Signed-off-by: Cody P Schafer <dev@codyps.com>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>

This might allow compilers that support the anotation to make better
choices when optimizing, and all these functions meet the requirements
for being marked pure.
Signed-off-by: Cody P Schafer <dev@codyps.com>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>

Reviwed-by: David Gibson <david@gibson.dropbear.id.au>
Signed-off-by: Cody P Schafer <dev@codyps.com>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>

Signed-off-by: Cody P Schafer <dev@codyps.com>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>

It now warns about sizeof(function-param-not-really-an-array).
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>

A simple printf logging infra where levels are determined by the
value of the "DEBUG" environment variable.

This is loosely based on the interfaces & functionality of Linux's
printk() and pr_*() wrapper macros.

Note that the current implementation uses "<N>" prefixes (where N is a
syslog level in ascii), allowing other programs that parse log output
(like systemd's journald) to know what the priority level is.
Signed-off-by: Cody P Schafer <dev@codyps.com>
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>

Signed-off-by: Cody P Schafer <dev@codyps.com>
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>

Signed-off-by: Cody P Schafer <dev@codyps.com>
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>

Config defines are disabled if a warning is emitted (we may want to
reconsider that), and warnings are emitted for define redefinition.
Signed-off-by: Cody P Schafer <dev@codyps.com>
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>

As for the type punning: gcc-5.1 with optimization (at least) warns about type punning in
the previous example. The new usage should be exactly equivalent to the
old, but just seperates the cast and deref into 2 statements. Frankly,
I'm suprised gcc's type-punning analysis is so limited.
Signed-off-by: Cody P Schafer <dev@codyps.com>
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>

Reviewing the previous patch it took me some time to work out what the
purpose of the compile_fail-BYTESTRING-2.c test.  Add a comment to avoid
that in future.
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>

bytestring: Module tests compile (tests_compile): FAIL
/home/x/g/ccc/ccan/ccan/bytestring/test/compile_fail-BYTESTRING-2.c:Compile gave warnings without -DFAIL:
/home/x/g/ccc/ccan/ccan/bytestring/test/compile_fail-BYTESTRING-2.c: In function ‘main’:
/home/x/g/ccc/ccan/ccan/bytestring/test/compile_fail-BYTESTRING-2.c:15:2: warning: ‘bs.len’ is used uninitialized in this function [-Wuninitialized]
  printf("%zd %s\n", bs.len, x);
  ^
Signed-off-by: Cody P Schafer <dev@codyps.com>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>

Without this, gcc warns about a sign mismatch in the comparison.
Signed-off-by: Cody P Schafer <dev@codyps.com>
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>

These leaks aren't really an issue since they are completely bounded,
but if one is building with leak sanitizer enabled (as
-fsanitize=address does in gcc-5.1), it kills the configurator, which
isn't very useful for us. Add the few free() calls it's looking for.

This is not an actual code issue, they just workaround
some optional compiler peculiarities.
Signed-off-by: Cody P Schafer <dev@codyps.com>
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au> (split off leak change)

Signed-off-by: A. Samy <f.fallen45@gmail.com>

Signed-off-by: A. Samy <f.fallen45@gmail.com>

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>

I really need to get rid of this...
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>

New module
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>

This adds a more complex testcase to the aga module.  This one is a trie
(basically a radix tree for strings).

It demonstrates different ways of constructing edge information from an
internal representation than the existing testcases.  Importantly, it also
demonstrates aga's ability to cope with the edge function lazily
constructing nodes on the fly.
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>

The aga algorithms can't be run concurrently, because they store state
information in the aga_node structures.  However, they are supposed to
detect an attempt to re-enter and safely report an error.  This adds a
testcase for this.
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>

This implements breadth first search for the abstract graph algorithms
module.
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>

This implements depth first search for the abstract graph algorithms
module.
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>

This adds code for a number of example graphs for use in tests of the aga
module.  They also demonstrate several different ways of constructing
graphs using the aga callbacks.

It adds one actual testcase, which just verifies that the example graph
look like what they're supposed to.  Specifically it computes a set of
adjacency lists for the example graphs from the callback code, then
compares that to a canned example of what it should be.
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>

New module.

This patch just adds the module, with some generic helper routines, no
actual algorithms are implemented yet.
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>

Our project builds the ccan tests with -Wextra, so we get warnings about
the unused variables.
Signed-off-by: Joel Stanley <joel@jms.id.au>
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>

The cp -a would copy nested submodules, and worse, if that submodule
was copied later it would end up duplicated in a subdir,
eg. "ccan/tal/str/str/".

Using modfiles handles this properly, and also uses git to filter
out ignored files & other junk.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>

It's generally a symlink, so the lstat & S_ISREG tests were eliminating it.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>

They currently get classed as "other_files"; they should be entirely
ignored.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>