July 06, 2012
On Friday, 6 July 2012 at 09:38:13 UTC, Jacob Carlborg wrote:
> It will not prevent the compiler for generating calls to the runtime.

It should – TypeInfo references will still be generated, though.

> So if you see calls to runtime functions you need to stub them out as Alex suggested or try to find a way to avoid them. Note that the compiler will link with the standard library by default as well. I don't know if the -noruntime flag prevents that.

This is a good point – it doesn't. Use -nodefaultlib (LDC's -defaultlib switch works a bit differently than DMD's in that it appends to a list of default libs instead of replacing, this should imho be fixed).

David
July 06, 2012
I swear you guys read my mind sometimes...  It's creepy.

I just had this very issue, doing the exact same thing, about an hour ago.

Have you tried with -nodefaultlib -noruntime ?  Cause that's what works for me...

I just got *something* to compile with no runtime or std.
Whether or not it actually does anything remains to be seen.


On Fri, 06 Jul 2012 01:53:10 -0500, BLM768 <blm768@gmail.com> wrote:

> I've been trying to write an OS kernel in D, and I'm having issues with the runtime. I'm trying to use LDC's -noruntime option, which is _supposed_ to prevent any runtime calls from being generated, but the linker keeps complaining about unresolved references to _d_assert_msg and other runtime functions. It seems that LDC is ignoring the switch and is generating runtime references anyway. Is there some other way to force it to stop trying to pull in the runtime? I'd rather not have to create/link a custom runtime at this point; I don't even have a memory allocator written, so I really don't want to mess with a runtime yet.


-- 
Using Opera's revolutionary email client: http://www.opera.com/mail/
July 07, 2012
On Friday, 6 July 2012 at 21:54:15 UTC, 1100110 wrote:
> I swear you guys read my mind sometimes...  It's creepy.
>
> I just had this very issue, doing the exact same thing, about an hour ago.
>
> Have you tried with -nodefaultlib -noruntime ?  Cause that's what works for me...
>
> I just got *something* to compile with no runtime or std.
> Whether or not it actually does anything remains to be seen.
>
>

No luck; it still references the runtime.
I've been stubbing out the runtime; it seems to be about done except for references to some functions I can't find. The symbols are __moddi3 and __divdi3; I assume they're C math library functions, as they're generated from code that uses div/mod operations. It also seems to be unable to properly find Object.toString() after I changed it to be nothrow, which I needed to do because I'm stubbing out the exception handling routines; it seems to have changed the mangling. It's a bit of a mess in there; I'm glad I don't always have to hack up the runtime :).


July 07, 2012
On Saturday, July 07, 2012 05:45:53 BLM768 wrote:
> On Friday, 6 July 2012 at 21:54:15 UTC, 1100110 wrote:
> > I swear you guys read my mind sometimes...  It's creepy.
> > 
> > I just had this very issue, doing the exact same thing, about an hour ago.
> > 
> > Have you tried with -nodefaultlib -noruntime ?  Cause that's what works for me...
> > 
> > I just got *something* to compile with no runtime or std. Whether or not it actually does anything remains to be seen.
> 
> No luck; it still references the runtime.
> I've been stubbing out the runtime; it seems to be about done
> except for references to some functions I can't find. The symbols
> are __moddi3 and __divdi3; I assume they're C math library
> functions, as they're generated from code that uses div/mod
> operations. It also seems to be unable to properly find
> Object.toString() after I changed it to be nothrow, which I
> needed to do because I'm stubbing out the exception handling
> routines; it seems to have changed the mangling. It's a bit of a
> mess in there; I'm glad I don't always have to hack up the
> runtime :).

Yes. nothrow is part of the name mangling, because it's part of the signature. In the long run, toString will be @safe const pure nothrow, but it's not there quite yet (const correctness and Object is still being sorted out would be one reason; a number of key string-related functions need to become pure for another). But if the compiler is expecting a specific signature, then that's the signature that you're going to have to give it, or the linker's not going to find the function when it goes to look for it.

- Jonathan M Davis
July 07, 2012
Wow, I haven't had that much trouble.  But I've tried to keep everything
at my level.  It's about half xomb. =P

I salute you for your bravery.

I had planned to stay far, far away from as much of that as I could.



On Fri, 06 Jul 2012 22:45:53 -0500, BLM768 <blm768@gmail.com> wrote:

> On Friday, 6 July 2012 at 21:54:15 UTC, 1100110 wrote:
>> I swear you guys read my mind sometimes...  It's creepy.
>>
>> I just had this very issue, doing the exact same thing, about an hour ago.
>>
>> Have you tried with -nodefaultlib -noruntime ?  Cause that's what works for me...
>>
>> I just got *something* to compile with no runtime or std.
>> Whether or not it actually does anything remains to be seen.
>>
>>
>
> No luck; it still references the runtime.
> I've been stubbing out the runtime; it seems to be about done except for references to some functions I can't find. The symbols are __moddi3 and __divdi3; I assume they're C math library functions, as they're generated from code that uses div/mod operations. It also seems to be unable to properly find Object.toString() after I changed it to be nothrow, which I needed to do because I'm stubbing out the exception handling routines; it seems to have changed the mangling. It's a bit of a mess in there; I'm glad I don't always have to hack up the runtime :).
>
>


