2nd try at this.

http://www.digitalmars.com/d/changelog.html

http://ftp.digitalmars.com/dmd.1.001.zip

Walter Bright wrote:
> 2nd try at this.
> 
> http://www.digitalmars.com/d/changelog.html
> 
> http://ftp.digitalmars.com/dmd.1.001.zip

I'm sure everyone will figure it out for themselves, but for everyone's one-clicking pleasure here's the correct link:

    http://ftp.digitalmars.com/dmd.1.002.zip

--bb

Walter Bright wrote:
> 2nd try at this.
> 
> http://www.digitalmars.com/d/changelog.html
> 
> http://ftp.digitalmars.com/dmd.1.001.zip

My problem's still not solved :(

String constants are pointing to wrong data at run-time. For example, std.string.split(a,"\t"), the "\t" points to random data.

I cannot seem to isolate it, and it works fine with 1.00.

L.

Lionello Lunesu wrote:
> My problem's still not solved :(
> 
> String constants are pointing to wrong data at run-time. For example, std.string.split(a,"\t"), the "\t" points to random data.
> 
> I cannot seem to isolate it, and it works fine with 1.00.

Try running the split earlier in the program.

Walter Bright wrote:
> Lionello Lunesu wrote:
>> My problem's still not solved :(
>>
>> String constants are pointing to wrong data at run-time. For example, std.string.split(a,"\t"), the "\t" points to random data.
>>
>> I cannot seem to isolate it, and it works fine with 1.00.
> 
> Try running the split earlier in the program.

It's in a unittest! :S
Anyway, I have it isolated, somewhat. Can I send you a .zip?

It also happens in linux, by the way. Just thought I'd test that, since in linux I would have gotten an access violation if something would overwrite string constants. Instead, I get the same assert failure.

L.

Walter Bright wrote:
> Lionello Lunesu wrote:
>> My problem's still not solved :(
>>
>> String constants are pointing to wrong data at run-time. For example, std.string.split(a,"\t"), the "\t" points to random data.
>>
>> I cannot seem to isolate it, and it works fine with 1.00.
> 
> Try running the split earlier in the program.

Here's some code that reproduces the problem. Note: the order of the modules on the dmd command line is important!

Also, the main module (m.d) does not use anything from the other two modules, but without them it works. Those two modules don't have static ctor/dtor, nor unittests.

You should get:
Error: AssertError Failure m(9)

L.

Lionello Lunesu wrote:
> Walter Bright wrote:
>> 2nd try at this.
>>
>> http://www.digitalmars.com/d/changelog.html
>>
>> http://ftp.digitalmars.com/dmd.1.001.zip
> 
> My problem's still not solved :(

Did you notice the direct link in the post is incorrect?
If you've accidentally downloaded v1.001 again, it's not so surprising it's not fixed...

Frits van Bommel wrote:
> Did you notice the direct link in the post is incorrect?

Yes, I got used to them : )

L.

Lionello Lunesu wrote:
> Frits van Bommel wrote:
>> Did you notice the direct link in the post is incorrect?
> 
> Yes, I got used to them : )

Me too, but since your post was essentially that you didn't see a difference (for your program) I couldn't just assume the same about you, now could I ;)?

Lionello Lunesu Wrote:

> Walter Bright wrote:
> > Lionello Lunesu wrote:
> >> My problem's still not solved :(
> >>
> >> String constants are pointing to wrong data at run-time. For example, std.string.split(a,"\t"), the "\t" points to random data.
> >>
> >> I cannot seem to isolate it, and it works fine with 1.00.
> > 
> > Try running the split earlier in the program.
> 
> Here's some code that reproduces the problem. Note: the order of the modules on the dmd command line is important!
> 
> Also, the main module (m.d) does not use anything from the other two modules, but without them it works. Those two modules don't have static ctor/dtor, nor unittests.
> 
> You should get:
> Error: AssertError Failure m(9)
> 
> L.
>