rookfighter / inifile-cpp Goto Github PK

Add a script that installs dependencies locally.

Change structure to be header only

Add static default fields to IniFile class

• defaultSeparator
• defaultComment
• defaultSectionOpen
• defaultSectionClose

I think that INI files with duplicated sections are confusing and error prone. For example:

[Foo]
var1 = hi
[Bar]
var2 = 12
[Foo]
var3 = 14

[Foo] will end up having both variables var1 and var3.
It is easy to spot this duplication in such a small example, but with bigger INI files this can be problematic.

I've made the pull request #14 to make the code throw an exception in this case.

Add option to change the char for comments.

@Rookfighter Thank you for this useful library.

I noticed that after updating the values and calling save option, all the comments and empty lines in the ini file gets removed.
Since most of these files are edited manually by users, retaining comments and white spaces are essential while saving.

Sample ini file

# Send a message to the world
[world]
message=Hello

# Just foo :)
[foo]
f1=123

Sample code

#include <inicpp.h>
int main()
{
    ini::IniFile myIni;
    myIni.load("some/ini/path");

    myIni["world"]["message"] = "Hello world";

    myIni.save("some/ini/path");
    
}

Result after saving the file

[world]
message=Hello world
[foo]
f1=123

Do let me know if there is already an option to avoid them being removed.

Many INI files are edited directly by end users, and some of them want to use spaces to make the contents easy to read.

For example:

[SECT1]
  #This comment is indented, as is the following lines, which are also aligned
  John   = 1
  Evelyn = 2
  Lucas  = 3

I've made the pull request #11 to implement that.

How to disable ALPHABET sorting in output file?

It always broke my file structure...

For example...

[General]
NUM_AITEMS = 5
AITEMS_ENABLED = false
[AITEM1]
DATA = X

After save it make fully unreadabled:

[AITEM1]
DATA = X
[General]
AITEMS_ENABLED = false
NUM_AITEMS = 5

Add helper functions

• hasSection
• hasField

I have a ini file:

[/home/xxx/.gnupg/]
Acls=user::rwx;group::---;other::---

[/home/xxx/.gnupg/.#lkxxx.29271]
Acls=user::rw-;group::r--;other::r--

[/home/xxx/.gnupg/openpgp-revocs.d/xxx.rev]
Acls=user::rw-;group::---;other::---

[/home/xxx/.gnupg/private-keys-v1.d/xxx.key]
Acls=user::rw-;group::---;other::---

[/home/xxx/.gnupg/private-keys-v1.d/xxx.key]
Acls=user::rw-;group::---;other::---

[/home/xxx/.gnupg/pubring.kbx]
Acls=user::rw-;group::r--;other::r--

[/home/xxx/.gnupg/pubring.kbx~]
Acls=user::rw-;group::---;other::---

[/home/xxx/.gnupg/trustdb.gpg]
Acls=user::rw-;group::---;other::---

[/home/xxx/.ssh]
Acls=user::rwx;group::r--;other::r--
[/home/xxx/.ssh/id_rsa]
Acls=user::rw-;group::---;other::---
[/home/xxx/.ssh/id_rsa.pem]
Acls=user::rw-;group::r--;other::r--
[/home/xxx/.ssh/id_rsa.pub]
Acls=user::rw-;group::r--;other::r--
[/home/xxx/.ssh/known_hosts]
Acls=user::rw-;group::---;other::---
[/home/xxx/.ssh/known_hosts.old]
Acls=user::rw-;group::r--;other::r--
[Defaults]
Excludes=
Includes=/home/xxx/.ssh;/home/xxx/.gnupg/;
SaveAcls=true

Then I get a error when I rerun my programe:

erminate called after throwing an instance of 'std::logic_error'
what(): l4: ini parsing failed, section not closed

But I can't find any error about this ini file :(

Allow ini files without sections, e.g.

foo=bar
hello=world

Add the option to change the file separator '='.

Most implementations consider the semicolor ";" as the start of a comment

I believe this #22 should be enough?

Hi, @Rookfighter. thank you for a great tool.

I'm trying to read an example.ini. And if the group-key-pair not exist in the ini-file, it returns "" that by converting to int gives 0 and by converting to bool gives false. It is in my opinion an unexpected behavior.

example.ini

[Group_1]
key_1 = 1;

main.cpp

#include <iostream>
#include "inicpp.h"
int main() {
	 ini::IniFile inifile;
	 inifile.load("example.ini");

	 int val_1 = inifile["Group_1"]["key_1"].as<int>();
	 int val_2 = inifile["Group_1"]["key_2"].as<int>();
	 
	 std::cout << "value 1 = " << val_1 << std::endl;
	 std::cout << "value 2 = " << val_2 << std::endl;
	 return 0;
}

Output:

value 1 = 1
value 2 = 0

My workaround for now is

#include <iostream>
#include <stdexcept>
#include "inicpp.h"

int readInt(ini::IniFile &inifile, const std::string &group, const std::string &key) {
	if (inifile[group][key].as<std::string>() == "")
		throw std::runtime_error(key + " is not found in " + group + "\n");
	else
		return inifile[group][key].as<int>();
}

int main() {
	ini::IniFile inifile;
	inifile.load("example.ini");

	try
	{
		int val_1 = readInt(inifile, "Group_1", "key_1");
		std::cout << "value 1 = " << val_1 << std::endl;
	} catch (...)
	{
		std::cout << "key_1 is not found in Group_1" << std::endl;
		inifile.clear();
		return 1;
	}

	try
	{
		int val_2 = readInt(inifile, "Group_1", "key_2");
		std::cout << "value 2 = " << val_2 << std::endl;
	} catch (...)
	{
		std::cout << "key_2 is not found in Group_1" << std::endl;
		inifile.clear();
		return 1;
	}

	return 0;
}

Hello, my library figcone depends on inifile-cpp, which is downloaded during CMake configuration using FetchContent. As inifile-cpp doesn't provide any version tags, I specified the master branch to download it. Recently (?) the master branch was removed, and it rendered all versions of figcone broken unless the user disables the support of the ini format. By default, all config formats are supported, so the build is failing during configuration while trying to download inifile-cpp.

It will teach me to only use forked repositories for projects that don't have any versioning, but can you please restore the master branch to fix all previous releases of my library? I'm not asking to make it the main branch again, just to create it in the repo. There were no incompatible changes, so it can be set on the latest commit.

Add type uint and float for Ini fields.

Вы можете использовать такой код (inicpp.h:234)?
void decode(const std::string &value, double &result) {
std::stringstream stream;
stream << value;
stream >> result;
}

Program crashes if invalid input found.

Need "bypass errors" option :)