Privacy and offline use
CAN databases are often confidential. CanViewer is built so that opening one does not share it with anyone.
Your files never leave your machine
- A file you open is read by the app in your browser (or in the desktop app). It is never uploaded: the app has no server to upload to.
- DBC files are parsed by JavaScript in the page. Other formats are read by canmatrix running in a Web Worker in the same browser, from a runtime the app ships itself.
- The last file you opened is kept in your browser's storage (a DBC file in local storage, an imported file in IndexedDB), so that it comes back after a reload. It stays on your machine. Close file in the settings menu forgets it, in the web app and in the desktop app; so does clearing the site's data in your browser, or deleting the desktop app's data folder (where).
- A log you replay in the Traces view, the other version on the Compare page, and the files you export or save as DBC are made in the page too. A log and the other version are not stored at all.
- Links to a view (
#msg=…) name a message, not its content, and the part after#is never sent to the web server.
No request outside the app's origin
The app loads everything from the place it was served from: its code, its fonts and icons, and the Python runtime with its packages, each package checked against a pinned SHA-256 checksum. It contacts no other host: no CDN, no font service, no analytics, no error reporting, no update check.
This is tested, not promised. Every change to CanViewer runs a check that drives the built app in Chromium, opens DBC and ARXML files online and offline, and records every request three ways (the browser's request events, a HAR file and Chromium's own network log). It fails on any request to another origin. The check runs against the app as it is published here too, under /app/, and the same kind of check crawls every page of this documentation site.
What the web host sees
This site and the web app are published with GitHub Pages. Like any web host, GitHub serves the app's files and so sees the requests for them: your IP address, your browser and the time, for the app's pages, scripts and styles, and for the Python runtime when you first open a non-DBC file. It does not see your files, their names, or anything you do in the app; none of it is sent. GitHub's own privacy statement covers what it logs.
The app is served from https://joshualeichsering.github.io/CanViewer/app/. Browsers keep storage per origin (scheme, host and port), not per path, so the copy of the last file the app keeps (local storage and IndexedDB, above) is in the storage of https://joshualeichsering.github.io, which every GitHub Pages site of that account shares: a page of any other site the account publishes there could read it. To leave nothing behind, use Close file when you are done, or use the desktop app, whose storage is its own.
If you would rather not load the app from GitHub at all, use the desktop app, or build the web app yourself and serve it from your own machine (see the developer guide).
Offline use
- Web app: after the first visit, a service worker keeps the app's files in your browser, so the app opens and reads DBC files without a network, and About CanViewer still shows the license texts and the third-party notices. The Python runtime is cached the first time you open a non-DBC file; after that, imports work offline too. A browser that has never imported a file cannot do so offline: the import then fails with a message, and the app keeps working for DBC files.
- Desktop app: everything is inside the program. It works offline from the start and makes no network requests at all. Its content security policy allows the window to load only the app's own files, and the worker that reads non-DBC formats runs under the same policy, so it cannot reach another host either. The web app has no such policy: GitHub Pages cannot send the header, and a policy in the page itself would not reach its import worker (why). There, the checks above stand in for it.