Re: Add a precompiled c++ obj file to dub
On Sunday, 8 October 2017 at 02:58:36 UTC, Fra Mecca wrote: On Saturday, 7 October 2017 at 23:54:50 UTC, user1234 wrote: On Saturday, 7 October 2017 at 19:56:52 UTC, Fra Mecca wrote: Hi all, I am writing a backend that is partly Vibe.d and partly clucene in c++. I have some object files written in c++ and compiled with g++ that are not considered by dub during the linking phase and throws `function undefined error ` every time. Is there a way to tell dub to let dmd handle that .o files? Yes, add this to your JSON: "sourceFiles-linux-x86_64" : [ "somepath/yourobject.o" ], I tried the sourceFiles approach, it failed and I could reproduce that in some days. At the end I added them as linking options (lflags) but it is kinda odd that it works given that everything is supplied to dmd as -Lobj.o Huh, i'm surprised but well, if it works for you. My advice was based on https://github.com/BBasile/dbeaengine/blob/master/dub.json (object file is passed to dmd) which works, I often use it.
Re: Add a precompiled c++ obj file to dub
On Saturday, 7 October 2017 at 23:54:50 UTC, user1234 wrote: On Saturday, 7 October 2017 at 19:56:52 UTC, Fra Mecca wrote: Hi all, I am writing a backend that is partly Vibe.d and partly clucene in c++. I have some object files written in c++ and compiled with g++ that are not considered by dub during the linking phase and throws `function undefined error ` every time. Is there a way to tell dub to let dmd handle that .o files? Yes, add this to your JSON: "sourceFiles-linux-x86_64" : [ "somepath/yourobject.o" ], I tried the sourceFiles approach, it failed and I could reproduce that in some days. At the end I added them as linking options (lflags) but it is kinda odd that it works given that everything is supplied to dmd as -Lobj.o
Re: Double ended arrays?
On 10/07/2017 05:02 PM, Steven Schveighoffer wrote: > https://github.com/schveiguy/dcollections/blob/master/dcollections/Deque.d > > It's implemented by maintaining 2 dynamic arrays, one that is "reversed" > at the front, and one that is normal at the back. When you prepend, it > appends to the "reverse" array. > > It's probably not the most efficient, but it does maintain the correct > complexities. I stole the idea from one of Chuck Allison's DConf talks[1] and used as the example for the Indexing Operators section here: http://ddili.org/ders/d.en/operator_overloading.html#ix_operator_overloading.opIndex > Note: that code is many years old, so it may not compile with the latest > compiler. Mine is supposed to compile with 2.076. > -Steve Ali [1] He knows about the theft. :)
Re: Double ended arrays?
On 10/7/17 3:38 AM, Chirs Forest wrote: I have some data that I want to store in a dynamic 2d array... I'd like to be able to add elements to the front of the array and access those elements with negative integers as well as add numbers to the back that I'd acess normally with positive integers. Is this something I can do, or do I have to build a container to handle what I want? Dcollections has something like this, a deque. It doesn't use negative integers to access the prepended elements, but I suppose it could be made to do this. See here: https://github.com/schveiguy/dcollections/blob/master/dcollections/Deque.d It's implemented by maintaining 2 dynamic arrays, one that is "reversed" at the front, and one that is normal at the back. When you prepend, it appends to the "reverse" array. It's probably not the most efficient, but it does maintain the correct complexities. Note: that code is many years old, so it may not compile with the latest compiler. -Steve
Re: Add a precompiled c++ obj file to dub
On Saturday, 7 October 2017 at 19:56:52 UTC, Fra Mecca wrote: Hi all, I am writing a backend that is partly Vibe.d and partly clucene in c++. I have some object files written in c++ and compiled with g++ that are not considered by dub during the linking phase and throws `function undefined error ` every time. Is there a way to tell dub to let dmd handle that .o files? Yes, add this to your JSON: "sourceFiles-linux-x86_64" : [ "somepath/yourobject.o" ],
Add a precompiled c++ obj file to dub
Hi all, I am writing a backend that is partly Vibe.d and partly clucene in c++. I have some object files written in c++ and compiled with g++ that are not considered by dub during the linking phase and throws `function undefined error ` every time. Is there a way to tell dub to let dmd handle that .o files?
exclude members at compile time
Ok, what I'm trying to do is the following:
take a type and a value of its type; given a known member of the
type, (re?)create a similar type, without this very known member.
