Size of try/catch blocks

[souldog]

If I have a very large function body which contains several
calls to functions which throw similar exceptions is it less
costly to have several try/catch blocks or one try/catch block
for the entire function body. Lets assume that the catch blocks
all perform the same function.

[Mick]

well if they perform the same function, then why not catch(...) unless you are looking at the exception....so I know soul this isn't what you are looking for, so explain a bit more....

[souldog]

what is mean is say F1, F2, ...., FN throw exceptions
and I have

void FOO()
{

          F1();
...

          F2();
...
...
...
          FN();
} 

As I understand it (which is not very well) I am going to pay
a price for having to keep track of code within try/catch for
stack unwinding. Now is this price worse if I use

void FOO()
{
    try
    {
          F1();
...

          F2();
...
...
...
          FN();
     }
     catch
     {
     }

}

as oppossed to 

void FOO()
{
    try
    {
          F1();
    }
    catch
     {
     }
...

    try
    {
           F2();
    }
    catch
     {
     }
... 
...
...
    try
    {
           FN();
     }
     catch
     {
     }

} 

Now maybe this is a dumb question and the difference is insignificant.

[Mick]

Ok...

just one try and multiple catchs...it matters not what is thrown, that a catch is there...you could disassembly it and look, but I would gather that the optimizer for the compiler would be the same....and no I haven't looked dumpbin /disasm .exe/.dll if your interested...

[souldog]

Got it Mick, thanks...

[mdmd]

Bobby Schmidt's 17 part exception series in the MSDN talks about
some overhead in section 4: even looks at the generated asm.
Maybe some details for you ? Probably different in VC7.

Handling Exceptions in C and C++, Part 4

[Mick]

Originally posted by mdmd
Bobby Schmidt's 17 part exception series in the MSDN talks about
some overhead in section 4: even looks at the generated asm.
Maybe some details for you ? Probably different in VC7.

Handling Exceptions in C and C++, Part 4

You should have stated at part 1....these go to 17,,,,

[mdmd]

Originally posted by Mick
You should have stated at part 1....these go to 17,,,,

I started from 17 and went backwards thinking he wouldn't be
"spelunking a tiny way into assembly language" untill the
end; I never made it to part 1.
Its good enough to read through again fully. Next time I'll start
from Handling Exceptions in C and C++, Part 1

[Mick]

Originally posted by mdmd
I started from 17 and went backwards thinking he wouldn't be
"spelunking a tiny way into assembly language" untill the
end; I never made it to part 1.
Its good enough to read through again fully. Next time I'll start
from Handling Exceptions in C and C++, Part 1

I only glanced through...seemed like a good read, and after all it's 17 parts..what can be wrong with that??? unless yer a liberal then it's really wrong

[souldog]

Thanks MDMD that is what I needed...information.