Add Clone() implementation for StringMap

[Headline]

We implemented a clone for ArrayList, but ArrayStack and StringMap were left without one. This pull request adds deep-copy cloning to said structures. Construct generally looks like

ArrayStack mystack = // something, lets assume it also has data
ArrayStack copy = mystack.Clone();

delete mystack;
delete copy;

Fixes #796

[asherkin]

This pull request adds deep-copy cloning to said structures.

The implementation is a shallow copy (which I think is all we can sanely do and matches ArrayList), was something else intended?

[Headline]

@asherkin unless i'm mistaken, these are deep copies. Any modifications to one structure will leave the other completely unaffected. Trie stuff like Entry::setString allocs input string len+1 and strcopy's the data, for example.

CCellArray::clone() also looks like a deep copy

sourcemod/core/logic/CellArray.h

Lines 183 to 197 in 3baa703

	ICellArray *clone()
	{
	CellArray *array = new CellArray(m_BlockSize);
	array->m_AllocSize = m_AllocSize;
	array->m_Size = m_Size;
	array->m_Data = (cell_t *)malloc(sizeof(cell_t) * m_BlockSize * m_AllocSize);
	if (!array->m_Data)
	{
	delete array;
	return NULL;
	}
	memcpy(array->m_Data, m_Data, sizeof(cell_t) * m_BlockSize * m_Size);
	return array;
	}

[KyleSanderson]

Is this actively desired? It seems like a thing that could be used but giving some lineage on the drive behind a PR helps better understand intentions.

[Headline]

@KyleSanderson I’m unsure if ArrayStack’s clone is used (or if people use ArrayStack at all), but I think being able to clone a StringMap is a really good addition into core. I can’t double check, but I don’t think it’s possible to distinguish a key’s mapped type so this couldn’t even be done native in sp, so a clone couldn’t be accomplished prior to this pr nevermind getters fail.

[KyleSanderson]

Stack is used by many; cloning a stack I'm not sure on as it likely predates most of us. The trick is to jump on BZ (or now here) and looking at why something was added. Nice-to-haves are legit, just wanted some lineage.

All good.

[Headline]

Ah, my bad. All should be well now.

[Headline]

For ref/lineage: simple tests seem to pass as-expected.

public void OnPluginStart()
{
	StringMap original = new StringMap();
	original.SetValue("bool", true);
	original.SetValue("handle", new ArrayList());
	original.SetString("string", "hello world");
	
	StringMap copy = original.Clone();
	delete original;
	
	bool ret = false;
	bool value;
	ret = copy.GetValue("bool", value);
	PrintToServer("bool test: %s", (ret && value)?"successful":"unsuccessful");
	
	ArrayList handle;
	ret = copy.GetValue("handle", handle);
	PrintToServer("hndl test: %s", (ret && handle != null)?"successful":"unsuccessful");

	char string[32];
	ret = copy.GetString("string", string, sizeof string);
	PrintToServer("string test: %s", (ret && strlen(string) > 0)?"successful":"unsuccessful");
	
	delete copy;
	delete handle;
}

[KyleSanderson]

That's it, mention the PR # like every other line item cause and you're probably golden.

Passing onto a peer.

[peace-maker]

@Headline now that #1304 was merged, this could be reduced to only adding a StringMap.Clone()?