What does work is this:
/// --- code ---
void main()
{
S s;
s.i = 42;
s.d = 73.0;
s.s.length = 5;
s.s[] = 1024;
s.c = 'c';
auto n = exclude!"i"(s);
static assert(!__traits(compiles, n.i));
assert(n.d == s.d);
assert(n.s == s.s);
assert(n.c == s.c);
n.s[0] = 55;
assert(n.s == s.s); // works, even by reference.
}
struct S
{
int i;
double d;
size_t[] s;
char c;
//auto fun(){} // <-- a function does not work, line 26
}
auto exclude(string without, T)(T t)
{
immutable b = [ __traits(allMembers, T) ];
struct Result
{
static foreach(i, _; b)
{
static if(without != b[i])
{
mixin(
typeof(__traits(getMember, T,
b[i])).stringof ~
" " ~
__traits(getMember, T, b[i]).stringof ~
";"
);
}
}
}
Result res;
static foreach(i, _; b)
{
static if(without != b[i])
{
__traits(getMember, res, b[i]) = __traits(getMember, t,
b[i]);
}
}
return res;
}
/// --- code ---
And the problem lies in line 26: If there is a method inside the
type, it won't work this way.
I assume, there should be a way, to copy the original type first
and to hide a member afterwards. Is there a simple approach for
this?
I think, the problem with my approach is, that if the member is
not a "data" member, than there are a plenty of possibilities to
check, like... function, template, well... there could be very
much types of indirections, as checking for "isCallable" is not
enough for recreation(?)
If these checks are the only way to handle this, than the answer
to my whole question is:
"recreation is restricted to PODs".
Then, I have to consider which checks to implement, and which
restrictions have to be applied to the input...
Re: Infuriating DUB/DMD build bug.
On Friday, 6 October 2017 at 23:02:56 UTC, Laeeth Isharc wrote: On Thursday, 5 October 2017 at 21:48:20 UTC, WhatMeWorry wrote: I've got a github project and using DUB with DMD and I keep running into this problem. I've tried deleting the entire ...\AppData\Roaming\dub\packages folder, but the problem repeats the very next build attempt. [...] See my post in learn on dmd path. The dmd path code was written in 1987 and could do with an update to process longer windows paths properly. We are working on this and I guess a chance a pull request on Monday but it depends on what else comes up. In any case it's a trivial fix. Presuming I am right about it being a path length problem. I did! But i didn't say anything because i wasn't sure if they were related. I'm pretty sure it is path related because the exact same dub.sdl files work fine on Linux and MacOS. (It's a cross platform project) Glad it is a trivial fix. Curious what it involves. Let me know if I can help out in any way. Mike Parker was kind enough to show me a manual dub local workaround for this issue. But I'll hold off now and see if your change does the fix. If it does, it will be the best timed bug fix ever :)
Re: Implementing swap for user-defined swaps
On 10/07/2017 07:55 PM, Balagopal Komarath wrote: I was implement my own range type that forwards all accesses to another range. I tried to write a `swap` function so that sort etc. could be called on my range. However, I cannot get `hasSwappableElements!ARange` to evaluate to true. But, when I copy pasted the definition of `hasSwappableElements` into my module and evaluated it, it was true. I assume this is something to do with how name lookup works in D. Here's simplified code that reproduces the same problem. https://dpaste.dzfl.pl/48a420cce261 You can't provide your own swap function. `sort` etc. use std.algorithm.swap, so your range has to work with that.
Re: @nogc formattedWrite
On Saturday, 7 October 2017 at 18:14:00 UTC, Nordlöw wrote: It would be nice to be able to formatted output in -betterC... Agreed. If you know the size of the buffer, you can use sformat, which might be @nogc, but I don't know if it's compatible with betterC. Also, you might check out Ocean, which might have nogc formatting. https://github.com/sociomantic-tsunami/ocean
@nogc formattedWrite
Is it currently possible to somehow do @nogc formatted output to
string?
I'm currently using my `pure @nogc nothrow` array-container
`CopyableArray` as
@safe pure /*TODO nothrow @nogc*/ unittest
{
import std.format : formattedWrite;
const x = "42";
alias A = CopyableArray!(char);
A a;
a.formattedWrite!("x : %s")(x);
assert(a == "x : 42");
}
but I can't tag the unittest as `nothrow @nogc` because of the
call to `formattedWrite`. Is this because `formattedWrite`
internally uses the GC for buffer allocations or because it may
throw?
It would be nice to be able to formatted output in -betterC...
