Tag Archives: enum classes

SDL2

[ SDL2 – Part 11 ] Text styling

4 Comments

Text styles using SDL2_ttf

In the last post we looked at how to render text. Now let’s take this a step further and change the appearance of the font. There are two ways you can change how the font looks. Font style and font outline.

Fon styles

SDL TTF allows you to set the style of the font. It supports the following font styles

  • bold
  • italic
  • underlined
  • strikthrough

You can combine these in any way you like. We’ll start off with just setting a single font style at a time, and then move on to see how we can apply several of them at once.

Font styles

Settnig font styles in TTF is easy, it just requires a single function. The function let us you set one or more font styles. Let’s start off by looking at how to set just one font style

Setting the font style

We can set the font style using the following function

Arguments :

  •  TTF_Font *font – the font to set the style on
  • int style       – the style to set on the font

As you can see, the style parameter is an int and not an enum. I’ll get back to why that is later, but for now let’s look at the possible values for style, these are all self-explanatory so I won’t be adding a description.

  • TTF_STYLE_NORMAL
  • TTF_STYLE_BOLD
  • TTF_STYLE_ITALIC
  • TTF_STYLE_UNDERLINE
  • TTF_STYLE_STRIKETRHOUGH

Any text you render after setting this font style will have the new effect, but it won’t change any text you have written with a different style. So when you set the style to TTF_STYLE_BOLD, all text you render from that point and until you set a different style will be bold. And as long as you pass any of the above values to the function, the font will only have the one last style you set.

Let’s do a simple example

Any text rendered at this point will be normal with no font styles

Any text rendered at this point will be bold

Any text rendered at this point will be in italics, but not bold

Any text rendered at this point will be normal with no font styles

Any text rendered at this point will be underlined

As you can see, this is pretty straight forwards. So let’s make things a little bit trickier by setting multiple font styles at once. To do this, we must first look a bit at the binary number system

Binary numbers

In order to learn about how to combine these flags, we need to look at binary numbers first of all. If you don’t already know about binary numbers, you should take a look at the above link. It’s not crucial, but it is highly recommended to know a little about them. I might create a blog post about them at some point. For now, I’ll just talk a tiny bit about the binary number system. But as I said, I highly recommend understanding it fully to the point where you can convert back and forth between binary and decimal numbers

The binary number system

On a daily basis, we use the decimal number system. The binary numbers system is just a different way of representing numbers. Any numbers can be converted from any other number system. So you can convert binary numbers to decimal numbers ( and the other way around ).

A computer stores numbers as individual bits ( 0‘s and 1‘s ). They correspond to on / off or true / false.

Let’s take a look at an 8 bit binary number ( 1 byte )

1010 0101

As you can see, it has 8 digits. So that’s eight different flags. Each of these flags have two different possible values : 0 / 1 or false / true. So that’s 8 bools for the price of a single byte!

Bitwise operations

So how do we use these 8 booleans? As you know, we have the following boolean operations in C++:

  • and ( && )
  • or ( || ) 

These work on an entire variable. An int, for instance will be false if its value is 0, otherwise its true.

But there are similar operations that does this on all bits of a variable. These are called bitwise operations, simply because they operate on a simple byte. To do a bitwise operation, we need two variables of equal size ( same number of digits ), for instance two bytes. The result of a bitwise operation is a third variable of the same size. So if we do a bitwise operation between two bytes, we get a third byte back as a result.

Let’s create two bytes, we’ll use these for a few examples

Byte 1 : 0101 0011 ( 64 + 16 + 2 + 1 = 85 )
Byte 2 : 0110 0010 ( 32 + 64 + 2 = 98 )

We’ll be referring to each digit as a position. So the digits in the first position is 0 in both or bytes. In the second position it’s 1 in both bytes and in the third it’s 0 in the first byte and 1 in the second byte.

