On travis, we're using VS2017
between onig and sass, it's really a pain in the neck
yup, the image used in the template is
windows-2019-vs2019 so I think that's why onig is failing
will do the job for now, so I can build master with pagination enabled
any idea why this does not work in liquid:
{% assign date = "2019-1" %}
<div>Date: {{ date | date: "%Y/%B" }}</div>I want 2019/January but I got 2019-1
I've looked at the code of the date filter and it seems it does not expect Ruby's Time.parse input (as specified by shopify https://shopify.github.io/liquid/filters/date/)
We've got an issue for parsing for formats
See https://github.com/cobalt-org/liquid-rust/issues/48#issuecomment-354367081
See https://github.com/cobalt-org/liquid-rust/issues/48#issuecomment-354367081
btw this is the image that zola uses to build with onig
https://github.com/getzola/zola/blob/next/azure-pipelines.yml#L14
https://github.com/getzola/zola/blob/next/azure-pipelines.yml#L14
Have we created an issue for the onig-sys crate about supporting vs2019?
looking more closely to onig-sys error: https://dev.azure.com/geobert/cobalt/_build/results?buildId=16&view=logs&j=e119be84-1450-57e0-2330-98c04fcd5c5d&t=03588b7d-ddf4-5e99-3837-03dffb31ba0b&l=315
"Unable to find libclang" so maybe we need a step to install clang?
Seems odd that it is using clang on windows
It could actually be right but I'm guessing its showing something is failing in the compiler detection, so its falling back to that
Seems to be building successfully!
https://dev.azure.com/cobalt-org/cobalt-org/_build?definitionId=2&_a=summary
I went to the extreme on testing combinations. I need to consider where to scale back
Hi there, I'm going to be the elephant in the room and ask a "rude" question: Where is cobalt heading? Obviously, Ed don't have the bandwidth to work on it :( In the meantime, I see Zola flourishing and we're even seeing people migrating to it and I have to admit it begins to annoy me a little bit :-/
I love working with you Ed, though, you teach me a bunch of stuff on Rust and even in computer engineering (even it's my fulltime job for 17 years now)
So I would be sad to give up on this
liquid has potential that despite it not being the most ideal template language, it is one of the few Rust runtime template engines that lets you fully customize it which I know at least
cargo-generate has appreciated
As for cobalt, I have project ideas that zola can't handle and I'm thinking tuning cobalt to be able to handle. I can't remember off my head what those features are for my "private, family focused facebook" but I remember coming across them
While family changes have limited my time quite a bit, another challenge is the number of projects I'm involved in. I am trying to identify pause points for those projects and work towards them.
committed: conventional 1.0 support before pausingtypos: benchmarks ran and formal announcementassert-rs: effectively at a pause pointcargo-release: reduce the number of commits and improve "is a release needed?" detectionstager/cargo-tarball: This is the area of most work needed
Whats not listed are some ideas I have around improving the licensing landscape. Those will also be a relatively big distraction if I move forward with them.
and I hope people are more excited by
typos than they were for committed
(
typos is a source code spell checker targeted at being run in CIs)
Funny enough,
typos was originally going to be my second Rust project for learning the language until Cobalt took over all my time
So that ideas has been sitting around for years
Especially when the goal it to automate Changelogs Oo
which is super duper great
Some people are against auto-generated changelogs. Sometimes it is because people are unfamiliar with how conventional helps. Other times, its because they aren't considering the size of the project, how thinly spread maintainers are, and that sometimes something is better than nothing (looking at you, all those Rust crates that broke compat without any changelog for me to reference to know how to update).