Implementing swap for user-defined swaps
Hello, I was implement my own range type that forwards all accesses to another range. I tried to write a `swap` function so that sort etc. could be called on my range. However, I cannot get `hasSwappableElements!ARange` to evaluate to true. But, when I copy pasted the definition of `hasSwappableElements` into my module and evaluated it, it was true. I assume this is something to do with how name lookup works in D. Here's simplified code that reproduces the same problem. https://dpaste.dzfl.pl/48a420cce261 Thanks, Balagopal.
Re: DLL hell :S
On Saturday, 7 October 2017 at 15:30:30 UTC, rikki cattermole wrote: A little from column a, a little from column b, but most because he might be able to do something for you. Thanks, I'll send him an email.
Re: DLL hell :S
On 07/10/2017 4:29 PM, Ian Hatch wrote: On Saturday, 7 October 2017 at 15:14:01 UTC, rikki cattermole wrote: Email Walter directly. I intend to campaign for next years (basically a soft TODO list) plan for what we want done. But until then, he and Andrei need to hear that this is the biggest limitation that D faces currently, not memory management. Hm, are you saying "if you email Walter he can tell you how to sort it" or "please make sure Walter knows this problem is important" (or both)? Definitely happy to give my feedback. A little from column a, a little from column b, but most because he might be able to do something for you. "not memory management" gives me an idea actually - if I ditch the GC, which I may want to do eventually anyway, I guess I won't have this issue.
Re: DLL hell :S
On Saturday, 7 October 2017 at 15:14:01 UTC, rikki cattermole wrote: Email Walter directly. I intend to campaign for next years (basically a soft TODO list) plan for what we want done. But until then, he and Andrei need to hear that this is the biggest limitation that D faces currently, not memory management. Hm, are you saying "if you email Walter he can tell you how to sort it" or "please make sure Walter knows this problem is important" (or both)? Definitely happy to give my feedback. "not memory management" gives me an idea actually - if I ditch the GC, which I may want to do eventually anyway, I guess I won't have this issue.
Re: DLL hell :S
Email Walter directly. I intend to campaign for next years (basically a soft TODO list) plan for what we want done. But until then, he and Andrei need to hear that this is the biggest limitation that D faces currently, not memory management.
DLL hell :S
Hello! I'm Ian, and I've been a programmer in games for 10 years. I've been poking at D for a year or so and I'm absolutely in love with the compile-time execution and inline unit testing in particular. I've been trying for a while to set up a project that I intend to build a lot of my future code on top of, but I'm stumbling at one of the first hurdles at the moment so I'm hoping I can straighten this out and be able to keep using D. My framework is supposed to load plugin dlls (and later sharedobjects) which operate on memory allocated by the executable, and sadly I've hit roadblock after roadblock trying to get this set up. Following https://wiki.dlang.org/Win32_DLLs_in_D I've gotten to a stage where I successfully load the DLL, find my functions, and call them passing the objects to be modified. Initially this was crashing immediately. After a lot of digging I found https://issues.dlang.org/show_bug.cgi?id=17326, so now I've added the gcopt to my dll to set the gc to manual, and I'm passing over my gc pointer and trying to set it as the proxy. Unfortunately this results in the DLL getting stuck in a spinlock in addRange within gc_setProxy() :( DLL side: https://pastebin.com/yBPs0A30 EXE side: https://pastebin.com/h2qLBqXA Thread lock call stack: https://pastebin.com/zUvH3Cnb I'm not really sure where to go from here. I'm doing this in my spare time and I've been stuck for months on something I know how to do in C++ (I really should have asked for help earlier). Is there an up to date example someone can point me to of how to set this up correctly so that the DLL uses the exe's runtime and/or garbage collector?
Re: Need importing dcompute.lib into my project
On Saturday, 7 October 2017 at 12:12:10 UTC, kinke wrote: On Saturday, 7 October 2017 at 09:04:26 UTC, kerdemdemir wrote: Error: static assert "Need to use a DCompute enabled compiler" Are you using latest LDC 1.4? The CUDA backend wasn't enabled for earlier versions. Yes I am. Actually I can build dcompute's source code so I am sure LDC version is good. Now I am trying to use dcompute within my project.
Re: Need importing dcompute.lib into my project
On Saturday, 7 October 2017 at 09:04:26 UTC, kerdemdemir wrote: Error: static assert "Need to use a DCompute enabled compiler" Are you using latest LDC 1.4? The CUDA backend wasn't enabled for earlier versions.
Re: Need importing dcompute.lib into my project
do you set "-mdcompute-targets=cuda-xxx" in the dflags for your
dub.json for your project?
I have added now after your comment.
But it seems it didn't changed anything.
