4 Commits

Include renames

Author	SHA1	Message	Date
sneak	6d9631a8fc	Build the image pipeline in pixad's fx graph (closes #313) ... The new package internal/providers builds the cache, the fetcher, the processor, the image service, the session manager and the encrypted URL generator from the config. pixad's fx graph provides them with providers.Constructors(), and handlers.New receives them built. The server tests build the same graph, with only the fetcher's dial function replaced. database.New opens the database, so these parts can read it while fx builds them; the cache's stop hook still runs before the database's. Each setting's default is declared in internal/config, and internal/providers works out the default cache_max_bytes. The cache's MaxBytes is the one value for the disk cache limit, and 0 turns the disk cache off. The fetcher and the processor take explicit values. Model: opus-5-5	2026-10-26 11:16:09 +00:00
clawbot	00d8ba26c5	One image request type instead of five (closes #307) ... The new package internal/imagerequest holds the request type, its two parts Source and Transform, the Format and Fit types, the dimension and fit checks, the /v1/image/ path parsing and the variant key; it imports no other package of pixa. imgcache.ParsedURL, imgcache.ImageRequest, signature.Request, imageprocessor.Request, the second Size, format and fit types and the conversions between them are gone. The image processor takes a Transform only. encurl keeps a private struct with the token's short CBOR field names. VariantKey moves with the function that makes it. Tests follow the moved and renamed code; the tests of the removed conversions go with them. Model: opus-5-5	2026-10-21 14:41:07 +02:00
clawbot	8b4dcda549	Keep a side computed for a size with one side 0 at 1 or more (closes #283) ... A size with one side 0 has that side worked out from the source's proportions by integer division, and nothing kept it at 1 or more. When it rounded down to 0, libvips was asked for a square: an 8192x1 source at 100x0 came out 100x100, at 1300x0 it asked for one block of about 39 GB, and some shapes answered 500. Each computed side, and each side contain, inside and outside work out again from the scale, is now at least 1. README.md now says what fit=outside does and what a 0 side means; with fit=outside an 8192x1 source at 100x0 comes out 8192x1, as covering the size requires. Model: opus-5-5	2026-10-16 04:17:59 +02:00
clawbot	afb5dc5d0d	Refuse a fit=outside size whose output exceeds the maximum (closes #257) ... fit=outside scales the source to cover the requested size, and the side scaled past it was held to no maximum: 1x8192?fit=outside of a 3000x600 source was resized to 40960x8192 and encoded. resize now refuses such an output with ErrOutputTooLarge before resizing, so both image routes answer it 400 with the error a size with one side 0 already gets. The other fit modes make the requested size or one within it. README.md and TODO.md say so. Model: opus-5-5	2026-10-13 10:44:45 +02:00