Rename "range" data types to "timestamps"

[razgraf]

Given we now have generic createWithTimestamps methods instead of the old variations (durations, deltas, ranges, milestones), shouldn't we rename the dedicated LockupLinear parameter for start / cliff / end from range to timestamps in the create method (and the DataType itself)?

CC: @sablier-labs/solidity

[PaulRBerg]

I was instinctually on board with this idea, but the following rationale made me change my mind:

• In the general context of Sablier Lockup, "range" is a different concept from "timestamps". It just so happens that in the case of LL, they are one and the same, but "timestamps" is a broader concept.
• In LD and LT
  • "timestamps" = start time + all segments/ tranches
  • "range" = start time + end time
• Thus, changing range to timestamps in LL would mean that the getRange getter turns into getTimestamps
  • This would introduce inconsistency: LL would have getTimestamps and LD and LT getRange
  • And it would be a breaking change WRT V2.1, and we wish to minimize breaking changes (as per the rationale you shared elsewhere, can't remember the discussion)

Nonetheless, it's a valid point, and I suggest we sleep on it some more. We should also ask the auditors for feedback.

[smol-ninja]

I am in favour of this idea (I have also been thinking about this for a while and one more refactor I will start a separate discussion for). The Range struct contains information on timestamps. Even in the case of LD and LT, the range indicates unix timestamps for the segment/tranche start and end time.

From Cambridge dictionary:

Range is the amount, number, or type of something between an upper and a lower limit.

So in my opinion, it makes more sense to use range in the context of durations than the actual timestamps.

My comments/questions on @PaulRBerg's points:

"timestamps" = start time + all segments/ tranches

Did you mean "timestamps" = start time + all segments/tranches timestamps? If so, that's more like a "current timestamp". I am not sure how it can be confused with the Timestamp struct.

"range" = start time + end time

I didn't understand this as well. Can you please explain again? Did you mean range = start unix time + end unix time? If yes then how is it different from the above context?

This would introduce inconsistency: LL would have getTimestamps and LD and LT getRange

We should also rename the range to timestamp in the context of Lockup Linear as well. In LL, the range struct contains three timestamps: start time, cliff time and end time. "Range" keyword can be used to determine stream range such as end time - start time, end time - cliff time and so on.

And it would be a breaking change WRT V2.1, and we wish to minimize breaking changes

That is something I agree with as well.

[PaulRBerg]

Unfortunately, I did not understand most of your points, @smol-ninja. I think the crux of the misunderstanding is this:

• Range = start time unix and end time unix
• Timestamps = start time unix, all segment/ tranche timestamps, and end time unix

The code base considers these as two different concepts.

Thus, let us answer these questions in order to expedite the debate:

1. What structs do we want to modify? Let's be precise.
2. What variables do we want to modify? Again, let's be precise.
3. In LD and LT, do we want to merge the concepts of 'range' and 'timestamps' into a single concept ('timestamps')?
4. If yes, what should the getTimestamps getter return in LD and LT? Just the start time and the end time, or those two + all the timestamps in between?

---

Let me start:

1. All Range structs defined in DataTypes would become Timestamp.
2. All range variables used as function parameters in the create functions.
3. I'm undecided about this. I guess it would be easier to merge the concepts (noting that I changed my mind, when my previous view was that we should continue to keep them separated)
4. It should return just the start time and the end time so that we don't have to define another struct.

[smol-ninja]

Sorry, I initially misunderstood the term "range" as referring to the amount of time that passes (similar to "duration"). I thought they both refer to the same thing in English language yet we use them in different contexts. After having a conversation with ChatGPT, I now understand that "Range" specifically refers to timestamps rather than the length of time. So please disregard my previous comment.

Regarding the proposal to refactor "Range" to "Timestamp," I agree with the first two points you made.

3. Regarding the merge, how do you plan to combine them? Since LL would still require a different struct because of cliff. I suggest keeping them separate until we figure out a way standardize the Range (or new Timestamp) struct within the Lockup library

4. Agreed.

[PaulRBerg]

Thank you for the follow-up, I understand you now.

3. There would be a Timestamps struct in each Lockup* model.

[PaulRBerg]

I suppose we don't have time to include this proposal in V2.2, do we?

[smol-ninja]

I think we can include this. Its just a refactor.

[PaulRBerg]

SGTM

[smol-ninja]

What should we call the variable?

1. LockupDynamic.Timestamp memory timestamp
2. LockupDynamic.Timestamp memory timestampLD

[PaulRBerg]

Simple timestamp should be fine

[smol-ninja]

Issue created:

• Rename Range struct to Timestamp struct #911