struct

Hwaro::Config::Options::BuildOptions

Inherits Struct / Value / Object

Constructors

new(output_dir : String = "public", base_url : String | Nil = nil, drafts : Bool = false, include_expired : Bool = false, include_future : Bool = false, minify : Bool = false, parallel : Bool = true, workers : Int32 = 0, cache : Bool = false, full : Bool = false, highlight : Bool = true, verbose : Bool = false, profile : Bool = false, debug : Bool = false, error_overlay : Bool = false, cache_busting : Bool = true, stream : Bool = false, memory_limit : String | Nil = nil, env : String | Nil = nil, skip_og_image : Bool = false, skip_image_processing : Bool = false, preserve_output : Bool = false, fast_start : Bool = false, fast_start_count : Int32 = 20, serve_mode : Bool = false)

Instance methods

base_url

base_url=(base_url : String | Nil)

batch_size

cache

cache=(cache : Bool)

cache_busting

cache_busting=(cache_busting : Bool)

debug

debug=(debug : Bool)

drafts

drafts=(drafts : Bool)

env

env=(env : String | Nil)

error_overlay

error_overlay=(error_overlay : Bool)

fast_start

When set (dev-server only), render only the homepage + the N most recent pages on the initial pass; the remaining pages are stashed on the Builder and rendered by a background fiber after the server is already serving. Drops "ready" time on large sites from O(all) to O(N). Always paired with fast_start_count.

fast_start=(fast_start : Bool)

When set (dev-server only), render only the homepage + the N most recent pages on the initial pass; the remaining pages are stashed on the Builder and rendered by a background fiber after the server is already serving. Drops "ready" time on large sites from O(all) to O(N). Always paired with fast_start_count.

fast_start_count

fast_start_count=(fast_start_count : Int32)

full

full=(full : Bool)

highlight

highlight=(highlight : Bool)

include_expired

include_expired=(include_expired : Bool)

include_future

include_future=(include_future : Bool)

memory_limit

memory_limit=(memory_limit : String | Nil)

minify

minify=(minify : Bool)

output_dir

output_dir=(output_dir : String)

parallel

parallel=(parallel : Bool)

preserve_output

Keep existing output files between rebuilds. Used by hwaro serve's watch-triggered rebuilds so repeated full rebuilds don't wipe the already-processed resized images (see ImageHooks#process_images mtime-skip logic, which only works when destinations survive).

preserve_output=(preserve_output : Bool)

Keep existing output files between rebuilds. Used by hwaro serve's watch-triggered rebuilds so repeated full rebuilds don't wipe the already-processed resized images (see ImageHooks#process_images mtime-skip logic, which only works when destinations survive).

profile

profile=(profile : Bool)

serve_mode

True when this build is being run as part of hwaro serve (dev server). Hooks can use this to change behavior (e.g. lazy OG generation).

serve_mode=(serve_mode : Bool)

True when this build is being run as part of hwaro serve (dev server). Hooks can use this to change behavior (e.g. lazy OG generation).

skip_image_processing

skip_image_processing=(skip_image_processing : Bool)

skip_og_image

skip_og_image=(skip_og_image : Bool)

stream

stream=(stream : Bool)

streaming?

verbose

verbose=(verbose : Bool)

workers

Number of concurrent render workers (fibers) for the parallel render phase. 0 = auto (CPU-based, the default). Lowering it (e.g. --jobs 2) reduces effective render parallelism, which on template/Crinja-heavy sites is often FASTER: those pages allocate many small objects, and past ~2 workers Boehm's global allocation lock contends harder than the extra cores help. Markdown-heavy sites keep scaling, so the default stays auto. Never changes output — only render concurrency.

workers=(workers : Int32)

Number of concurrent render workers (fibers) for the parallel render phase. 0 = auto (CPU-based, the default). Lowering it (e.g. --jobs 2) reduces effective render parallelism, which on template/Crinja-heavy sites is often FASTER: those pages allocate many small objects, and past ~2 workers Boehm's global allocation lock contends harder than the extra cores help. Markdown-heavy sites keep scaling, so the default stays auto. Never changes output — only render concurrency.