Decode ISO-8859-1 filename* in Content-Disposition - #78

When a server sends an RFC 5987 filename* in ISO-8859-1, the downloaded file is named after the raw percent-escapes.

Content-Disposition: attachment; filename*=iso-8859-1'en'%A3%20rates.txt

Before: the file is saved as %A3%20rates.txt. After: it is saved as £ rates.txt.

parseContentDispositionFilename passes every filename* value to decodeURIComponent. That function only reads UTF-8, so a Latin-1 byte such as %A3 throws a URIError, and the catch returns the undecoded value. RFC 5987 requires recipients to support both UTF-8 and ISO-8859-1, and the header above is the RFC's own example. Servers that still default to Latin-1 send it this way.

Change

• src/main/download/probe.ts: the regex now captures the charset. For iso-8859-1, each %XX is decoded as one Latin-1 character. UTF-8 and other charsets go through the same path as before.
• e2e/download.spec.ts: new case in the file names @smoke table.

How I checked

• The new e2e case fails on main (fileName is %A3%20rates.txt) and passes with the fix.
• npm run lint, npm run typecheck, and npm run format:check pass.
• npm run test:e2e:smoke: 90 passed, on Linux under xvfb.