weird iOS 5.0 SDK GCC side effect on 4.2.x devices

Question

I'm building a library against the 5.0 sdk GCC and running the code on a 4.2.x device.

I'm not using 5.0 objective-c specific calls in this layer and the project is compatible to ship on 4.0+.

I'm seeing some behavior in my library that is really odd with my if blocks.

typically this should work

BYTE    byteVal : 1;

byteVal = FALSE;

if (byteVal)
    // ALWAYS RUNS

The problem being that the code in the if block is always executing.

This is causing me problems with zlib gzip functionality. Is the 4.2.x OS using some offset or different register alignment that isn't standard with building with the newer GCC?

I'm at a loss as to what is going on here and why this fails always on 4.2.x devices.

Any thoughts?

Genuine question (not implying it's incorrect or anything - not an expert) - Is that the whole code (I mean, isn't the `BYTE byteVal : 1` in an `extern {...}` kind of block?)? — Romain, Jan 20 '12 at 16:41
I'm also having problems with read() in zlib gzip where I cannot read files on disk. It always returns 0 bytes read. — cynistersix, Jan 20 '12 at 18:17
Could you post the actual struct definition and the relevant real code you use it in? Did you turn on all warnings? — Mat, Jan 22 '12 at 12:46
Sorry I cannot post the structure. What did fix it was to use byteVal == 1 instead of just the if (byteVal) — cynistersix, Feb 22 '12 at 19:44

score 0 · Answer 1 · answered Feb 22 '12 at 19:45

0

use

if (byteVal == 1)

There is some problem with using single bit wide member variables that the if (byteVal) is always true even when it is not.

answered Feb 22 '12 at 19:45

cynistersix

1,215
1
16
30

weird iOS 5.0 SDK GCC side effect on 4.2.x devices

1 Answers1