Bitwise OR

A bitwise OR operation means we look at all positions as check if either of them is 1 if so, we set the digit in that position to 0. If no digit in that position is 1, we set it to 0

The operator for bitwise OR in C++ is | ( just one |, not two )

Here is a simple example of bitwise OR between two bytes

0101 0011
OR
0110 0010
=
0111 0011

Bitwise AND

In a bitwise AND operation, we look at each position and see if both of them are 1. If so, we set the digit in that position to 1, otherwise we set it to 0. So in OR we set it to 1 if any of the two is 1, here we only set it to 1 if both are 1.

The operator for bitwise AND in C++ is & ( just one &, not two )

Here’s the a simple example :

0101 0011
AND
0110 0010
=
0100 0010

Bitwise XOR

XOR or exclusive OR is slightly less known than OR and AND. In an XOR operation, we check if the two values are different. So this is equivalent to != in C++.

  • ( true  != false ) = true
  • ( true  != true  ) = false
  • ( false != true  ) = true
  • ( false != false ) = false

Simply put, an XOR operation is true if the two parts are different. So in a bitwise XOR operation, we look at each position and see if the two digits are different. If so we set the digit at that position to 1, otherwise we set it to 0.

The operator for bitwise XOR in C++ is !=

Here is an example :

0101 0011
XOR
0110 0010
=
0011 0001

Bitwise NOT

We also have the a bitwise version of the NOT opeartion this is done by using the ~ operator in C++. If we used ! we would get the result of NOT on the entire variable, not the individual bits which is what we want. This operation only takes a single element and flips all bits ( turns 1‘s into 0‘s and 0‘s into 1‘s. ). Let’s test it on our two bytes

The operator for bitwise NOT in C++ is !

Byte 1 :

NOT 0101 0011
=   1010 1100

Byte 2 :

NOT 0110 0010
=   1001 1101

Setting and checking individual bits

So now that we know how to do bitwise operations, we need a way of checking and setting the individual bits. This is done simply by using OR, AND and XOR. Before we take a look at how to do this, let’s define a few values to check.

Remember that the different font styles are ints? This is because they are used to perform bitwise operations to set and unset different bits. Here they are again, this time with their values. For simplicity, I’ll only list the last four bits ( the others are always 0 ). The values are in decimal with the binary representation in parenthesis

  • TTF_STYLE_NORMAL = 0 ( 0000 )
  • TTF_STYLE_BOLD = 1 ( 0001 )
  • TTF_STYLE_ITALIC = 2 ( 0010 )
  • TTF_STYLE_UNDERLINE = 4 ( 0100 )
  • TTF_STYLE_STRIKETRHOUGH = 8 ( 1000 )

As you can see, they all have only one ( or zero ) bit set. This means we can use AND, OR or XOR on just one bit.

Setting a bit

To set a bit ( without affect any other bit ) we use the ORoperation. So say that we have four bits set to 0, 0000 and we want to set the bit for bold on it ( 0001 ). In other words, we want the result 0001. What we do is : that we take our original 4 bits ( 0001 ) and set it to the original 4 bits ( 0001 ) OR‘ed with the bitmask for bold ( 0001 ) :

0000
OR
0001 ( value of TTF_STYLE_BOLD )
=
0001

Simple as that! This woks for any of the other flags in the same way. They all will end up setting one bit.

Note that this will not change any other bits. If we try set the italics font style on the above variable we get :

0001
OR
0010 ( value of TTF_STYLE_ITALIC )
=
0011 ( TTF_STYLE_BOLD and TTF_STYLE_ITALIC set )

Let’s make a simple function that adds a style to a mask.

Unsetting a bit

Sometimes we want to unset a bit. Say for instance we want to remove the italics from the font above. How do we do that without affection the other values? This is a bit more complex, because it requires two operations. What we are trying to do is the following :

Say we have a bitmask ( 0000 1011 ) and we want to unset the bit for bold text, but leave the rest unchanged. So we need to be able to go :

From 1011 to 1010

To do this, we need to use an AND operation. This is because we can’t turn of a bit using OR, and XOR would only flip it back and forth between 0 and 1

But we can’t use AND with the flag we want to unset alone, because that would keep the flag at 1 and change every other to 0!