Here is the dub.json file I have:
{
"name": "dsharpear",
"authors": [
"Erdem"
],
"dflags" : ["-mdcompute-targets=cuda-210" ,"-oq", "-betterC"],
"dependencies": {
"dcompute": ">=0.0.0-alpha0 <0.1.0"
},
"description": "Beamforming with D ",
"copyright": "Copyright © 2017, Erdem",
"license": "proprietary"
}
And running "dub build --compiler=D:\LDCDownload\bin\ldc2.exe
--force" fails with:
Performing "debug" build using D:\LDCDownload\bin\ldc2.exe for
x86.
derelict-util 2.1.0: building configuration "library"...
derelict-cl 2.0.0: building configuration "library"...
derelict-cuda 2.0.1: building configuration "library"...
..\..\..\..\..\AppData\Roaming\dub\packages\derelict-cuda-2.0.1\derelict-cuda\source\derelict\cuda\runtimeapi.d(816,5):
Deprecation: constructor derelict.cuda.runtimeapi.dim3.this
all parameters have default arguments, but structs cannot have
default constructors.
dcompute 0.0.0-alpha0: building configuration "library"...
Targeting 'i686-pc-windows-msvc' (CPU 'pentium4' with features '')
Building type: real
Building type: uint
Building type: char
Building type: ubyte
Building type: ulong
Building type: int
Building type: double
Building type: long
Building type: ushort
Building type: wchar
Building type: byte
Building type: short
Building type: float
Building type: dchar
..\..\..\..\..\AppData\Roaming\dub\packages\dcompute-0.0.0-alpha0\dcompute\source\dcompute\std\package.d(6,5):
Error: static assert "Need to use a DCompute enabled compiler
See https://github.com/thewilsonator/ldc/tree/dcompute;
D:\LDCDownload\bin\ldc2.exe failed with exit code 1.
Re: Need importing dcompute.lib into my project
On Saturday, 7 October 2017 at 09:04:26 UTC, kerdemdemir wrote: You should add DCompute as a DUB dependancy. Hi, I inited my project with Dub by unsing "dub init DSharpEar" and I added dependency "dcompute". Even I give extra parameters for using ldc. dub build --compiler=D:\LDCDownload\bin\ldc2.exe --force I am getting : Error: static assert "Need to use a DCompute enabled compiler" Any suggestions? Regards Erdem do you set "-mdcompute-targets=cuda-xxx" in the dflags for your dub.json for your project?
Re: Need importing dcompute.lib into my project
You should add DCompute as a DUB dependancy. Hi, I inited my project with Dub by unsing "dub init DSharpEar" and I added dependency "dcompute". Even I give extra parameters for using ldc. dub build --compiler=D:\LDCDownload\bin\ldc2.exe --force I am getting : Error: static assert "Need to use a DCompute enabled compiler" Any suggestions? Regards Erdem
Re: Double ended arrays?
On Saturday, October 07, 2017 07:38:47 Chirs Forest via Digitalmars-d-learn wrote: > I have some data that I want to store in a dynamic 2d array... > I'd like to be able to add elements to the front of the array and > access those elements with negative integers as well as add > numbers to the back that I'd acess normally with positive > integers. Is this something I can do, or do I have to build a > container to handle what I want? Dynamic arrays only support concatenating to the end, and they use size_t for indices and length, and size_t is unsigned. The standard library does the same with any containers that it has as do most 3rd party libraries. You'll need to either implement what you want yourself or use something from somewhere like code.dlang.org. Based on Ilya's post, it sounds like you may be able to use Mir for what you want, but I'd be very surprised to see any libraries use negative indices for anything - especially since most everything uses size_t for indices. - Jonathan M Davis
Re: Double ended arrays?
On Saturday, 7 October 2017 at 07:38:47 UTC, Chirs Forest wrote: I have some data that I want to store in a dynamic 2d array... I'd like to be able to add elements to the front of the array and access those elements with negative integers as well as add numbers to the back that I'd acess normally with positive integers. Is this something I can do, or do I have to build a container to handle what I want? Mir Algorithm [1] has 2D arrays. Elements can be added to the front/back of each dimension using `concatenation` routine [2]. In the same time it does not support negative indexes. [1] https://github.com/libmir/mir-algorithm [2] http://docs.algorithm.dlang.io/latest/mir_ndslice_concatenation.html#.concatenation Best regards, Ilya Yaroshenko
Double ended arrays?
I have some data that I want to store in a dynamic 2d array... I'd like to be able to add elements to the front of the array and access those elements with negative integers as well as add numbers to the back that I'd acess normally with positive integers. Is this something I can do, or do I have to build a container to handle what I want?
