Remove error: true from config (#529)

* Remove `error: true` from config

* Try caching dependencies

* Cache in preview workflow too

* Try a no-op commit

.github/workflows/preview.yml

@@ -25,6 +25,9 @@ jobs:
         with:
           version: '1.10'
 
+      - name: Load Julia packages from cache
+        uses: julia-actions/cache@v2
+
       # Note: needs resolve() to fix #518
       - name: Instantiate Julia environment
         run: julia --project=. -e 'using Pkg; Pkg.instantiate(); Pkg.resolve()'

.github/workflows/publish.yml

@@ -24,6 +24,9 @@ jobs:
         with:
           version: '1.10'
 
+      - name: Load Julia packages from cache
+        uses: julia-actions/cache@v2
+
       # Note: needs resolve() to fix #518
       - name: Instantiate Julia environment
         run: julia --project=. -e 'using Pkg; Pkg.instantiate(); Pkg.resolve()'

_quarto.yml

@@ -5,6 +5,7 @@ project:
     port: 4200
     browser: true
 
+
 # These cannot be used as variables. They are reserved for the project configuration.
 website:
   title: "Turing.jl"

tutorials/_metadata.yml

@@ -6,7 +6,6 @@ format:
 execute:
   echo: true
   output: true
-  error: true
 include-in-header:
   - text: |
       <style>
@@ -16,4 +15,4 @@ include-in-header:
         a:hover {
           text-decoration: underline;
         }
-      </style>
+      </style>

tutorials/docs-17-implementing-samplers/index.qmd

@@ -448,6 +448,7 @@ logpdf(Beta(3, 3), 10.0)
 and so the samples that fall outside of the range are always rejected. But do notice how much worse all the diagnostics are, e.g. `ess_tail` is very poor compared to when we use `unconstrained=true`. Moreover, in more complex cases this won't just result in a "nice" `-Inf` log-density value, but instead will error:
 
 ```{julia}
+#| error: true
 @model function demo()
     σ² ~ truncated(Normal(), lower=0)
     # If we end up with negative values for `σ²`, the `Normal` will error.