Search:

To illustrate the difference:

Compile time allocation:

static char fred[128];

Run time equivalent:

char * fred;
fred = malloc(128);

Normally, I wouldn't respond to this type of rant, but for the last time:

You are not even remotely special. True - never thought any different.
We've seen your kind a hundred times. Err -...

And it is.

I made two stupid mistakes:

1) I got my globals wrong and defined them as static in main.c rather than in gui/guiglb.c with extern definitions in gui/guiglb.h. How I made that...

True, the compiler bug thought was just clutching at straws.

I did try to replicate the problem in a simpler form, but it worked as expected. Maybe it is too simple, so I'll try with a more...

Can anyone tell me what the complaints procedure is against this sort of abuse? Do the members here really want to be remembered for this type of behaviour?

Ah, abuse now. Look at my last post.

Really, if you can't say anything constructive, please STFU.

Nope, clang does the same. It must be me...

Oh, come on!

The code and results are exactly as I displayed them. Saying the results are impossible isn't helpful. They are what they are - I haven't made them up.

I'm beginning to wonder if...

No, I meant what I said - AFTER the assignment loop, in other words after the closing brace.

#define NFLAGS 8
The breakpoint is just after the assignment loop.

I'm not inexperienced in C, but this one has me baffled. It involves GTK+3.0.

The relevant bits of code are:


char FlagEnt[NFLAGS][8] = ...

static GtkWidget * flentry[NFLAGS];
static...