Re: ini library in OSX
On Monday, 13 October 2014 at 16:06:42 UTC, Robert burner Schadek wrote: On Saturday, 11 October 2014 at 22:38:20 UTC, Joel wrote: On Thursday, 11 September 2014 at 10:49:48 UTC, Robert burner Schadek wrote: some self promo: http://code.dlang.org/packages/inifiled I would like an example? go to the link and scroll down a page How do you use it with current ini files ([label] key=name)?
Re: DUB build questions
On Sat, 2014-12-20 at 05:46 +, Dicebot via Digitalmars-d-learn wrote: On Saturday, 20 December 2014 at 04:15:00 UTC, Rikki Cattermole wrote: b) Can I do parallel builds with dub. CMake gives me Makefiles so I can make -j does dub have a similar option? No Worth noting that it is not actually a dub problem as much, it is simply not worth adding parallel builds because separate compilation is much much slower with existing D front-end implementation and even doing it in parallel is sub-optimal compared to dump-it-all-at-once. From previous rounds of this sort of question (for the SCons D tooling), the consensus of the community appeared to be that the only time separate module compilation was really useful was for mixed D, C, C++, Fortran systems. For pure D systems, single call of the compiler is deemed far better than traditional C, C++, Fortran compilation strategy. This means the whole make -j thing is not an issue, it just means that Dub is only really dealing with the all D situation. The corollary to this is that DMD, LDC and GDC really need to make use of all parallelism they can, which I suspect is more or less none. Chapel has also gone the compile all modules with a single compiler call strategy as this enables global optimization from source to executable. -- Russel. = Dr Russel Winder t: +44 20 7585 2200 voip: sip:russel.win...@ekiga.net 41 Buckmaster Roadm: +44 7770 465 077 xmpp: rus...@winder.org.uk London SW11 1EN, UK w: www.russel.org.uk skype: russel_winder
Re: Derelict SDL2 library not loading on OS X
On 12/20/2014 11:46 AM, Joel wrote: To uninstall SDL, do I just wipe the framework, and SDL dylib? Sorry, I can't help you there. I have no idea how things are done on Mac. And I think it depends on how it got there in the first place. You may want to take that question to the SDL mailing list (web interface at [1]) if no one answers here. [1] https://forums.libsdl.org/
Re: Get vars in current scope at compile time?
Rikki Cattermole: No way to do this. But perhaps it's worth supporting as future enhancement with a __traits. What are the use cases? Bye, bearophile
Re: Get vars in current scope at compile time?
On 20/12/2014 11:03 p.m., bearophile wrote:
Rikki Cattermole:
No way to do this.
But perhaps it's worth supporting as future enhancement with a __traits.
What are the use cases?
Bye,
bearophile
Short answer, I'm not keen on the idea, at least not yet.
I would far more comfortable once I've categorized the different types
of CTFE implements there can be. Since nobody has really done this before.
What makes me more uncomfortable to is DIP50. Looking over it again,
this is getting rather close to macro systems in LISP family of language
who do have full AST access/modification capabilities. So in other
words, its just asking for trouble without some proper thought.
Long answer, lets consider this code:
void myfunc() {
int x, y, z;
myvalues();
}
void myvalues(string file = __FILE__, string func = __FUNCTION__, int
line = __LINE__)() {
__traits(functionValues, file, func, line, z) =
__traits(functionValues, file, func, line, x) +
__traits(functionValues, file, func, line, y)
}
This was meant to be a little off from what was asked. Because I know
this will be wanted eventually.
Its nasty ambiguous code.
A better way might be to pass in a super function state e.g.
void myvalues(SSTATET)(SSTATET SSTATE = __SUPERSTATE__) {
with(SSTATE) {
z = x + y;
}
}
Same idea, but could be used for e.g. mixin templates and what have you.
Want all members of SSTATE? __traits(allMembers to the rescue!
Either way, not really D.learn material.
Re: Unittest in a windows app
I managed to isolate the problem to the following. Program 1
below works (displays unit test failure when run), while program
2 does not.
* Program 1 *
import std.stdio;
unittest
{
assert(false);
}
void main()
{
writeln(Hello D-World!);
}
* Program 2 *
module winmain;
import core.sys.windows.windows;
unittest {
assert(false);
}
extern (Windows)
int WinMain(HINSTANCE hInstance, HINSTANCE hPrevInstance, LPSTR
lpCmdLine, int nCmdShow)
{
return 0;
}
Re: Unittest in a windows app
Can it be because Windows main wrapper consumes exceptions or spawn a separate thread for that or something like that?
