https://gcc.gnu.org/bugzilla/show_bug.cgi?id=116885
--- Comment #21 from Jaroslav Fojtík ---
Anyway, all unit tests are passing now, including this one named "struct", on
all Debian duilders:
https://buildd.debian.org/status/package.php?p=wp2latex&suite=sid
I did not note any problem before from
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=116885
--- Comment #20 from Jonathan Wakely ---
(In reply to Jaroslav Fojtík from comment #0)
> jara@cfarm421:~/mintest$ make
> gcc -I./ -Wall -g -O2 -c struct.c -o struct.o
> gcc -I./ -Wall -g -O2 -c mintest.c -o mintest.o
> gcc -lm -o mintest str
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=116885
--- Comment #19 from Andrew Pinski ---
(In reply to Jaroslav Fojtík from comment #18)
> May be that this report points to a similar bug:
> https://gcc.gnu.org/bugzilla/show_bug.cgi?id=116621
>
> I cannot consider this to be a bug because of t
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=116885
--- Comment #18 from Jaroslav Fojtík ---
May be that this report points to a similar bug:
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=116621
I cannot consider this to be a bug because of types mismatch i my case.
"Normaly" the va_arg is imp
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=116885
--- Comment #17 from Jaroslav Fojtík ---
The only reasonable source of this crash was that doubles are passed different
way than uint64_t. May be that doubles are newly passed through different
stack/pipe. Only looking to assembly might reveal d
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=116885
Jeffrey Walton changed:
What|Removed |Added
CC||noloader at gmail dot com
--- Comment
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=116885
--- Comment #15 from Jaroslav Fojtík ---
When I have finished university where I has been learned C/C++ in 2000 there
was no problem to typecast uint64_t* to double* or any other type I need. We
were learned that this is correct ;).
Now I see t
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=116885
Sam James changed:
What|Removed |Added
CC||sjames at gcc dot gnu.org
--- Comment #14 f
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=116885
--- Comment #13 from Jaroslav Fojtík ---
Thank you for help.
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=116885
--- Comment #12 from Jaroslav Fojtík ---
this solution should be correct according to documentation:
typedef union
{
double d;
uint64_t u64;
} DblUint64;
case 'e':{DblUint64 du;
du.d = va_arg(ap,double);
wri
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=116885
--- Comment #11 from Vincent Lefèvre ---
(In reply to Jaroslav Fojtík from comment #9)
> Accessing to an incompatible type throug pointer is not clean in general.
> But I need here to do a dirty job to flip endianity of double type. There
> are
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=116885
--- Comment #10 from Jaroslav Fojtík ---
I can call 8x fwrite with 1 byte to camouflage this. It will produce slower and
longer code and type punning will be only hidden. I cannot avoid this shuffle.
// Needed to store double in LoE
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=116885
--- Comment #9 from Jaroslav Fojtík ---
https://wiki.sei.cmu.edu/confluence/display/c/EXP39-C.+Do+not+access+a+variable+through+a+pointer+of+an+incompatible+type
Accessing to an incompatible type throug pointer is not clean in general. But I
ne
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=116885
Vincent Lefèvre changed:
What|Removed |Added
CC||vincent-gcc at vinc17 dot net
--- Com
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=116885
--- Comment #7 from Jaroslav Fojtík ---
Just for clarification.
This worked till GCC-14 - The code has been verified and proven functionable
many times:
case 'e':{const uint64_t e = va_arg(ap,uint64_t);
writed+=WrQWORD_LoEnd(
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=116885
Jaroslav Fojtík changed:
What|Removed |Added
Status|WAITING |RESOLVED
Resolution|---
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=116885
Thomas Koenig changed:
What|Removed |Added
Keywords|ABI |
--- Comment #5 from Thomas Koenig ---
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=116885
--- Comment #4 from Thomas Koenig ---
(In reply to Jaroslav Fojtík from comment #3)
> It seems to me that this problem has something to do with memcpy optimised
> for AVX, that newly requests everything to be alligned.
As shown by the valgrind
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=116885
--- Comment #3 from Jaroslav Fojtík ---
It seems to me that this problem has something to do with memcpy optimised for
AVX, that newly requests everything to be alligned.
--
<<>> (c)1998-2024
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=116885
--- Comment #2 from Jaroslav Fojtík ---
May be that I did something wrong, but this code worked for 20 years without
any problem. It has been also tested unded Borland and VisualStudio without any
notable problem.
I have noted problem only for
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=116885
Thomas Koenig changed:
What|Removed |Added
Status|UNCONFIRMED |WAITING
CC|
21 matches
Mail list logo
