olaf/Link-Log
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases 4 Wiki Activity

New Entry page improvements on link detection

Browse Source
This commit is contained in:
olaf
2026-08-28 09:04:24 +02:00
parent cf83c32b25
commit 50d61371e1
8 changed files with 159 additions and 37 deletions
Download Patch File Download Diff File Expand all files Collapse all files
VIBE/CHAT_LOG.md
+36
View File
@@ -1424,3 +1424,39 @@ Don't let the frontend and plugin pull from the google foundry but make sure any
### Assistant outcome
Removed the web frontend's Google Fonts import and added a Docker build-time font bundling script. Image builds now download pinned Asset, DM Sans, and Space Grotesk files into `frontend/static/fonts`, generate a local `fonts.css`, and serve them through LinkLog at `/static/fonts`. The Firefox extension did not make Google Font requests and continues to use bundled assets with local fallback fonts. A clean image build confirmed all nine local font assets are present.
### User
On the new-entry page follow the logic for existing pages that already exists in the plugin and display "This link already exists. Comment and tags can be updated, and by submitting publishing will be retriggered." when the link exists. Also assess if the Auto-fill title button is necessary, or if the functionality can be automated while the URL is entered.
### Assistant outcome
Added a duplicate-check request (`GET /api/links/check`) to `new-entry.js`, triggered on URL/title blur, that shows the requested warning in a new `#duplicate-status` element and clears it as the fields are edited. Replaced the manual "Auto-fill Title" button with automatic scraping on URL blur (only filling the title when it is still empty and the URL actually changed), removing the now-unused button and CSS.
### User
Also: detect as duplicate when the URL has arguments (not already stripped). And: add a button to re-fetch the title for if the URL has been fetched but the user still edits the field.
### Assistant outcome
Added client-side stripping of known tracking parameters (`utm_*`, `gclid`, `fbclid`, etc.) mirroring the browser extension, applied consistently to the scrape lookup, duplicate check, and submission. Added a "Re-fetch Title" button next to the title field that force-refreshes the scraped title regardless of the auto-fill guard.
### User
If the link already exist with URL arguments then also display the warning, but add that the link has been stored additional parameters.
### Assistant outcome
Extended the duplicate warning to note when the stored link had additional URL parameters stripped (e.g. tracking parameters) compared to the entered URL.
### User
I want this last change to be more precise: I do not want to be warned when parameters are stripped, but when the URL entered has arguments/parameters and the URL in the database has none or different ones (even after tracking parameters are stripped).
### Assistant outcome
Added `find_owned_link_by_title` in `link_service.py` and extended `GET /api/links/check` to return `{exists, url_matches, stored_url}`, detecting a title match with a differing URL in addition to an exact match. `new-entry.js` now only warns about a differing stored URL when `url_matches` is false, no longer warning solely because tracking parameters were stripped from an otherwise matching URL.
### User
In the check duplicate I have a `<br>` tag, but that shows as formatted text on the page, I want a real break to occur at that position.
### Assistant outcome
Reworked `setStatus` in `new-entry.js` to split messages on `\n` and build real `<br>` elements via DOM APIs instead of relying on `textContent` (which escaped the literal `<br>`) or unsafe `innerHTML`.
### User
Update VIBE and Changelog.
### Assistant outcome
Appended the new-entry duplicate-detection and title auto-fill work to `VIBE/PROMPTS.md` and `VIBE/CHAT_LOG.md`, and added a Version v0.1.3 section to `CHANGELOG.MD`.