- 15 Feb, 2011 18 commits
-
-
Rusty Russell authored
If we need to clean up the children, they didn't exit cleanly. This takes a bit more care when writing tests, but found a leak in tdb2.
-
Rusty Russell authored
tdb2 has various places where it recovers from failure (eg. falling back when it can't open /dev/urandom, or allocation for error logging). We want to test those paths, but doing so thoroughly causes cominatorial explosion. Add FAIL_PROBE for such cases: in this case it goes only 3 more calls deep.
-
Rusty Russell authored
We trap them, might as well put them in history. This also makes tracking open file descriptors more robust.
-
Rusty Russell authored
-
Rusty Russell authored
ie. SIGUSR1 means it timed out.
-
Rusty Russell authored
If children and parents aren't isolated properly (ie. failtest is buggy) we really want to run gdb on a failing child. This hack allows that.
-
Rusty Russell authored
open fails in fairly normal cases; don't double execution time for this!
-
Rusty Russell authored
We insist they write the same things to pipes, since we can't "undo" them, but strictly speaking we don't care if they write different things into files. Note: it may indicate a bug if they do...
-
Rusty Russell authored
-
Rusty Russell authored
The idea of saving files as we do writes doesn't work with mmap: this just saves the entire contents of all open files before forking child, and restores them afterwards.
-
Rusty Russell authored
Use the same structure we use in the history.
-
Rusty Russell authored
Each function in the history stores a cleanup function, rather than storing extra structures. In particular, we save writes and file offsets using this.
-
Rusty Russell authored
Append a "+" if you want to start forking as normal.
-
Rusty Russell authored
Useful for showing exactly where we inserted failures, and how long each one took.
-
Rusty Russell authored
Correctly mark which calls we failed.
-
Rusty Russell authored
-
Rusty Russell authored
-
Rusty Russell authored
Catch fcntl, particularly fcntl locks.
-
- 17 Feb, 2011 2 commits
-
-
Rusty Russell authored
-
Rusty Russell authored
-
- 22 Feb, 2011 7 commits
-
-
Rusty Russell authored
-
Sam Vilain authored
If 'localhost' exists in /etc/hosts with a v4 and a v6 address, the hack which joins two addrinfo entries together is not necessary. Detect to see if 'localhost' returned a v6 address, and if so, just return the single linked list result from getaddrinfo() (Reworked by Rusty Russell) Signed-off-by:
Sam Vilain <sam@vilain.net>
-
Andreas Schlick authored
-
Andreas Schlick authored
-
Rusty Russell authored
-
Rusty Russell authored
-
-
- 18 Feb, 2011 1 commit
-
-
Rusty Russell authored
-
- 11 Feb, 2011 10 commits
-
-
Stewart Smith authored
move daemon-with-notify to daemon_with_notify as dashes aren't allowed in CCAN module names (and ccanlint segfaults on them)
-
Stewart Smith authored
-
Stewart Smith authored
-
Stewart Smith authored
-
Stewart Smith authored
-
Stewart Smith authored
-
Stewart Smith authored
-
Stewart Smith authored
-
Stewart Smith authored
-
Stewart Smith authored
-
- 10 Feb, 2011 1 commit
-
-
Stewart Smith authored
-
- 07 Feb, 2011 1 commit
-
-
Joel Stanley authored
The type checking function does not use its parameter: ccan/opt/opt.h: In function ‘_check_is_entry’: ccan/opt/opt.h:328:53: warning: unused parameter ‘e’ Annotate the function using UNUSED from compiler.h. This commit also adds compiler.h as a dependency for opt.
-