@nxgt/material 2.5.0 → 2.6.0

package/docs/troubleshooting.md

@@ -138,6 +138,9 @@ words in your error.
 - [The zoom changes by itself while a long PDF is open](#the-zoom-changes-by-itself-while-a-long-pdf-is-open)
 - [A link in a PDF replaces the page the reader is on](#a-link-in-a-pdf-replaces-the-page-the-reader-is-on)
 - [`onPageChange` is not called when `initialPage` is past the end, or after a new file](#onpagechange-is-not-called-when-initialpage-is-past-the-end-or-after-a-new-file)
+- [`onPageChange` reports another page after opening on `initialPage`](#onpagechange-reports-another-page-after-opening-on-initialpage)
+- [`FilePreview` shows a file card with a Download link for a PDF](#filepreview-shows-a-file-card-with-a-download-link-for-a-pdf)
+- [`FilePreview` says "This file could not be previewed."](#filepreview-says-this-file-could-not-be-previewed)
 - [Reporting a bug](#reporting-a-bug)
 
 ---
@@ -2430,6 +2433,61 @@ before 2.2.0.
 to itself. Since 2.2.0 both are reported, once.
 **Fix:** update `@nxgt/material`. A page you kept in sync by hand can go.
 
+### `onPageChange` reports another page after opening on `initialPage`
+
+**When:** the reader opens on `initialPage` and, without the user scrolling,
+`onPageChange` reports another page, most often the last one, and the page field
+shows it. Seen on slow devices and busy machines, near the end of a
+document, with `@nxgt/material` 2.5.0 and earlier.
+**Why:** after placing the opening page, the reader ignored scrolling for
+only 150 ms. When the placement scroll landed later (a late frame), it was
+read as the user's, and at the end of the pages that is the last page.
+**Fix:** update `@nxgt/material`. The opening page now keeps the page number while
+the view stays on it; the user's own scroll, find in page or a screen
+reader moving the view still report the page they reach.
+
+### `FilePreview` shows a file card with a Download link for a PDF
+
+**When:** a PDF given to `FilePreview` shows its name, size and a Download
+link instead of the reader.
+**Why:** `FilePreview` imports `PdfReader` the first time it previews a PDF,
+and falls back to the file card when that import fails: the viewer's chunk
+could not be fetched (a deploy replaced the chunks the open page was built
+with, or a network error), or the bundle has no `react-pdf` and `pdfjs-dist` — they are
+dependencies of `@nxgt/material`, not optional peers, so this is a bundler
+configuration that left them out. Nothing is thrown; the next PDF previewed
+tries again.
+**Fix:** reload the page after a deploy; otherwise check that your bundler
+keeps `react-pdf` and `pdfjs-dist` and pdf.js's worker (see `PdfReader` in
+[Media & files](components/media-and-files.md#pdfreader)). The same file in a
+`PdfReader` of its own shows the underlying error. A PDF the reader loaded
+but cannot open (a corrupt or password-protected file, a URL that answers an
+error page) is different: the preview stays `ready` and `PdfReader` shows its
+own error inside it.
+
+### `FilePreview` says "This file could not be previewed."
+
+**When:** a `FilePreview` given a URL, or a file it cannot show, renders that
+message (and `onError` receives the error).
+**Why:** a text, JSON, CSV or Markdown URL is fetched with `fetch`: an HTTP
+error is a `FilePreviewHttpError` with its `status`, and a cross-origin URL
+without CORS headers is a `TypeError: Failed to fetch`. An image, video or
+audio file the browser cannot decode (a corrupt file, a codec the browser
+lacks, a URL that answers an error page) fails in its own element. A URL
+that is not `http(s):`, relative, `blob:`, or a `data:` URL of an image,
+video, audio file, plain text, CSV, Markdown, JSON or PDF (`javascript:`,
+`data:text/html`, and also a `data:` URL of XML, TSV or YAML, though
+`FilePreview` would show those kinds from a `Blob`) is refused before
+anything is linked, loaded or fetched: the error is a
+`FilePreviewUnsafeUrlError`, "Refused to preview a URL with an unsafe scheme", and there is no
+Download link.
+**Fix:** serve the file with `Access-Control-Allow-Origin` for your page's
+origin, or fetch it yourself (with your credentials) and pass the `Blob`;
+pass `type` when the server answers `application/octet-stream` for a file
+whose URL has no extension; pass a refused URL's file as a `Blob`, or from
+an `https:` URL. A text file larger than `maxTextBytes` is not an error: its
+beginning is shown, with "Showing the first … of this file." below.
+
 ---
 
 ## Reporting a bug

package/package.json

@@ -1,6 +1,6 @@
 {
 	"name": "@nxgt/material",
-	"version": "2.5.0",
+	"version": "2.6.0",
 	"license": "MIT",
 	"bugs": {
 		"email": "stevegires@gmail.com"
@@ -113,7 +113,7 @@
 		"docs:previews": "vitest run --project previews"
 	},
 	"dependencies": {
-		"@nxgt/material-core": "^0.2.0",
+		"@nxgt/material-core": "^0.3.0",
 		"@base-ui/react": "^1.6.0",
 		"@dnd-kit/core": "^6.3.1",
 		"@dnd-kit/sortable": "^10.0.0",

package/dist/lib/components/ui/fields/slider/slider-ticks.d.ts

@@ -1 +0,0 @@
-export declare function tickValues(min: number, max: number, ticks: boolean | number, step?: number): number[];