serokell / log-warper

Logging library to provide more convenient, extremely configurable but simple monadic interface with pretty output
MIT License
19 stars 11 forks source link

[LW-36] v2.0: Redesign log-warper #108

Closed chshersh closed 6 years ago

chshersh commented 6 years ago

New version of log-warper. Comes under version 2.0.

log-warper has been split into two packages:

Not all parts are finished. But I just want to share this work to agree on final design.

There's v1 branch which is log-warper snapshot. That branch should be used for maintaining older version of log-warper. And, hopefully, will be removed in future.

chshersh commented 6 years ago

@int-index @volhovm I fixed code according to review comments.

Regarding extensible configuration: nice thing to have. But I propose to think about it in parallel with other changes. Personally I don't see fast an easy solution for this problem (since I'm not so experienced in type-level magic; existential constructors like for old HandlerT is not really good solution to me...). So I can open issues for other things that need to be done and then they can be done in parallel, maybe @vrom911 could take some issue or maybe somebody else want to hack on new log-warper.

chshersh commented 6 years ago

Oh, and of course after I fix CI.

gromakovsky commented 6 years ago

There's v1 branch which is log-warper snapshot

Should be protected, I think.

chshersh commented 6 years ago

@avnik

can we finally drop hslogger derived GPL3 licensed stuff?

When log-warper will be completely rewritten in near future there will be no sign of old hslogger and GPL3 license as a consequence. hslogger code is not used, modules not exported, it's just dead code. But it's convenient to have to rewrite and not to forget some things. There's a lot of useful stuff in log-warper at the moment!

chshersh commented 6 years ago

@volhovm

But it makes sense to me to revise this idea before we actually release the 2.0. I mean, we can merge this particular PR, and implement other things after just to minimize PR complexity. I do not find current solution acceptable.

Merging PR to master doesn't mean releasing 2.0 on Hackage. It just opens opportunities for other things to be implemented in parallel.

chshersh commented 6 years ago

@gromakovsky

Should be protected, I think.

Done.

chshersh commented 6 years ago

@volhovm @int-index @gromakovsky @avnik I fixed review comments and made CI pass. I think I can now press Squash and Merge button after somebody approves. And then I can create issues for further plans.