I don’t know for sure how it is impacted yet, as I’m still learning. It’s still been better memory management than most other languages. When I was still using it in a Ubuntu docker container, was getting around 75 MB a day or so, even with doing my own calls to GC and using streams and other niceties that ai recommended.
I’m now hosting on FreeBSD in a jail, so will be monitoring how well it does there. Also made some more adjustments with image fetching that could improve memory things too. I had just assumed XML parsing in general had some memory issues.
I don’t think I have the capacity to work as a maintainer.
But you’re welcome to fork the repo and continue development under a different user/org. And I’m happy to support such efforts with feedback, reviews, etc.
I was just not wanting to offend if I did the fork and rename path, so wanted to check with you prior. I somewhat want to have a learning experience too. Already have an idea for the name.
Have some interesting feature ideas, but we will see how realistic they are. I’ve done a little bit of a SSG of project in python for an online learning platform, so have a little bit of idea of core concepts at least.
Still working on getting specs all passing and such, but got it compiling. Believe it or not, ended up making a new sass library based on the new dart sass. Since at the minimum I’d like to get my old Jekyll sites compiling, forgot they use sass and there are tons of active Jekyll themes using it too. I have some other ideas that I’m exploring that might be cool if it works out. Not sure if either shard is ready for use yet, but thought I’d give an update.
I certainly respect the previous work on this stuff. Even with SWE ai agents helping, taking a bit to progress. I’m trying to get this fork to be kind of a drop in replacement for Jekyll. So you would run the executable and it would compile it without config file changes. We shall see how well it goes.
Kind of splitting out Jekyll plugin equivalent parts to handle popular plugins. Learned we don’t have circular dependency functionality. I would have liked to have it in an optional plugin or plugins potentially but have it still require my app in its setup.
Is there any mechanism to have special dependencies like a shards install plugins or something like that? Not sure if a makefile would be ideal since you have to have the dependencies listed prior in the shard file. Would be interesting if you could still somewhat taylor a binary to what a user needs. Guess this is a compiled language problem.
Also found some possible small issues in the liquid.cr library, but waiting till I’m finished with this rewrite to confirm whether they are truly bugs/problems or not.
