LÖVE

I want my players to name things and save them to a file. I've got saving and loading working, and naming works somewhat. I have the following:

if(key == "backspace" or key == "delete") then
if(string.len(tempString) > 0) then
tempString = string.sub(tempString,1,string.len(tempString) - 1)
elseif(string.len(tempString) < 12) then
tempString = tempString..key
else
end

This works pretty well, but if players press lshift, or tab, or some non-character key they get "lshift" (or whatever) entered into the text box. Obviously this is suboptimal.

Is there a better way to have players enter strings? Is there some functionality I'm not applying? Or is there an easy filter for "that key is a character" which I'm not using?

--Mr. Strange

I'm assuming you're doing this in the love.keypressed() callback?

In the most recent versions of LÖVE, there is a second argument that gets passed to love.keypressed - love.keypressed(k, u), where u is the Unicode codepoint of the key that was pressed. You can then use Lua's built-in function string.char(b) to produce the printable version of that key. So it'd look something like the following: This should prevent lshift and its brethren from showing up, though as a disclaimer, I haven't tested the above recently.

Hmmm... doesn't really work. For some characters (like tab) I get a box, while others (lshift, ctrl) simply seem to provide a null character, which love then gets confused about, and prints nothing after. I can still delete characters (as per the first part of the if) and get back to a good state, but it's not what I want.

Hmmm - maybe I should add a check like this:

if(string.len(string.char(key)) > 1) then
-- don't append the character
else
-- do append.
end

That will probably work, though it's not especially elegant.

Indeed, the following code in keypressed(key, unicode) works fine. Using the unicode gives me capitals and other good characters, so that's a nice improvement.

The null characters are easy, add this:

no u

I check the length of the key like you have mentioned. I think that's quite nice enough. Oh, and I use a look-up table to handle the shifted characters. You could see my code in the programme posted here: [RLC]Data Execution Prevention?.

I hope that's helpful.

rhezalouis wrote:Oh, and I use a look-up table to handle the shifted characters.

That's not necessary anymore -- we have the unicode argument now, which works for every key and (IIRC) for different keyboard layouts as well.

Robin wrote:

rhezalouis wrote:Oh, and I use a look-up table to handle the shifted characters.

That's not necessary anymore -- we have the unicode argument now, which works for every key and (IIRC) for different keyboard layouts as well.

Ow, sorry, I am still working in 0.6.0 and haven't read the new features. Waa, I really have to catch up.

rhezalouis wrote:Ow, sorry, I am still working in 0.6.0 and haven't read the new features. Waa, I really have to catch up.

The unicode argument was added in 0.6.0, so unless you meant you are still working in 0.5.0...