Re: Bug in std.json or my problem

2020-04-22 Thread Craig Dillabaugh via Digitalmars-d-learn
On Wednesday, 22 April 2020 at 18:35:49 UTC, CraigDillabaugh 
wrote:

On Wednesday, 22 April 2020 at 18:23:48 UTC, Anonymouse wrote:
On Wednesday, 22 April 2020 at 17:48:18 UTC, Craig Dillabaugh 
wrote:

clip


File an issue if you have the time, maybe it will get 
attention. Unreported bugs can only be fixed by accident.


I thought it might be worth filing a bug, but wanted to confirm 
if others thought this was actually a bug.  I had encountered 
an identical issue with vibe-d years ago.


Thanks for the feedback.


After some digging it appears there is already a fix for this 
(though a bit old) … maybe I am the only person who cares about 
std.json after all :o)


https://github.com/dlang/phobos/pull/5005/commits/e7d8fb83d2510b252cd8cfd2b744310de6fa84e5



Re: Bug in std.json or my problem

2020-04-22 Thread CraigDillabaugh via Digitalmars-d-learn

On Wednesday, 22 April 2020 at 18:23:48 UTC, Anonymouse wrote:
On Wednesday, 22 April 2020 at 17:48:18 UTC, Craig Dillabaugh 
wrote:
The crash is caused because the 'income' field with value 0.0 
is
output as 0 (rather than 0.0) and when it is read this is 
interpreted

as an integer.

Shouldn't this work?


Yes, it's just buggy.

Giving it a value of an even 1.0 will make it throw the same 
exception (js["income"].type is JSONType.integer), but a value 
of 1.1 will make it pass (.type properly becomes 
JSONType.float_).


I don't know of a solution other than to check 
js["income"].type beforehand and use .floating or .integer as 
befits, or simply use an alternative JSON library (asdf?).


File an issue if you have the time, maybe it will get 
attention. Unreported bugs can only be fixed by accident.


I thought it might be worth filing a bug, but wanted to confirm 
if others thought this was actually a bug.  I had encountered an 
identical issue with vibe-d years ago.


Thanks for the feedback.


Re: Bug in std.json or my problem

2020-04-22 Thread Anonymouse via Digitalmars-d-learn
On Wednesday, 22 April 2020 at 17:48:18 UTC, Craig Dillabaugh 
wrote:

The crash is caused because the 'income' field with value 0.0 is
output as 0 (rather than 0.0) and when it is read this is 
interpreted

as an integer.

Shouldn't this work?


Yes, it's just buggy.

Giving it a value of an even 1.0 will make it throw the same 
exception (js["income"].type is JSONType.integer), but a value of 
1.1 will make it pass (.type properly becomes JSONType.float_).


I don't know of a solution other than to check js["income"].type 
beforehand and use .floating or .integer as befits, or simply use 
an alternative JSON library (asdf?).


File an issue if you have the time, maybe it will get attention. 
Unreported bugs can only be fixed by accident.


Bug in std.json or my problem

2020-04-22 Thread Craig Dillabaugh via Digitalmars-d-learn
So perhaps I am the only person in the world using std.json, but 
I was wondering

if the following code should work.

=
import std.json;
import std.conv;
import std.stdio;

struct Person {
string name;
float income;

this (string name, float income) {
this.name = name;
this.income = income;
}

this(JSONValue js) {
this.name = to!string(js["name"]);
/* This next line crashes with .. JSONValue is not 
floating type.

 *  to!float( js["income"].toString()) works.
 */
this.income = js["income"].floating;
}

JSONValue toJSON() {
JSONValue json;
json["name"] = JSONValue(this.name);
json["income"] = JSONValue(this.income);
return json;
}
}


int main(string[] argv) {
Person bob = Person("Bob", 0.0);

string bob_json = bob.toJSON().toString();

Person sonofbob = Person(parseJSON(bob_json));

writeln(sonofbob.toJSON().toPrettyString());

return 0;
}
===

The crash is caused because the 'income' field with value 0.0 is
output as 0 (rather than 0.0) and when it is read this is 
interpreted

as an integer.

Shouldn't this work?