| commit | e5435409ac46c068b007c42664040b453e9f8a40 | |
| author | Steffen Nurpmeso <steffen@sdaoden.eu> | |
| Sat, 10 Aug 2019 22:52:29 +0000 (11 00:52 +0200) | ||
| committer | Steffen Nurpmeso <steffen@sdaoden.eu> | |
| Sat, 17 Aug 2019 22:12:43 +0000 (18 00:12 +0200) | ||
| tree | 64be0d5a735087b3f7122ab571116c4e3df34e7a | treesnapshot (tar.gz zip) |
| parent | 75122526df5ec98f53538a72f6cfbea21b596f83 | commitdiff |
mx-test.sh: no, use my initial approach plus Robert Elz..
The FIFO approach does not work out. Robert Elz moved once again
and hinted keeping an fd open (exec 4<t.fifo), but that still is
much too complicated.
Use my initial ID-file based approach with some of the hints of
Robert Elz (for example signal communication in between parent and
timeout process). It is still a little bit racy, but worked on
all platforms without problems.
The FIFO approach does not work out. Robert Elz moved once again
and hinted keeping an fd open (exec 4<t.fifo), but that still is
much too complicated.
Use my initial ID-file based approach with some of the hints of
Robert Elz (for example signal communication in between parent and
timeout process). It is still a little bit racy, but worked on
all platforms without problems.
| mx-test.sh | diffblobblamehistory |