Closed jaredallard closed 1 year ago
I probably should've created an issue for this first 😓, please let me know if you'd like me to do that, otherwise I'm happy to chat about this on here or on the Slack!
This sounds pretty great. I'm busy this weekend but I can take a look tomorrow. Thanks a lot for this proposal!
This sounds pretty great. I'm busy this weekend but I can take a look tomorrow. Thanks a lot for this proposal!
Sorry to ping, but any update on a review? Totally understand if you're busy!
FYI I'm going to be using mage
a lot more, I'd be happy to help drive some of the improvements here if you all have a priority list of some I could tackle! Thanks for the awesome project! ✨
Thank you so much @natefinch! Is there a timeline I could get for releasing this? Eager to get this through asdf
👀
I can make a release tonight:)
What this PR does: This PR removes the usage of
go list
in favor of parsing the files with thego/build
package. This speeds up the discovery of go files from ms to us, which combined withMAGEFILE_HASHFAST
can reduce the time for a re-run ofmage
to ~200ms.I was getting a little irritated with every run of mage being at least 500-600ms due to
go list
taking a min of200ms
and being called twice, go list does a lot more then we were using it for. So, I propose we parse it instead.Notes for my reviewer: I did this pretty quickly, and without that much regard for style. I'm happy to split this up and clean it up if we're OK with the approach. Thank you for reviewing :sparkles:
Related to https://github.com/magefile/mage/issues/404
Performance Measurements
This PR:
1.13.0: