On 20/03/2014 19:14, Tim Wescott wrote:
On Thu, 20 Mar 2014 11:51:34 -0700, Jim Thompson wrote:

Anyone know how to interpret this IF (I think) statement...

PMI=(PM=0)?((PM==0)?1000:PM):PMC

it's in a notation-style I don't understand.

It looks like C-style conditional operator (whose use is frowned upon,
particularly in cascade like that).

In C and C++,

ANSWER = THING ? THIS : THAT;

means that if THING is true, ANSWER = THIS. Of THING is false, ANSWER =
THAT.

In more traditional if-else notation, your statement expands to:

if (PM = 0)
{
if (PM == 0)
{
PMI = 1000;
}
else
{
PMI = PM;
}
}
else
{
PMI = PMC;
}

Although to some extent it would be more readable for a non-programmer
by rewriting it in order of ascending value of PM namely

if (PM0) PMI = PMC
else
{
if (PM==0) PMI = 1000; // in heaven's name why????
else PMI = PM
}

either way it has CCI=3 DP=2
PMI = -1, 0, +1 to test all execution paths

Ironically most machine level compares can do this branch network as a
single comparison and then branching on the sign bit and zero bit.

Which packs much less confusion per line, and is therefor much less
desirable from a geeky programmer point of view.

If there was a serious intention to obfuscate then the variables would
all be called O,I,l,O1,Ol,IO,I0,lO,l0,II,I1,Il, etc.

Those limp-wristed busybodies at MISRA have ruled out its use in mission-
critical software, if you can imagine that.

Unfortunately they can't rule out the possibility of a typo and writing

if (PM=0) doit // which never works

when the intention is

if (PM==0) doit

which is why it is favoured to put the constant as the lhs.

if (0=PM) is then a hard compiler error.

Most modern compilers will catch if (a=b) as potentially dangerous code.

--
Regards,
Martin Brown