AVM: Modify StackType to provide additional information

[barnjamin]

This pr modifies the existing StackType to allow more precise specification of types, especially those that are common, including a name and value or length bounds.

e.g.
address is a []byte that is exactly 32 bytes long
boolean is a uint64 that is in the range 0-1 (inclusive)

This change also has the benefit of allowing some additional assembly time checks. For example the TEAL snippet

bytes 0xdeadbeef
itxn_field Sender

should fail at assembly time because the byte string we're passing couldn't possibly be an address type.

Major Changes:

1. The existing StackType is now called avmType and is a member of the new StackType struct.
2. The proto function that provides typing for the ops has been updated to account for the new StackTypes
3. A number of the OpSpecs have had their proto argument updated to denote they work on the more specific type.
4. A number of Field specs have been updated to specify that they return the more specific type.
5. The opdoc program has been updated to make use of the new types and fully document them in the langspec and README

[codecov]

Codecov Report

Merging #5130 (d9ca502) into master (f4f5ec6) will increase coverage by 1.57%.
The diff coverage is 74.40%.

@@            Coverage Diff             @@
##           master    #5130      +/-   ##
==========================================
+ Coverage   53.78%   55.35%   +1.57%     
==========================================
  Files         450      454       +4     
  Lines       56201    63898    +7697     
==========================================
+ Hits        30226    35373    +5147     
- Misses      23626    26127    +2501     
- Partials     2349     2398      +49     

Impacted Files	Coverage Δ
agreement/agreementtest/simulate.go	86.66% <ø> (+1.14%)	⬆️
cmd/goal/application.go	18.19% <ø> (+0.40%)	⬆️
config/localTemplate.go	60.60% <ø> (-2.66%)	⬇️
daemon/algod/api/server/v2/handlers.go	0.86% <0.00%> (+0.02%)	⬆️
daemon/algod/api/server/v2/utils.go	11.94% <0.00%> (+1.02%)	⬆️
daemon/algod/server.go	4.30% <0.00%> (+0.38%)	⬆️
data/bookkeeping/genesis.go	26.38% <0.00%> (-0.28%)	⬇️
data/transactions/application.go	61.53% <0.00%> (-4.42%)	⬇️
data/transactions/logic/fields.go	75.50% <ø> (+0.37%)	⬆️
data/transactions/logic/tracer.go	50.00% <0.00%> (-7.15%)	⬇️
... and 44 more

... and 397 files with indirect coverage changes

📣 We’re building smart automated test selection to slash your CI/CD build times. Learn more

[jannotti]

I have to go to a meeting, so this isn't much feedback yet. Main idea that I think would be helpful is if we didn't have to name some of these types. It forced you to shoehorn some things into "hash" or "addr" that really aren't. I'd rather they appear in docs as [32]byte (which would still be a nice improvement over []byte.

I wonder about some others, like key and bigint, would they be better as [..64]byte or something like that? I haven't yet seen, but I assume you have a new section in docs that explains these types?

[jasonpaulos]

Seems like a good improvement, I just have minor comments

[jannotti]

Should we remove the [32]byte line and just hard code a line like:
| [N]byte | len(x) == N | []byte |

I'd probably remove the method time and use [4]byte directly if we did that.

Just looking for thoughts, no need to make changes here. I want to get this in.

[jasonpaulos]

Looks good, just one nit, but still looks good to merge

[jasonpaulos]

nit: I think this function should return false if the stack type is any, based on prior discussion