If we get EOPNOTSUPP from open() failing, then it may just be that we're
looking at a unix(4) socket instead of a normal file/device or whatnot.
Fallback to trying to open it as a unix(4) socket, which is a useful
feature that doesn't add much complexity.
Details
Details
Diff Detail
Diff Detail
- Repository
- rG FreeBSD src repository
- Lint
Lint Skipped - Unit
Tests Skipped - Build Status
Buildable 61326 Build 58210: arc lint + arc unit
Event Timeline
Comment Actions
I'd like to look at incorporating something like https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=117733, too, though the patch isn't available anymore.
Comment Actions
Should we mention this in the man page? Not sure it's necessary, but if this is something that other tee implementations don't do, or something that they document, we ought to.
Comment Actions
Add a note to the manpage that we'll also accept UNIX sockets. No other
implementation that I looked at seems to do this, but it would be handy.
usr.bin/tee/tests/tee_test.sh | ||
---|---|---|
85 | Since the value of $unixsock is hard-coded here, maybe just get rid of the variable entirely? |
usr.bin/tee/tests/tee_test.sh | ||
---|---|---|
70 | No blank line before variables |