When {{ with partial "foo" }} returns a falsy value (false, nil, ""),
the with block is skipped but _popPartialDecorator was only called
inside the with block. This left an orphan entry on the decorator
stack, causing "partial decorator ID mismatch" panic on subsequent
partial calls.
Add _popPartialDecorator call to the else branch to ensure the
decorator stack is always balanced regardless of the partial's
return value.
Fixes#14419
The fetchResourceForSpec function was reusing a single ReadSeekCloser
across multiple goroutines, causing a race condition when
TestImageTransformConcurrent runs with imageProcWorkers > 1.
This was exposed by commit e569dd59a ("Misc webp performance work")
which increased imageProcWorkers from 1 to 2.
Fix by creating a new reader on each call to the open function.
Fixes race condition causing segfault on Apple Silicon.
Co-authored-by: Bjørn Erik Pedersen <bjorn.erik.pedersen@gmail.com>
This adds a new file cache named 'modulequeries' with a 24h maxAge
that caches JSON responses from 'go mod download -json' when querying
module versions with constraints (e.g., version = "<v3.0.0").
On subsequent builds, if the cached JSON exists and the module
directory it references is non-empty, the cached result is used
directly, skipping the costly VCS interaction entirely.
Fixes#14417
There are 2 fundamental changes related to alias handling in this PR:
1. the `Page.Aliases`` method returns a ready-to-use list of aliases, not the raw input data. This means that it's prefixed with correct dimension values (e.g. language code) and output format paths, and ready to use in e.g. Netlify's _redirects file.
2. We only render aliases for output formats that is both defined as `IsHTML` and `Permalinkable`.
See #14402
*- Dynamic pool sizing for WebP and Katex based on CPU count
- WebP decode: use RGB (3 bytes) for opaque images instead of RGBA (4 bytes)
- Add new imaging.config with new encoder method option; this is default 4 (what we had before), if have tested with value 2 for a 30% speedup on my MacBook, but that does sacrifice quality in some cases (quality/speed trade-off 0-6)
- Track hasAlpha in decode response to enable RGB optimization
- Scale image processing workers (1->2) on high-core machines
- Add WebP benchmark test
On my MacBook Pro M1 Pro, I now get these numbers compared to Hugo v0.152.0:
```
goos: darwin
goarch: arm64
pkg: github.com/gohugoio/hugo/internal/warpc
cpu: Apple M1 Pro
│ cmp152.bench │ fix-webpbench.bench │
│ sec/op │ sec/op vs base │
Webp-10 162.6m ± 32% 145.3m ± 21% -10.60% (p=0.026 n=6)
│ cmp152.bench │ fix-webpbench.bench │
│ B/op │ B/op vs base │
Webp-10 41.69Mi ± 0% 123.79Mi ± 0% +196.92% (p=0.002 n=6)
│ cmp152.bench │ fix-webpbench.bench │
│ allocs/op │ allocs/op vs base │
Webp-10 9.963k ± 3% 18.797k ± 2% +88.67% (p=0.002 n=6)
```
For general image resize benchmarks compared to Hugo v0.152.0 (note that that's not the result of this branch):
```bash
goos: darwin
goarch: arm64
pkg: github.com/gohugoio/hugo/resources/images
cpu: Apple M1 Pro
│ cmp152.bench │ fix-webpbench.bench │
│ sec/op │ sec/op vs base │
ImageResize-10 84.12m ± ∞ ¹ 40.70m ± ∞ ¹ -51.61% (p=0.029 n=4)
¹ need >= 6 samples for confidence interval at level 0.95
│ cmp152.bench │ fix-webpbench.bench │
│ B/op │ B/op vs base │
ImageResize-10 32.37Mi ± ∞ ¹ 29.25Mi ± ∞ ¹ -9.64% (p=0.029 n=4)
¹ need >= 6 samples for confidence interval at level 0.95
│ cmp152.bench │ fix-webpbench.bench │
│ allocs/op │ allocs/op vs base │
ImageResize-10 11.34k ± ∞ ¹ 12.03k ± ∞ ¹ +6.07% (p=0.029 n=4)
¹ need >= 6 samples for confidence interval at level 0.95
```
Closes#14370
Rename resources/images/exif to resources/images/meta and extend
metadata support to include XMP and IPTC fields. Deprecate .Exif
in favor of .Meta. Update imagemeta dependency and add configuration
options for metadata extraction.
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Closes#13146
Go's stdlid now has a maps package, which is very useful. We have been using imports on the form `xmaps "maps"`, but auto import tools doesn't handle this, which is annoying.
I have been thinking about this for some time, but have been holding back because of all the import changes. However, now is a good time to do this, with very little unmerged code.
This has 2 main benefits:
1. When you only need width/height, we don't need to spin up a WASM runtime.
2. When all processed images are cached in the file cache, we don't need to spin up a WASM runtime.
The above is only really visible for smaller sites, but it's annoying to potentially have a site building in 15ms take e.g. 200ms because of WASM startup time.
Fixes#14371
* The fix for #14357 yesterday sadly had the assumption that default language/version/role always was the first site in the sites matrix. This is common, but not always true.
* Also, that fix forgot to add a redirect the other way when `defaultContentLanguageInSubdir` was disabled, e.g. from `/en/` to `/`.
* This commit also renames config option `DisableDefaultDimensionRedirect` to `DisableDefaultSiteRedirect`. This is stricly a breaking change, but it's only been out for a day, and the old name didn't make much sense.
Fixes#14361
When a new shortcode file was added during `hugo server`, the shortcode
was detected but not properly registered in the template cache. This caused
subsequent content edits using the shortcode to fail with "template for
shortcode not found" until the server was restarted.
The issue was that when a shortcode was added, only a glob identity for
dependent content was added to the changes list, but not the shortcode
file's own pathInfo. This meant RefreshFiles filtered it out and never
inserted it into the shortcodesByName cache.
The fix adds the shortcode file's pathInfo to changes (in addition to the
glob for dependent content), and corrects the glob pattern from
`shortcodes/...` to `/_shortcodes/...` to match the actual path format.
Fixes#14207
* The alias handling was left as is when we added new dimensions in v0.152.0, which meant that if you had `defaultContentVersionInSubDir=true` set, the alias page in the root would not be correctly created.
* Also, if you had multihost setup with multiple dimensions, the publish would be incorrect: The language key was added last instead of first.
* The home page alias handling is reworked and made more robust, which also fixes some subtle issues:
* It now supports redircects for multiple HTML output formats for the home page.
* We skip creating aliases for disabled home pages.
* With potentially multiple output formats with one of them canonical, we added a new `Canonical` method to the `Page.OutputFormats`, which allows you to do this in a template:
```handlebars
{{ with .OutputFormats.Canonical }}<link rel="{{ .Rel }}" href="{{ .Permalink }}">{{ end }}
```
Fixes#14354Fixes#14356
Bjørn Erik Pedersen