It seems that Free Pascal 3.0.4 Sort has something wrong?

[shyakocat]

Today I found a fault(maybe?) of fpc 3.0.4 when I was solving an ACM problem.

I uses TOrderingArrayUtils to Sort an TVector,but it turns to RunTimeError at times.Then I view the source of garrayutils.pp on github.

Code: Pascal  [Select][+][-]

1.     while (cur <> 0) AND (TCompare.c(Arr[l],Arr[cur])) do
2.     begin
3.       temp:=Arr[cur];
4.       Arr[cur]:=Arr[l];
5.       Arr[l]:=temp;
6.       cur:=l;
7.       l:=Parent(cur-Start)+Start;
8.     end;

I think the

Code: Pascal  [Select][+][-]

1. cur <> 0

should be

Code: Pascal  [Select][+][-]

1. cur > Start

.Otherwise it will access a wrong position of the heap because (cur-Start) doesn't have a Parent.

But I'm not sure.Can anyone prove it and contact the authority?

[guest64953]

Today I found a fault(maybe?) of fpc 3.0.4 when I was solving an ACM problem.

I uses TOrderingArrayUtils to Sort an TVector,but it turns to RunTimeError at times.Then I view the source of garrayutils.pp on github.

Code: Pascal  [Select][+][-]

1.     while (cur <> 0) AND (TCompare.c(Arr[l],Arr[cur])) do
2.     begin
3.       temp:=Arr[cur];
4.       Arr[cur]:=Arr[l];
5.       Arr[l]:=temp;
6.       cur:=l;
7.       l:=Parent(cur-Start)+Start;
8.     end;

I think the

Code: Pascal  [Select][+][-]

1. cur <> 0

should be

Code: Pascal  [Select][+][-]

1. cur > Start

.Otherwise it will access a wrong position of the heap because (cur-Start) doesn't have a Parent.

But I'm not sure.Can anyone prove it and contact the authority?

I don't think this is the right sub forum to post this.

[Martin_fr]

Moved the topic to better board

[BrunoK]

Where is the better board ?

[Martin_fr]

It is now in "Free Pascal » General", instead of "Third Party announcements"
Please go back to the orig topic, or otherwise open your own thread for "moving topic" related questions"