Release 1.6.0

[nicolas-grekas]

No description provided.

[dbu]

looks good to me. suggested linking to the RFC for the encoded characters.

as the userinfo encoding is a change in behaviour, and a potential BC break in some super edge cases, i wonder if we should wait a few days if php-fig/fig-standards#1298 coalesces into an agreement and if we change behaviour only change it once...

[nicolas-grekas]

I have zero confidence in the speed of anything when the FIG is involved (except maybe for negative speed ;) )
About the BC break, there can't be any, because the resulting URLs are invalid right now, so this is only a bugfix (as discussed on php-http/discovery#222)

[nicolas-grekas]

I think we should instead cover those edge cases in https://github.com/php-http/psr7-integration-tests/, and encourage mainstream implementations to adopt this test suite. On this we can have a leverage as a community. Not on the FIG.

[dbu]

About the BC break, there can't be any, because the resulting URLs are invalid right now, so this is only a bugfix (as discussed on php-http/discovery#222)

in the discussion in php-fig/fig-standards#1298 it was mentioned that encoding immediately changes what is returned in getUserInfo. that said, as that method concatenates username and password, separated by :, if somebody was going that route they would still have had a bug when the password contains a :...

i guess the changelog is good enough in case somebody hits that edge case. i think my concern is more about xkcd than anything relevant... so lets go.

[nicolas-grekas]

Thanks for jumping into the topic and for the reviews!

[Nyholm]

Thank you

[nicolas-grekas]

Thanks for merging and tagging!

[dbu]

cheers: https://github.com/php-http/psr7-integration-tests/actions/runs/4636945648 👍