0000 0101
AND
0000 0001
=
0000 0001

This is the opposite of what we want! Wait? Opposite you say? Well how about we use the NOT operator here to get the opposite result? This works perfectly, because NOT 0000 0001 is 1111 1110. And, as we saw earlier, doing AND 1 won’t change the element we’re AND‘ing with. So we get :

0000 0101
AND
1111 1110
=
0000 0100

Success! Only the bit we were trying to unset has changed. So let’s make a function that does this for us :

Checking a bit

To check a bit, we also need to use the bitwise AND operation. And since we are only checking and not setting, we don’t have to store the value anywhere which means we don’t have to worry about changing anything.

To check a bitmask, simply do an AND operation with the value you want to check for ( in this case, any of the TTF_STYLE_.... values ). So, to check if a text is bold, we do an AND between our mask and TTF_STYLE_BOLD :

0011 ( our bit mask, TTF_STYLE_BOLD and TTF_STYLE_ITALIC set )
AND
0001
=
0001

As you can see, we only check the bit that’s set in our variable ( TTF_STYLE_ITALIC set ) the others will be 0 no matter what our mask is. The value 0001 is not 0, and thus this evaluates to true and we now know that the font is bold.

If our mask didn’t have the bold bit set ( only the italic one ), our mask would be 0010. An AND between 0010 AND 0001 is false ( they have no bit set to 1 in common ) and the result is 0 aka false.

So let’s create a function for that too!

Conclusion

With a little knowledge about binary numbers and bitwise operations, we can easily set, add, remove and check various font styles in SDL_TTF.

Since it does involve a little low level code, I made a simple class that does the apparitions for us in a more intuitive way. I strongly suggest using this opposed to “raw” TTF_Fount*

Feel free to comment if you have anything to say or ask questions if anything is unclear. I always appreciate getting comments.

You can also email me : olevegard@headerphile.com
Share
C++11

[ C++11 – Part 1 ] Enum class

Leave a comment

Introduction to the series

I thought I’d start of by talking about my favorite thing that’s new about C++11 standard. C++11 has a lot of useful new features : auto, initializer lists, lambda, shared_ptr, unique_ptr, move semantics, etc. Some of these are quite comprehensive and complicated, and I’m still working on understanding all of them. But this one is actually quite simple.

About enum classes

Enum classes are basically enumerations that can’t implicitly be converted to integer types. So basically you can’t write :

Why is this such a good thing?

First of all it helps code clarity. A simple question : Say you have an enum of card suits, which would you prefer?

Using and enforcing enum classes means you’ll instantly know what the value is, so you’ll never have to find the actual definition of the enum and look up what the different int values does.

So it helps code clairty. Is that all?

No! There are more serious issues. Say you have the following code

This might look okay, they both set the card to Suit::Clover. But here comes the big issue. Someone changes the order of the elements in Suit.

What’s the issue?

It might seem a very subtle change. But you let’s say you implemented reset method a long time ago. It’s been working fine ever since, and you’ve completely forgotten that it uses and integer value to set suit. It’ll now do something different from the constructor! So you end up with cards being different when you reset them from when you create them. And issues like this can be very confusing and hard to debug.

This is just a minimal example to showcase the issue. But this can happen just as easily in huge projects where the enum is used hundreds of places. Debugging such issues will waste a huge amount of time and cause a whole lot of pain and agony. All because of using an integer to assign an enum! Using enum class prevents these issues.

Conclusion

Use enum classes! It might be a bit annoying having to write the full enumeration name sometimes. But it’s a lot better than spending hours upon hours debugging when someone changes the order of the enumeration items or adds a new value in front of others!

Feel free to comment if you have anything to say or ask questions if anything is unclear. I always appreciate getting comments.

You can also email me : olevegard@headerphile.com
Share