Please pardon me if anyone has mentioned this before:
compiler bugs should be separated from library bugs in DStress.

zwang wrote:
| Please pardon me if anyone has mentioned this before:
| compiler bugs should be separated from library bugs in DStress.

Talking from a Java point of view:
Only compiler and VM bugs are included in DStress.

Talking form a D point of view:
All those parts in Phobos that are part of the language standard
- - like GC, switch/case and int[].sort - are tested.
General lib features like std.string.toString aren't tested.

"VM" bugs can't be separated from "pure" compiler bugs.
- - int[].sort could be created by the compiler or call a library
function
- - a simple compiler might use generic lib features to do the GC but an
advanced compiler could create an explicit GC optimized for that very
source code

Thomas

Thomas Kühne wrote:
<snip>
> Talking form a D point of view:
> All those parts in Phobos that are part of the language standard - - like GC, switch/case and int[].sort - are tested.
> General lib features like std.string.toString aren't tested.
<snip>

By "language standard" do you mean the parts of the D standard that are about the language itself rather than the standard library?

Stewart.

-- 
My e-mail is valid but not my primary mailbox.  Please keep replies on on the 'group where everyone may benefit.

Stewart Gordon wrote:
| Thomas Kühne wrote:
| <snip>
|
|> Talking form a D point of view:
|> All those parts in Phobos that are part of the language standard - -
|> like GC, switch/case and int[].sort - are tested.
|> General lib features like std.string.toString aren't tested.
|
| <snip>
|
| By "language standard" do you mean the parts of the D standard that
| are about the language itself rather than the standard library?

Yes. The border case is http://dstress.kuehne.cn/run/align_08.d
GC is a language feature but direct access to the GC isn't.

In general everything in phobos/internal and phobos/std/typeinfo qualifies.

Thomas