On inheritance and polymorphism in cats and dogs (and other beasts too)
So, I have this pet project where classes Cat and Dog inherit
from the more generic Beast class.
All beasts prosper and multiply and so do cats and dogs. The
breeding routine is fairly constant across species, with minor
variations. So I'd like to define the breed method in the Beast
class and overload it in the Cat and Dog classes.
Something like :
Cat rita = new Cat(female);
Cat ringo= new Cat(male);
Cat junior=rita.breed(ringo);
with
Class Cat:Beast{
...
Cat breed(Cat sire){
// do what all beasts do
//then add cat-specific genetics
}
...
}
Now, what I can't seem to figure out is the // do what all beast
do part. this current object is obviously an instance of the
Cat class. How do I de-specialize it so it can behave as an
instance of the more generic Beast class ? Then, the offspring
will in turn be an instance of Beast : how to cast it as a Cat ?
So far, all I've been able to do is to dance around the issue by
writting ad-hoc constructors like
Beast b=new Beast( someCat )
or
Cat c=new Cat(someBeast)
but this seems awkward and inefficient.
There must be some more clever and straightforward way to do
this, surely ?
Also : D is not my first OO language, but polymorphism and
inheritance still are advanced concepts for me.
Re: On inheritance and polymorphism in cats and dogs (and other beasts too)
On Saturday, 20 December 2014 at 15:40:32 UTC, Derix wrote:
So, I have this pet project where classes Cat and Dog inherit
from the more generic Beast class.
All beasts prosper and multiply and so do cats and dogs. The
breeding routine is fairly constant across species, with minor
variations. So I'd like to define the breed method in the
Beast
class and overload it in the Cat and Dog classes.
Something like :
Cat rita = new Cat(female);
Cat ringo= new Cat(male);
Cat junior=rita.breed(ringo);
with
Class Cat:Beast{
...
Cat breed(Cat sire){
// do what all beasts do
//then add cat-specific genetics
}
...
}
Now, what I can't seem to figure out is the // do what all
beast
do part. this current object is obviously an instance of the
Cat class. How do I de-specialize it so it can behave as an
instance of the more generic Beast class ? Then, the offspring
will in turn be an instance of Beast : how to cast it as a Cat ?
So far, all I've been able to do is to dance around the issue by
writting ad-hoc constructors like
Beast b=new Beast( someCat )
or
Cat c=new Cat(someBeast)
but this seems awkward and inefficient.
There must be some more clever and straightforward way to do
this, surely ?
Also : D is not my first OO language, but polymorphism and
inheritance still are advanced concepts for me.
What you do is to implement the breed method in Beast class,
then override it in both Cat and Dog, but then call
super.breed() method which belongs to Beast.
Re: On inheritance and polymorphism in cats and dogs (and other beasts too)
On Saturday, 20 December 2014 at 15:40:32 UTC, Derix wrote:
// do what all beasts do
You'll want to call the function in the base class, which is done
with the super keyword in D.
I wouldn't make the super function return the new instance
though, that can't be as easily customized. I'd separate it into
two functions: breed and genetic mix
class Beast {
// call this on a new zygote
// it is protected because it doesn't make sense to mix
genetics
// in public
protected void mixGenetics(Beast mother, Beast father) {
// modify this to be a combination of generic traits
// for example
this.furColor = mother.furColor + father.furColor;
}
}
Now, make the specific breed functions on the subclasses. I
didn't put that in the interface because you typically don't want
to substitute parents with generic beasts - cats and dogs can't
breed together, but if there was a breed(Beast) function in the
base class, that would be permitted. (in OO theory, this is the
liskov substitution principle)
class Cat : Beast {
Cat breed(Cat sire) {
Cat offspring = new Cat();
offspring.mixGenetics(this, sire);
return offspring;
}
protected override void mixGenetics(Cat mother, Cat father) {
super(mother, father); // calls the generic
Beast.mixGenetics
// now we can do cat-specific stuff
this.pickiness = mother.pickiness * father.pickiness;
}
}
Dog would look similar to cat.
