cmd/scion: add a ping flag that allows setting the final packet size

[bunert]

The new flag packet-size allows specifying the size of the packet, which takes the variable header size into account.
The new flag overwrites the payload-size flag similar to how the max-mtu flag handles it, while the max-mtu flag
now overrides the payload-size and the packet-size flag.

---

This change is 

[uniquefine]

Reviewed 1 of 1 files at r2, all commit messages.
Reviewable status: all files reviewed, 3 unresolved discussions (waiting on @bunert and @oncilla)

---

scion/cmd/scion/ping.go line 192 at r2 (raw file):

				if overhead > int(flags.pktSize) {
					return serrors.New("desired packet size smaller than header overhead")

I think it makes sense to display the minimum size.

Suggestion:

					return serrors.New("desired packet size smaller than header overhead", "minimum_packet_size", overhead)

---

scion/cmd/scion/ping.go line 196 at r2 (raw file):

				pldSize = int(flags.pktSize - uint(overhead))

drop this empty line

---

scion/cmd/scion/ping.go line 276 at r2 (raw file):

		`choose the payload size such that the sent SCION packet including the SCION Header,
SCMP echo header and payload are equal to the MTU of the path. This flag overrides the
'payload_size' and 'packet_size' flag.`)

Suggestion:

flags

[oncilla]

Reviewed all commit messages.
Reviewable status: all files reviewed, 7 unresolved discussions (waiting on @bunert)

---

-- commits line 2 at r2:
Let's be concise in the title. It should be easy to parse on a glance.
This is especially important if you skim read the git log

Suggestion:

cmd/scion: add a ping flag which allows setting the final packet size

---

-- commits line 4 at r2:
In networking, the idomatic term is packet, not package.

Suggestion:

packet

---

-- commits line 4 at r2:
we use kebap-case flags.

Suggestion:

max-mtu

---

-- commits line 8 at r2:
This does not add any value really. You can drop it.

I think we need to adapt the "good commit message" example in our internal documentation.
https://go.dev/doc/contribute#commit_messages is what we aim for in the end.

Code quote:

  Changes:
  - additional check if the 'packet-size' is set
  - if desired size is smaller than the required overhead for headers, return an error
  - 'packet-size' overwrites 'payload-size'
  - 'maxMTU' overwrites both

[bunert]

Reviewed 1 of 1 files at r3, all commit messages.
Reviewable status: all files reviewed, 7 unresolved discussions (waiting on @oncilla and @uniquefine)

---

-- commits line 2 at r2:

Previously, oncilla (Dominik Roos) wrote…

Let's be concise in the title. It should be easy to parse on a glance.
This is especially important if you skim read the git log

Done.

---

-- commits line 4 at r2:

Previously, oncilla (Dominik Roos) wrote…

we use kebap-case flags.

Done.

---

-- commits line 4 at r2:

Previously, oncilla (Dominik Roos) wrote…

In networking, the idomatic term is packet, not package.

Done.

---

-- commits line 8 at r2:

Previously, oncilla (Dominik Roos) wrote…

This does not add any value really. You can drop it.

I think we need to adapt the "good commit message" example in our internal documentation.
https://go.dev/doc/contribute#commit_messages is what we aim for in the end.

Done.

---

scion/cmd/scion/ping.go line 192 at r2 (raw file):

Previously, uniquefine (Sebastian Leisinger) wrote…

I think it makes sense to display the minimum size.

Done.

---

scion/cmd/scion/ping.go line 196 at r2 (raw file):

Previously, uniquefine (Sebastian Leisinger) wrote…

drop this empty line

Done.

---

scion/cmd/scion/ping.go line 276 at r2 (raw file):

		`choose the payload size such that the sent SCION packet including the SCION Header,
SCMP echo header and payload are equal to the MTU of the path. This flag overrides the
'payload_size' and 'packet_size' flag.`)

Done.

[uniquefine]

Reviewed 1 of 1 files at r4, all commit messages.
Reviewable status: all files reviewed, 4 unresolved discussions (waiting on @oncilla)

---

-- commits line 4 at r2:

Previously, bunert wrote…

Done.

The title of the PR is still mentions "package"

---

-- commits line 4 at r2:

Previously, bunert wrote…

Done.

The PR description still mentions maxMTU

[oncilla]

Reviewed 1 of 1 files at r4, all commit messages.
Reviewable status: complete! all files reviewed, all discussions resolved (waiting on @bunert)

[uniquefine]

Reviewable status: complete! all files reviewed, all discussions resolved (waiting on @bunert)

[bunert]

Reviewable status: complete! all files reviewed, all discussions resolved (waiting on @bunert)

---

-- commits line 4 at r2:

Previously, uniquefine (Sebastian Leisinger) wrote…

The title of the PR is still mentions "package"

Fixed it. But a question: should the PR title include the first line (https://go.dev/doc/contribute#commit_messages) or is this just about the PR description while the title is just for the PR?

---

-- commits line 4 at r2:

Previously, uniquefine (Sebastian Leisinger) wrote…

The PR description still mentions maxMTU

Should be correct now.