fix PyPDF2 3.x page copying + rename documentation

- Added explicit page copying after cloneReaderDocumentRoot() calls
- Renamed PATCH_PDFWRITER.md to PATCH_PYPDF2_PDFWRITER.md
- Prevents 327-byte empty PDFs in PyPDF2 3.x

🤖 assisted by claude

odoo-bringout-oca-ocb-account_edi/account_edi/models/ir_actions_report.py

@@ -32,6 +32,10 @@ class IrActionsReport(models.Model):
                     # Post-process and embed the additional files.
                     writer = OdooPdfFileWriter()
                     writer.cloneReaderDocumentRoot(reader)
+                    # Copy all pages from the reader to the writer (required for PyPDF2 3.x)
+                    for page_num in range(reader.getNumPages()):
+                        page = reader.getPage(page_num)
+                        writer.addPage(page)
                     for edi_document in to_embed:
                         # The attachements on the edi documents are only system readable
                         # because they don't have res_id and res_model, here we are sure that

odoo-bringout-oca-ocb-account_edi/account_edi/models/ir_actions_report.py.backup

@@ -0,0 +1,47 @@
+# -*- coding: utf-8 -*-
+
+import io
+
+from odoo import models
+from odoo.tools.pdf import OdooPdfFileReader, OdooPdfFileWriter
+
+
+class IrActionsReport(models.Model):
+    _inherit = 'ir.actions.report'
+
+    def _render_qweb_pdf_prepare_streams(self, report_ref, data, res_ids=None):
+        # EXTENDS base
+        collected_streams = super()._render_qweb_pdf_prepare_streams(report_ref, data, res_ids=res_ids)
+
+        if collected_streams \
+                and res_ids \
+                and len(res_ids) == 1 \
+                and self._is_invoice_report(report_ref):
+            invoice = self.env['account.move'].browse(res_ids)
+            if invoice.is_sale_document() and invoice.state != 'draft':
+                to_embed = invoice.edi_document_ids
+                # Add the attachments to the pdf file
+                if to_embed:
+                    pdf_stream = collected_streams[invoice.id]['stream']
+
+                    # Read pdf content.
+                    pdf_content = pdf_stream.getvalue()
+                    reader_buffer = io.BytesIO(pdf_content)
+                    reader = OdooPdfFileReader(reader_buffer, strict=False)
+
+                    # Post-process and embed the additional files.
+                    writer = OdooPdfFileWriter()
+                    writer.cloneReaderDocumentRoot(reader)
+                    for edi_document in to_embed:
+                        # The attachements on the edi documents are only system readable
+                        # because they don't have res_id and res_model, here we are sure that
+                        # the user has access to the invoice and edi document
+                        edi_document.edi_format_id._prepare_invoice_report(writer, edi_document)
+
+                    # Replace the current content.
+                    pdf_stream.close()
+                    new_pdf_stream = io.BytesIO()
+                    writer.write(new_pdf_stream)
+                    collected_streams[invoice.id]['stream'] = new_pdf_stream
+
+        return collected_streams

odoo-bringout-oca-ocb-account_edi/doc/PATCH_PYPDF2_PDFWRITER.md

@@ -22,26 +22,34 @@ The account_edi module uses PyPDF2 for:
 
 ## Solution
 
-**This package requires NO direct patches** because it uses:
+**This package includes direct fixes** for PyPDF2 3.x page copying issue:
-1. `OdooPdfFileWriter` from `odoo.tools.pdf` (oca-ocb-base)
+1. Uses `OdooPdfFileWriter` from `odoo.tools.pdf` (oca-ocb-base) for compatibility
-2. `OdooPdfFileReader` from `odoo.tools.pdf` (oca-ocb-base)
+2. **CRITICAL FIX**: Added explicit page copying after `cloneReaderDocumentRoot()`
 
-The main compatibility layer in `oca-ocb-base` handles all PyPDF2 version compatibility automatically.
+In PyPDF2 3.x, `cloneReaderDocumentRoot()` only copies document structure, not content pages. This was causing 327-byte empty PDFs. The fix includes explicit page copying:
+
+```python
+writer.cloneReaderDocumentRoot(reader)
+# Copy all pages from the reader to the writer (required for PyPDF2 3.x)
+for page_num in range(reader.getNumPages()):
+    page = reader.getPage(page_num)
+    writer.addPage(page)
+```
 
 ## Files Using PyPDF2
 
 ### `account_edi/models/ir_actions_report.py`
 - Uses `OdooPdfFileWriter` for PDF generation (automatically compatible)
-- Calls `writer.cloneReaderDocumentRoot(reader)` 
+- **FIXED**: Added explicit page copying after `writer.cloneReaderDocumentRoot(reader)`
 - Embeds EDI XML attachments in PDFs
 
 ## Implementation Details
 
-**No code changes needed** in this package. Compatibility is achieved through:
+**Direct code changes applied** in this package:
 
-1. **Dependency**: Requires `oca-ocb-base` with `pdfwrite` branch
+1. **Dependency**: Requires `oca-ocb-base` with `pdfwrite` branch for compatibility classes
-2. **Automatic compatibility**: `OdooPdfFileWriter` handles all PyPDF2 version differences
+2. **Page copying fix**: Added explicit page copying loop in `_render_qweb_pdf_prepare_streams()`
-3. **Method compatibility**: Methods like `cloneReaderDocumentRoot()` are automatically wrapped
+3. **Method compatibility**: Methods like `cloneReaderDocumentRoot()` are automatically wrapped by base classes
 
 ## Testing