fix: merge per-glyph CID text and extract Form XObject content

CHANGELOG.md

@@ -9,6 +9,21 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ---
 
+## [0.8.2] - 2026-05-21
+
+### Fixed
+- **Per-glyph text merging** — PDFs generated by wkhtmltopdf that emit one `Tj` operator per glyph now produce correctly merged words instead of `"D O M I C I L I O"` (#74)
+  - New `mergeAdjacentElements()` collapses consecutive single-glyph TextElements sharing font/size/line into word-level elements
+  - New `AssembleText()` replaces naive concatenation with spatial gap detection for word and line boundaries
+- **Form XObject text extraction** — PDFs generated by TCPDF that place all page content inside Form XObjects (`Do` operator) now extract text correctly instead of returning empty strings (#74)
+  - New `processFormXObject()` with resource stack push/pop and 8-level recursion guard
+  - Handles nested XObjects and mixed direct/XObject content on the same page
+
+### Added
+- **Test fixtures** — Real-world CID TrueType shipping label PDFs from @iv7dev added to `testdata/pdfs/cid-fonts/` for regression testing
+
+---
+
 ## [0.8.1] - 2026-05-21
 
 ### Fixed

ROADMAP.md

@@ -130,6 +130,14 @@ Page sizes, custom dimensions, landscape orientation, and text rotation:
 
 ## Current Development
 
+### v0.8.2
+
+**Released**: May 2026
+
+- **Per-glyph text merging** (#74) — wkhtmltopdf single-Tj-per-glyph output now merged into words via spatial gap detection
+- **Form XObject extraction** (#74) — TCPDF content inside `Do` Form XObjects now extracted with resource stack and recursion guard
+- Real-world CID TrueType test fixtures from @iv7dev
+
 ### v0.8.1
 
 **Released**: May 2026

document.go

@@ -306,11 +306,7 @@ func (d *Document) ExtractTextFromPage(pageNum int) (string, error) {
 		return "", fmt.Errorf("failed to extract text from page %d: %w", pageNum, err)
 	}
 
-	var result string
-	for _, elem := range elements {
-		result += elem.Text + " "
-	}
-	return result, nil
+	return extractor.AssembleText(elements), nil
 }
 
 // ExtractTextElementsFromPage extracts positioned text elements from a specific page.