fix: allow mutating transactions

[jstarry]

Problem

After deserializing a transaction, it's not possible to sign and send because sending the transaction could mutate the transaction's blockhash and cause a mutation error.

Summary of Changes

Allow mutating the recent blockhash of a deserialized transaction

Fixes #25137

[codecov]

Codecov Report

Merging #25141 (d70c81b) into master (05de0e3) will decrease coverage by 6.9%.
The diff coverage is n/a.

@@             Coverage Diff             @@
##           master   #25141       +/-   ##
===========================================
- Coverage    82.0%    75.1%     -7.0%     
===========================================
  Files         610       38      -572     
  Lines      167945     2243   -165702     
  Branches        0      323      +323     
===========================================
- Hits       137868     1686   -136182     
+ Misses      30077      444    -29633     
- Partials        0      113      +113     

[jordaaash]

Hmm, by the same logic, I could see an argument for allowing the fee payer, or even instructions, to be mutated, as long as the transaction hasn't been signed yet. Maybe the change that should be made is not to remove these fields but to make the check specific to signed transactions?

[jstarry]

Hmm, by the same logic, I could see an argument for allowing the fee payer, or even instructions, to be mutated, as long as the transaction hasn't been signed yet. Maybe the change that should be made is not to remove these fields but to make the check specific to signed transactions?

Good point. What do you think about allowing anything in the transaction to be modified? We just trust devs to know that if they modify the transaction, any existing signatures would be invalidated. The main thing that #23720 fixed was ensuring that a deserialized transaction would serialize to the same thing if it wasn't touched.

[jordaaash]

We just trust devs

So crazy it just might work!

But yeah, throwing an error isn't great either, so maybe we just reserialize if changed?

[jstarry]

But yeah, throwing an error isn't great either, so maybe we just reserialize if changed?

Yeah, that's exactly what I'm thinking. Let me know what you think of the updated PR

[jordaaash]

LGTM!