyPI of the current latest version of ptf, with only a minor version increment from the latest release.
• Simultaneously do these things:
  • change the default packet manipulation module to bf_pktpy
  • increase the major version number
  • make a release with those changes with the new major version number.

Some discussion about reasons for this: p4lang/p4c#5145 (comment)

[fruffy]

What's the status of this issue, should we make the switch soon?

[vgurevich]

Switching from a superior, maintained module to an inferior and unmaintained one is a big decision to make...

[fruffy]

Yes, but that decision has been made quite some time ago. The org has come to the conclusion that we should minimize the use of GPL in the repositories, which includes scapy. By making bf_pktpy the default you stop polluting downstream code with GPL-licensed code.

The impact should be low, downstream users just need to switch the module.

[fruffy]

PoC here: p4lang/p4c#5418