However, since mixGenetics really only makes sense when making a
new Cat, you might want to make it a constructor instead of a
method, then follow the same pattern.
Cat class. How do I de-specialize it so it can behave as an
instance of the more generic Beast class ? Then, the offspring
will in turn be an instance of Beast : how to cast it as a Cat ?
Generally speaking, de-specialization happens automatically. The
super keyword calls the parent class' version which takes care of
all that.
You can cast Beasts back into Cats with cast(Cat) beast. Be sure
to check for null - if beast is a dog, that cast will return null.
So far, all I've been able to do is to dance around the issue by
writting ad-hoc constructors like
that makes some sense, you'd just want to make sure the specific
ones are done in specific child constructors, then call the
parent with super()
Re: Unittest in a windows app
On 19.12.2014 22:39, Dan Nestor wrote:
Hello everybody, this is my first post on this forum.
I have a question about unit testing a Windows application. I
have slightly modified Visual D's default Windows application
stub to the following:
[...]
try
{
Runtime.initialize();
Runtime.initialize() no longer calls the unittests since a few versions.
You have to call runModuleUnitTests() explicitely, e.g.
try
{
Runtime.initialize();
if (runModuleUnittests())
result = myWinMain(hInstance, hPrevInstance, lpCmdLine,
nCmdShow);
Runtime.terminate();
}...
Why call function rise exception?
The problem with 31 string. http://www.everfall.com/paste/id.php?jgsdz7mdbrnm If I uncomment it, and server response is return 404 code I code at runtime throw exception with text: Can't parse config: HTTP request returned status code 404 I thought that checkLinkCode is throw any king exception, and tried to wrap it's in try-catch block. But without result. So look like it's ok. I can't understand in which moment and why code is throwing to those exception.
Re: Why call function rise exception?
Oh sorry I understood where problem.
Re: dco how to specify Jpath?
On Thursday, 20 November 2014 at 10:48:17 UTC, Suliman wrote: I am playing with dco. And it's look very helpful for tiny projects. I can't understand is it's possible to add to dco.ini Jpath? I am talking about something like: dflags=-JD:\code\d\App1\source\ but when I am trying to compile code with dco it's can't find import. I looked at source code and it's look like it's do not support dflags... https://github.com/FrankLIKE/dco/tree/master/source Ok,I will add it. Thank you.
Re: dco how to specify Jpath?
On Thursday, 20 November 2014 at 10:48:17 UTC, Suliman wrote: I am playing with dco. And it's look very helpful for tiny projects. I can't understand is it's possible to add to dco.ini Jpath? I am talking about something like: dflags=-JD:\code\d\App1\source\ but when I am trying to compile code with dco it's can't find import. I looked at source code and it's look like it's do not support dflags... https://github.com/FrankLIKE/dco/tree/master/source -Jpath where to look for files for ImportExpressions. This switch is required in order to use ImportExpressions. path is a ; separated list of paths. Multiple -J's can be used, and the paths are searched in the same order. ImportExpression: import ( AssignExpression ) But -Ipath -ID:\code\d\App1\source\ It's ok. So dco now can do it.
Re: dco how to specify Jpath?
On Thursday, 20 November 2014 at 10:48:17 UTC, Suliman wrote: I am playing with dco. And it's look very helpful for tiny projects. I can't understand is it's possible to add to dco.ini Jpath? I am talking about something like: dflags=-JD:\code\d\App1\source\ but when I am trying to compile code with dco it's can't find import. I looked at source code and it's look like it's do not support dflags... https://github.com/FrankLIKE/dco/tree/master/source Default importPath is importPath=-I$(DMDInstallDir)windows/import You can modify it to your source. or add your source to the end. to do it: importPath=-I$(DMDInstallDir)windows/import -ID:\code\d\App1\source\
Re: dco how to specify Jpath?
On Thursday, 20 November 2014 at 10:48:17 UTC, Suliman wrote: I am playing with dco. And it's look very helpful for tiny projects. I can't understand is it's possible to add to dco.ini Jpath? I am talking about something like: dflags=-JD:\code\d\App1\source\ but when I am trying to compile code with dco it's can't find import. I looked at source code and it's look like it's do not support dflags... https://github.com/FrankLIKE/dco/tree/master/source In dco.ini importPath= -I..\\source // modity it to your source
