• strtr <strtr spam.com> Feb 13 2010
• daoryn <asd asdf.wqa> Feb 13 2010
• strtr <strtr spam.com> Feb 13 2010
• Don <nospam nospam.com> Feb 14 2010

strtr <strtr spam.com> writes:

or: Why does dmd keep on crashing on my code :)
(I blame it on the average D programming; they don't write enough crappy code ) 

After reading the -w post I tried using the -w switch again and it also
prevents me from seeing all my warning, but for a different reason : it crashes
dmd :(

Before crashing a few correct warnings are shown :
warning - tests.d(232): Error: implicit conversion of expression (i1) of type
uint to ubyte can cause loss of data

Somehow I used size_t like this:
struct S {
  ubyte[4] ub;
  void opIndex(size_t i1, size_t i2, size_t i3, size_t i4) {
    ub[0] = i1;
...
}
I have no clue as to why I put size_t there but replacing them with ubyte seems
to fix the crashing problem. 

I can't replicate the crash in something small using only the struct or
something  :(
Somehow it only works when at least twenty modules are imported in (indirectly)
and a lot of them have circular dependencies.

daoryn <asd asdf.wqa> writes:

strtr Wrote:

 or: Why does dmd keep on crashing on my code :)
 (I blame it on the average D programming; they don't write enough crappy code
) 
 
 After reading the -w post I tried using the -w switch again and it also
prevents me from seeing all my warning, but for a different reason : it crashes
dmd :(
 
 Before crashing a few correct warnings are shown :
 warning - tests.d(232): Error: implicit conversion of expression (i1) of type
uint to ubyte can cause loss of data
 
 Somehow I used size_t like this:
 struct S {
   ubyte[4] ub;
   void opIndex(size_t i1, size_t i2, size_t i3, size_t i4) {
     ub[0] = i1;
 ...
 }
 I have no clue as to why I put size_t there but replacing them with ubyte
seems to fix the crashing problem. 
 
 I can't replicate the crash in something small using only the struct or
something  :(
 Somehow it only works when at least twenty modules are imported in
(indirectly) and a lot of them have circular dependencies.
 


size_t is the default type to use as index on arrays, since the "length"
property of arrays is of size_t. Since you are overloading the Index operator,
its only logical the type of the index to be size_t.

The code you provided isnt sufficient to help you, since we can't reproduce the
bug/problem. Either trim your code to a small example, or paste it at dpaste.com

strtr <strtr spam.com> writes:

daoryn Wrote:
 
 size_t is the default type to use as index on arrays, since the "length"
property of arrays is of size_t. Since you are overloading the Index operator,
its only logical the type of the index to be size_t.
 



 The code you provided isnt sufficient to help you, since we can't reproduce
the bug/problem. Either trim your code to a small example, or paste it at
dpaste.com
 


Also, I'm not posting the 15 modules of copyrighted code ;)

Don <nospam nospam.com> writes:

strtr wrote:
 or: Why does dmd keep on crashing on my code :)
 (I blame it on the average D programming; they don't write enough crappy code
) 
 
 After reading the -w post I tried using the -w switch again and it also
prevents me from seeing all my warning, but for a different reason : it crashes
dmd :(


 
 Before crashing a few correct warnings are shown :
 warning - tests.d(232): Error: implicit conversion of expression (i1) of type
uint to ubyte can cause loss of data
 
 Somehow I used size_t like this:
 struct S {
   ubyte[4] ub;
   void opIndex(size_t i1, size_t i2, size_t i3, size_t i4) {
     ub[0] = i1;
 ...
 }
 I have no clue as to why I put size_t there but replacing them with ubyte
seems to fix the crashing problem. 
 
 I can't replicate the crash in something small using only the struct or
something  :(
 Somehow it only works when at least twenty modules are imported in
(indirectly) and a lot of them have circular dependencies.