Apex good first issues

A bug in the Salesforce document generation service causes rich-text or long-text field values containing bare HTML tables or headings to be incorrectly escaped as plain text. This results in raw markup appearing in generated documents instead of rendered HTML, leading to inconsistent rendering depending on how the content was created.

Hyperlinks inserted directly in Microsoft Word are not clickable in the generated PDF. While they appear styled as links, the URL is dropped during the DOCX to HTML conversion process because the system only recognizes custom `w:docgen-url` attributes for creating clickable links, not standard Word hyperlink references.

The user is asking if the login component supports redirecting to a specific URL after successful login, using a URL parameter like `startURL`. They are looking for a way to achieve this functionality, similar to how other standard login components might behave.

When a Word document contains multiple floating text boxes within a single paragraph, such as in headers or footers, only the first text box is rendered. Subsequent text boxes and their content, including merge tags, are silently dropped. This issue is particularly problematic for documents authored in LibreOffice, where header/footer elements are often structured as multiple text boxes.

HTML templates with embedded `data:` URI images render blank because the PDF generation engine (Flying Saucer) cannot decode them. A workaround exists for the Designer upload flow, but it's not applied to templates imported via bundles, AI generation, or other methods, leading to broken PDF output without errors.

The `extractLoopBody` method in `DocGenService` uses a whole-document regex matcher that hits a 'Regex too complicated' limit around 1MB for large documents. This is a recurring issue with similar regex patterns in the system, and while it's a different path than previous fixes, it still impacts the giant-query path.

This issue proposes enabling custom font embedding in generated PDFs using the `pdf-lib` library. Currently, PDFs are limited to a few basic fonts, preventing the use of brand fonts and causing issues with bold text and special characters. The solution involves adding the `@pdf-lib/fontkit` plugin to `pdf-lib`, which is already partially integrated.

This issue reports that Arabic text renders correctly in DOCX output but not in PDF generated by Portwood. The problem is likely due to PDF generation using Flying Saucer, which has limited font support, and may require explicitly setting the font to 'Arial Unicode MS' for Arabic characters. Even with the correct font, bidirectional reordering and contextual shaping for Arabic might not be supported by the PDF renderer.

This issue highlights a critical problem where several browser platform APIs are being used in the LWC layer without proper feature detection or fallbacks. This has led to multiple customer-reported failures across different Salesforce sandbox environments, as these environments block APIs inconsistently. The issue calls for a comprehensive audit and fix for all such instances to prevent future regressions.

Managed package upgrades are failing to automatically add new picklist values, such as 'PDF' and 'Canvas', to existing fields. This requires administrators to manually add these values, preventing users from accessing new features. The issue proposes implementing an in-app repair mechanism using Apex's native metadata deployment capabilities, which the package already utilizes for other operations.

This issue requests a way for Flows to react to document generation completion. Currently, there's no direct mechanism to know when a document has been generated by the `docGenRunner` on a Flow screen, leading to potential data drift. The proposed solution involves adding outputs to `docGenRunner` and dispatching Flow events to signal generation success.

A user reported that when using a PowerPoint template and setting the output format to PDF, the system silently generates a .pptx file instead of an error. The underlying code correctly identifies this limitation but handles it inconsistently, leading to a poor user experience where the limitation is not discoverable until after generation.

When exporting a document template from one Salesforce organization to another, the page size setting (e.g., A4) is not preserved during the import process. Instead, the imported template defaults to the Letter page size, causing layout issues and incorrect document generation. This is likely due to the page size field not being included in the export payload.

This issue highlights an ambiguity in currency formatting where multiple Nordic currencies (SEK, NOK, DKK) and East Asian currencies (JPY, CNY) render with the same symbol ('kr' and '¥' respectively). The proposed solution is to add a modifier to the currency formatting mechanism to allow users to explicitly render the ISO code (e.g., SEK instead of kr) to avoid confusion, especially in multi-currency environments.

The Canvas designer's "Import HTML" functionality incorrectly uses a general HTML converter instead of a specific canvas document deserializer. This causes exported Canvas documents to be flattened into a single element upon re-import, breaking the intended round-trip functionality for templates.