import std.stdio ;

void foo ( int[] array ) {
	foreach ( i ; array ) { writeln ( i ) ; }
}

void main () {
	foo ( [ 1, 2, 3 ] ) ;
}

On both DMD 2.062 and 2.063 this compiles OK but causes a segfault. I'm running Kubuntu Raring 64-bit. Any hints?

-- 
Shriramana Sharma ஶ்ரீரமணஶர்மா श्रीरमणशर्मा

Shriramana Sharma:

> On both DMD 2.062 and 2.063 this compiles OK but causes a segfault.
> I'm running Kubuntu Raring 64-bit. Any hints?

I have no hints. But on Windows32 I see no segfault. And the online 64 bit Linux (dpaste) gives no errors.

Bye,
bearophile

On Friday, 31 May 2013 at 17:14:46 UTC, Shriramana Sharma wrote:
> import std.stdio ;
>
> void foo ( int[] array ) {
> 	foreach ( i ; array ) { writeln ( i ) ; }
> }
>
> void main () {
> 	foo ( [ 1, 2, 3 ] ) ;
> }
>
> On both DMD 2.062 and 2.063 this compiles OK but causes a segfault.
> I'm running Kubuntu Raring 64-bit. Any hints?

Can't reproduce this. Any chance of a stacktrace and/or disassembly?

On Fri, May 31, 2013 at 11:46 PM, John Colvin <john.loughran.colvin@gmail.com> wrote:
>
> Can't reproduce this. Any chance of a stacktrace and/or disassembly?

Please give instructions as to how I can get either of those two for you. Would you like the binary produced?

-- 
Shriramana Sharma ஶ்ரீரமணஶர்மா श्रीरमणशर्मा

On Friday, 31 May 2013 at 18:22:44 UTC, Shriramana Sharma wrote:
> On Fri, May 31, 2013 at 11:46 PM, John Colvin
> <john.loughran.colvin@gmail.com> wrote:
>>
>> Can't reproduce this. Any chance of a stacktrace and/or disassembly?
>
> Please give instructions as to how I can get either of those two for
> you. Would you like the binary produced?

given a program named "test" in the current directory

gdb ./test
run               #will segfault
bt full

and copy the output.

On Sat, Jun 1, 2013 at 12:07 AM, John Colvin <john.loughran.colvin@gmail.com> wrote:
> given a program named "test" in the current directory
>
> gdb ./test
> run               #will segfault
> bt full
>
> and copy the output.

Did it (after installing libc6-dbg just in case) but it doesn't seem
to be very useful:

(gdb) bt full
#0  0x000000000043dd20 in _d_dso_registry ()
No symbol table info available.
#1  0x000000000043862f in ?? ()
No symbol table info available.
#2  0x0000000000000001 in ?? ()
No symbol table info available.
#3  0x0000000000463898 in
TypeInfo_PS3std5array17__T8AppenderTAxaZ8Appender4Data.__init() ()
No symbol table info available.
#4  0x0000000000460bc8 in ?? ()
No symbol table info available.
#5  0x0000000000460db8 in ?? ()
No symbol table info available.
#6  0x0000000000460610 in ?? ()
No symbol table info available.
#7  0x0000000000460bc8 in ?? ()
No symbol table info available.
#8  0x0000000000000040 in ?? ()
No symbol table info available.
#9  0x000000000045790d in __libc_csu_init ()
No symbol table info available.
#10 0x00007ffff760ee35 in __libc_start_main (main=0x43b0d0 <main>,
argc=1, ubp_av=0x7fffffffdf58, init=0x4578b0 <__libc_csu_init>,
    fini=<optimized out>, rtld_fini=<optimized out>,
stack_end=0x7fffffffdf48) at libc-start.c:219
        result = <optimized out>
        unwind_buf = {cancel_jmp_buf = {{jmp_buf = {0,
140737351945365, 140733193388033, 0, 1, 140737343639840, 0, 0},
mask_was_saved = 4392416}},
          priv = {pad = {0x7ffff7def725 <_dl_runtime_resolve+53>,
0x1c, 0x4578b0 <__libc_csu_init>, 0x7fffffffdf58}, data = {
              prev = 0x7ffff7def725 <_dl_runtime_resolve+53>, cleanup
= 0x1c, canceltype = 4552880}}}
        not_first_call = <optimized out>
#11 0x0000000000430609 in _start ()
No symbol table info available.

-- 
Shriramana Sharma ஶ்ரீரமணஶர்மா श्रीरमणशर्मा

On Friday, 31 May 2013 at 17:14:46 UTC, Shriramana Sharma wrote:
> import std.stdio ;
>
> void foo ( int[] array ) {
> 	foreach ( i ; array ) { writeln ( i ) ; }
> }
>
> void main () {
> 	foo ( [ 1, 2, 3 ] ) ;
> }
>
> On both DMD 2.062 and 2.063 this compiles OK but causes a segfault.
> I'm running Kubuntu Raring 64-bit. Any hints?

Works for me ubuntu 64.

Do you have both versions installed?
Would it compile if you used the wrong druntime / phobos libraries?

On Sat, Jun 1, 2013 at 6:02 AM, Anthony Goins <neontotem@gmail.com> wrote:
>
> Works for me ubuntu 64.
> Do you have both versions installed?

Of course not. The later install from the distributed DEB (from dlang.org) overwrites the earlier one by apt install.

> Would it compile if you used the wrong druntime / phobos libraries?

I don't have any D-related packages on my system other than the stock DMD AMD64 package. Even libphobos2-4.6-dev which is available for Raring is not installed.

-- 
Shriramana Sharma ஶ்ரீரமணஶர்மா श्रीरमणशर्मा

On Sat, Jun 1, 2013 at 6:39 AM, Shriramana Sharma <samjnaa@gmail.com> wrote:
> I don't have any D-related packages on my system other than the stock DMD AMD64 package. Even libphobos2-4.6-dev which is available for Raring is not installed.

That was actually slightly incorrect. I had installed gdc apart from dmd -- just to see how the two work out. It's only after that I installed gdc (I had only used dmd) that D programs segfaulted. Now I went into /var/log/ and extracted yesterday's installation details from dpkg.log and purged all the packages that gdc pulled in -- mostly it was old gcc-4.6 and related files.

Nevertheless, the horrible fact is that not only this program I initially posted on this thread but any and all D programs I compile produce a segfault!

Please help!

-- 
Shriramana Sharma ஶ்ரீரமணஶர்மா श्रीरमणशर्मा

On Sat, Jun 1, 2013 at 1:27 PM, Shriramana Sharma <samjnaa@gmail.com> wrote:
>
> Nevertheless, the horrible fact is that not only this program I initially posted on this thread but any and all D programs I compile produce a segfault!

All programs compiled by *DMD* produce a segfault. Programs compiled by GDC work just fine. What is this, some sort of incompatibility between compilers? Like GDC went and told my system "hey look only D programs compiled by me should work"?!

I couldn't try LDC since they don't have Ubuntu packages and I couldn't figure out how to install the binary tarball.

Please help!

-- 
Shriramana Sharma ஶ்ரீரமணஶர்மா श्रीरमणशर्मा