conversion: revert json-marshalling change in #144, marshal hex-format

[holiman]

This undoes the change in json-marshalling, making it so that we (again) marshal uint256.Int in hex-format. This is more widely useful in go-ethereum.

I originally intended to let it be this way, but I changed my mind due to this comment: #144 (comment).

This would be an alternative to ethereum/go-ethereum#28628

cc @fjl @karalabe

[codecov]

Codecov Report

Merging #145 (add7529) into master (f24ed59) will not change coverage.
The diff coverage is 100.00%.

Additional details and impacted files

@@            Coverage Diff            @@
##            master      #145   +/-   ##
=========================================
  Coverage   100.00%   100.00%           
=========================================
  Files            5         5           
  Lines         1643      1643           
=========================================
  Hits          1643      1643           

[karalabe]

But why?

…

On Thu, 30 Nov 2023 at 21:46, codecov[bot] ***@***.***> wrote: Codecov <https://app.codecov.io/gh/holiman/uint256/pull/145?src=pr&el=h1&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=Martin+Holst+Swende> Report Merging #145 <https://app.codecov.io/gh/holiman/uint256/pull/145?src=pr&el=desc&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=Martin+Holst+Swende> (add7529 <add7529>) into master <https://app.codecov.io/gh/holiman/uint256/commit/f24ed59bea89c23941cf073aeb3f702514f3b371?el=desc&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=Martin+Holst+Swende> (f24ed59 <f24ed59>) will *not change* coverage. The diff coverage is 100.00%. Additional details and impacted files @@ Coverage Diff @@## master #145 +/- ## ========================================= Coverage 100.00% 100.00% ========================================= Files 5 5 Lines 1643 1643 ========================================= Hits 1643 1643 — Reply to this email directly, view it on GitHub <#145 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAA7UGLCAG2FEIQHLR452ADYHDPBXAVCNFSM6AAAAABABXLKW2VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTQMZUGQ2DENJRGU> . You are receiving this because you were mentioned.Message ID: ***@***.***>

[holiman]

Because in many places we jump through hoops in order to force our big.Int's to output hex. E.g. ethereum/go-ethereum#28628.

[holiman]

As @fjl noted:

However, let me just say I find the upstream change is not the best. We always want hex output in go-ethereum. So now we need to add a conversion every time just like for big.Int. I thought uint256 is purpose built for our use, why would we make it harder for ourselves?

[lmittmann]

What is the status here? will this change from hex to decimal format be reverted or not? Having decimal marshaling is kind of annoying in the ethereum context and would require a proxy type as hexutil.Big everywhere.

[holiman]

We discussed this, and agreed to not revert it. I implemented a hexutil.U256 type in go-ethereum, and closed this PR.

[lmittmann]

Okay. Just out of curiosity, what is the reason for changing this in the first place?

To give some more context: My lib is effected by this breaking change (I presume it will be relatively easy to fix though). But still, this package has hundreds of direct dependents which code might break or act unexpected. This should be a v2 release IMO or kept as is.

[holiman]

Okay. Just out of curiosity, what is the reason for changing this in the first place?

To be more like big.Int. Or exactly which of the changes are you referring to?

To give some more context: My lib is effected by this breaking change (I presume it will be relatively easy to fix though). But still, this package has hundreds of direct dependents which code might break or act unexpected. This should be a v2 release IMO or kept as is.

Well, depends on how you see it. It doesn't break the programing API, but yes, it does break some places where marshalling is concerned. But many places where you would unmarshal a hex-string, would also accept an int-string.

Anyway, go-ethereum is the main driver for this library.