default cidv1 to base32

[Stebalien]

ipfs/kubo#6220

[Stebalien]

@whyrusleeping or @anorth,

We'd like to make go-cid return base32 CIDv1 CIDs by default instead of base58 (CIDv0 CIDs will continue to use base58). The argument is that:

1. Not many users are using CIDv1 yet so this shouldn't be to obvious of a change.
2. We'd like to switch to base32 by default eventually as it's case-insensitive. There's no reason not to switch now.

Proposal here: ipfs/kubo#6220

---

As far as I can tell, the go-filecoin project doesn't test against any hard-coded base32 CIDv1 CIDs so this shouldn't have any impact (other than maybe changing the CLI output in some cases). However, I figured you'd like to be aware of this change and have a chance to raise objections.

If you do have any repos with a bunch of base58 CIDv1 CIDs that you'd like to convert, the following perl script works pretty well (but double check the results, it has false positives):

perl -i -e 'sub base32 { my ($cid) = @_; my $res=`ipfs cid base32 $cid`; chomp $res; return $res; }' -pe 's/\b(z[a-zA-Z0-9]{10,})/"@{[base32($1)]}"/ge' **/*

[whyrusleeping]

SGTM, ship it!

[anorth]

No objections, thanks for the heads-up!

…

On Tue, 7 May 2019 at 18:34, Steven Allen ***@***.***> wrote: @whyrusleeping <https://github.com/whyrusleeping> or @anorth <https://github.com/anorth>, We'd like to make go-cid return base32 CIDv1 CIDs by default instead of base58 (CIDv0 CIDs will continue to use base58). The argument is that: 1. Not many users are using CIDv1 yet so this shouldn't be to obvious of a change. 2. We'd like to switch to base32 by default eventually as it's case-insensitive. There's no reason not to switch now. Proposal here: ipfs/kubo#6220 <ipfs/kubo#6220> ------------------------------ As far as I can tell, the go-filecoin project doesn't test against any hard-coded base32 CIDv1 CIDs so this shouldn't have any impact (other than maybe changing the CLI output in some cases). However, I figured you'd like to be aware of this change and have a chance to raise objections. If you *do* have any repos with a bunch of base58 CIDv1 CIDs that you'd like to convert, the following perl script works pretty well (but double check the results, it has false positives): perl -i -e 'sub base32 { my ($cid) = @_; my $res=`ipfs cid base32 $cid`; chomp $res; return $res; }' -pe 's/\b(z[a-zA-Z0-9]{10,})/"@{[base32($1)]}"/ge' **/* — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#85 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AADMW6Q65CMWDZWZ7UFC4K3PUE5IXANCNFSM4HLB5RJA> .

[lidel]

To remove confusion, we should replace Base58 one in README:

-zdvgqEMYmNeH5fKciougvQcfzMcNjF3Z1tPouJ8C7pc3pe63k
+bafzbeigai3eoy2ccc7ybwjfz5r3rdxqrinwi4rwytly24tdbh6yk7zslrm

And the one in tests (multiple occurences in cid_test.go):

-zb2rhhFAEMepUBbGyP1k8tGfz7BSciKXP6GHuUeUsJBaK6cqG
+bafkreie5qrjvaw64n4tjm6hbnm7fnqvcssfed4whsjqxzslbd3jwhsk3mm

[Stebalien]

@lidel I fixed the readme but couldn't find the CIDs you're referring to in the tests.

[lidel]

@Stebalien my bad, I was looking at master. All looks good now!

[hsanjuan]

Heads up: breaking change tagged with patch release. People may rely on what cid strings look like (for me it just breaks sharness tests).

[Stebalien]

Apologies, I should have bumped the minor.