-- 
Using Opera's revolutionary email client: http://www.opera.com/mail/
July 07, 2012
On 07-07-2012 06:26, Jonathan M Davis wrote:
> On Saturday, July 07, 2012 05:45:53 BLM768 wrote:
>> On Friday, 6 July 2012 at 21:54:15 UTC, 1100110 wrote:
>>> I swear you guys read my mind sometimes...  It's creepy.
>>>
>>> I just had this very issue, doing the exact same thing, about
>>> an hour ago.
>>>
>>> Have you tried with -nodefaultlib -noruntime ?  Cause that's
>>> what works for me...
>>>
>>> I just got *something* to compile with no runtime or std.
>>> Whether or not it actually does anything remains to be seen.
>>
>> No luck; it still references the runtime.
>> I've been stubbing out the runtime; it seems to be about done
>> except for references to some functions I can't find. The symbols
>> are __moddi3 and __divdi3; I assume they're C math library
>> functions, as they're generated from code that uses div/mod
>> operations. It also seems to be unable to properly find
>> Object.toString() after I changed it to be nothrow, which I
>> needed to do because I'm stubbing out the exception handling
>> routines; it seems to have changed the mangling. It's a bit of a
>> mess in there; I'm glad I don't always have to hack up the
>> runtime :).
>
> Yes. nothrow is part of the name mangling, because it's part of the signature.
> In the long run, toString will be @safe const pure nothrow, but it's not there
> quite yet (const correctness and Object is still being sorted out would be one
> reason; a number of key string-related functions need to become pure for
> another). But if the compiler is expecting a specific signature, then that's
> the signature that you're going to have to give it, or the linker's not going
> to find the function when it goes to look for it.
>
> - Jonathan M Davis
>

Is overriding a @safe function with @trusted allowed/meant to be allowed? I hope so, otherwise this is going to be a severe limitation.

-- 
Alex Rønne Petersen
alex@lycus.org
http://lycus.org


July 07, 2012
On Saturday, 7 July 2012 at 04:27:07 UTC, Jonathan M Davis wrote:
> On Saturday, July 07, 2012 05:45:53 BLM768 wrote:
>> On Friday, 6 July 2012 at 21:54:15 UTC, 1100110 wrote:
>> > I swear you guys read my mind sometimes...  It's creepy.
>> > 
>> > I just had this very issue, doing the exact same thing, about
>> > an hour ago.
>> > 
>> > Have you tried with -nodefaultlib -noruntime ?  Cause that's
>> > what works for me...
>> > 
>> > I just got *something* to compile with no runtime or std.
>> > Whether or not it actually does anything remains to be seen.
>> 
>> No luck; it still references the runtime.
>> I've been stubbing out the runtime; it seems to be about done
>> except for references to some functions I can't find. The symbols
>> are __moddi3 and __divdi3; I assume they're C math library
>> functions, as they're generated from code that uses div/mod
>> operations. It also seems to be unable to properly find
>> Object.toString() after I changed it to be nothrow, which I
>> needed to do because I'm stubbing out the exception handling
>> routines; it seems to have changed the mangling. It's a bit of a
>> mess in there; I'm glad I don't always have to hack up the
>> runtime :).
>
> Yes. nothrow is part of the name mangling, because it's part of the signature.
> In the long run, toString will be @safe const pure nothrow, but it's not there
> quite yet (const correctness and Object is still being sorted out would be one
> reason; a number of key string-related functions need to become pure for
> another). But if the compiler is expecting a specific signature, then that's
> the signature that you're going to have to give it, or the linker's not going
> to find the function when it goes to look for it.
>
> - Jonathan M Davis

The problem is that if I leave the try/catch block, it's referring to stubbed-out functions, and if I get rid of it, LDC complains that toString() isn't nothrow. I guess I'd just better leave it in and hope that toString() never gets called. I could throw in an assert(false), which I'll have wired up to just cause a kernel panic.


July 07, 2012
On Saturday, July 07, 2012 07:27:52 Alex Rønne Petersen wrote:
> Is overriding a @safe function with @trusted allowed/meant to be allowed? I hope so, otherwise this is going to be a severe limitation.

Well, an overriding function in a derived class cannot be any looser than the one it's overriding in the base class, but @trusted is essentially the same as @safe from the caller's perspective, so I would expect it to work (I wouldn't expect @system to work though, since that _is_ looser). In fact, I'm pretty sure that there was a post about it not too long ago where someone was having issues because the function in a derived class was inferred as @safe, and they had to mark it as @trusted to get it to work. So, I think that you're fine on that count. However, even if you _couldn't_ override an @safe function with an @trusted one for some reason, all you would have to do is create a helper function which was @trusted, and you could get around the limitation quite easily.

- Jonathan M Davis
July 07, 2012
On 2012-07-06 21:42, David Nadlinger wrote:
> On Friday, 6 July 2012 at 09:38:13 UTC, Jacob Carlborg wrote:
>> It will not prevent the compiler for generating calls to the runtime.
>
> It should – TypeInfo references will still be generated, though.

So what happens with code like this:

auto o = new Object;

It won't generate any code for that? Then what happens with "o", uninitialized?

-- 
/Jacob Carlborg


1 2
Next ›   Last »