{{ message }}
Repository navigation
v4: FileDownload silently started verifying TLS certificates #4240
Description
v3 FileDownload extended HttpCrawler and inherited ignoreSslErrors: true. v4 FileDownload extends BasicCrawler (packages/http-crawler/src/internals/file-download.ts), has no ignoreTlsErrors option and sends with { session } only, so downloads from hosts with invalid certificates now fail where they used to succeed.
The upgrade guide and skill now document the workaround (httpClient: new ImpitHttpClient({ ignoreTlsErrors: true })). Decide before GA whether to:
- keep the new strict default and the documented workaround, or
- add an
ignoreTlsErrorsoption toFileDownloaddefaulting totrue, matchingHttpCrawler, and forward it tosendRequest.
[🤖] Claude Code using Fable 5.1
Activity
barjin commentedon Oct 13, 2026
I believe the original was some sort of a WCC hack. We shouldn't imo disable the TLS verification in general.
The
ignoreTlsErrorsoption is now glued to theproxyInfostruct (so, e.g., Apify SDK can disable the TLS verification for MITM proxies so they can work properly). We should make sure we respect thatproxyInfo.ignoreTlsErrorsoption (and explicit user setting), but we imo should be verifying certificates by default.feat(http): add an ignoreTlsErrors option to FileDownload
feat(http): add an ignoreTlsErrors option to FileDownload