Saw a tweet where someone put rumble on text printing... seems niche, but would be nice to support some way for people to do this.

Best I can think is to have custom self-closing tags (and maybe not self closing?) that just call a function.

I find the two methods PrintCompleted and CharacterPrinted to be a little ambiguous. I get the latter, but the meaning of PrintCompleted only makes sense when you know it can't be the event for character printed since that would be redundant.

Suggestions - rename "PrintCompleted" to "TypewritingCompleted" or "TextCompleted"

Repro:
The following is observable in master right now by running the test scene and skipping + advancing.

1. Print text to a TMProUGui component that includes sprites. Ex: "This is a sprite !"
2. Change the text to something else.
3. Call UpdateGeometry on each MeshInfo. Ex:

        var textInfo = this.TextComponent.textInfo;
        for (int i = 0; i < textInfo.meshInfo.Length; i++)
        {
            textInfo.meshInfo[i].mesh.vertices = textInfo.meshInfo[i].vertices;
            this.TextComponent.UpdateGeometry(textInfo.meshInfo[i].mesh, i);
        }

The issue is sort of with TMPro. As an optimization they leave the SubMesh objects around that are used to render the sprites. They use the same SubMesh objects for fallback fonts. Then, in TextAnimation, calling TextComponent.UpdateGeometry(this.textInfo.meshInfo[i].mesh, i); causes the garbage sprite mesh info to be resupplied to the SubMesh.

We should be able to use TextInfo.ClearMeshInfo() before every print to fix.

There could be areas we are generating needless garbage. Especially consider:

• String concatenation
• Animating text mesh
• any case of "new"

Using a sprite tag causes the text to be truncated (i.e. not fully typed out).

It looks like this is because the sprites are not included in the totalPrintedChars count, which is derived from the length of the text with all tags removed. So each time you use a sprite tag your text will end up truncated by 1 character.

The text component this prints to must currently use rich tags because we hide characters using a color tag.

At bare minimum we should error if they don't have RichText checked. As far as supporting it, the best thing I can think is to implement the naive solution of printing each character, if they don't have RichText enabled.

It also seems to skip an incrementation of the number of letters that should be displayed when trying to use <\br> as an alternative.
(Meaning "<\br>" is taking up the slot of a character that is being displayed)

The TextTyper has a couple bad smells that I'd like to get rid of:

• It modifies the iterator (i) when applying tags in the loop
• It modifies the displayedText member as it loops. This is critical for keeping track of what's been printed, but it can be cleaned up.

Instead I think we should add a function, SplitForTyping that would take a string and an index. It would return a struct that contains a string for the display text, a string for the hidden text, and a list of outstanding tags (which need to be applied but only if they haven't already been applied). Every loop the TextTyper would simply call this function, and act on the data accordingly.

It might not even be a function, but rather a complete class, and it should be unit tested.

Release UnityPackages typically don't include tests. Therefore, since we will want to include the example scene and sound effects I should create an Examples folder for these.

These will all need to be modifiable at runtime, so that delay could be a user setting. Also punctuation characters would need to change based on the language.

One character per frame is simply too slow for anything heavy on dialogue.
Sorry if this has been addressed already. I don't understand the majority of the other tickets.

those UTF-32 characters ( like U+1F340) are stored in 4-bytes sequences, starting with 0xD8xx ... thus in one moment Unity Text components receives only a first-half of the UTF-32 character, which leads to crash like this:

ArgumentException: invalid utf-16 sequence at -1819518832 (missing surrogate tail) Parameter name: string

UnityEngine.TextGenerator.Populate_Internal (System.String str, UnityEngine.Font font, Color color, Int32 fontSize, Single scaleFactor, Single lineSpacing, FontStyle style, Boolean richText, Boolean resizeTextForBestFit, Int32 resizeTextMinSize, Int32 resizeTextMaxSize, Int32 verticalOverFlow, Int32 horizontalOverflow, Boolean updateBounds, TextAnchor anchor, Single extentsX, Single extentsY, Single pivotX, Single pivotY, Boolean generateOutOfBounds, Boolean alignByGeometry, System.UInt32& error)
UnityEngine.TextGenerator.Populate_Internal (System.String str, UnityEngine.Font font, Color color, Int32 fontSize, Single scaleFactor, Single lineSpacing, FontStyle style, Boolean richText, Boolean resizeTextForBestFit, Int32 resizeTextMinSize, Int32 resizeTextMaxSize, VerticalWrapMode verticalOverFlow, HorizontalWrapMode horizontalOverflow, Boolean updateBounds, TextAnchor anchor, Vector2 extents, Vector2 pivot, Boolean generateOutOfBounds, Boolean alignByGeometry, UnityEngine.TextGenerationError& error)
UnityEngine.TextGenerator.PopulateAlways (System.String str, TextGenerationSettings settings)
UnityEngine.TextGenerator.PopulateWithError (System.String str, TextGenerationSettings settings)
UnityEngine.TextGenerator.Populate (System.String str, TextGenerationSettings settings)
UnityEngine.TextGenerator.GetPreferredWidth (System.String str, TextGenerationSettings settings)
UnityEngine.UI.Text.get_preferredWidth ()
UnityEngine.UI.LayoutUtility.m__2 (ILayoutElement e)
UnityEngine.UI.LayoutUtility.GetLayoutProperty (UnityEngine.RectTransform rect, System.Func2 property, Single defaultValue, ILayoutElement& source) UnityEngine.UI.LayoutUtility.GetLayoutProperty (UnityEngine.RectTransform rect, System.Func2 property, Single defaultValue)
UnityEngine.UI.LayoutUtility.GetPreferredWidth (UnityEngine.RectTransform rect)
UnityEngine.UI.LayoutUtility.GetPreferredSize (UnityEngine.RectTransform rect, Int32 axis)
UnityEngine.UI.HorizontalOrVerticalLayoutGroup.GetChildSizes (UnityEngine.RectTransform child, Int32 axis, Boolean controlSize, Boolean childForceExpand, System.Single& min, System.Single& preferred, System.Single& flexible)
UnityEngine.UI.HorizontalOrVerticalLayoutGroup.CalcAlongAxis (Int32 axis, Boolean isVertical)
UnityEngine.UI.VerticalLayoutGroup.CalculateLayoutInputHorizontal ()
UnityEngine.UI.LayoutRebuilder.m__2 (UnityEngine.Component e)
UnityEngine.UI.LayoutRebuilder.PerformLayoutCalculation (UnityEngine.RectTransform rect, UnityEngine.Events.UnityAction`1 action)
UnityEngine.UI.LayoutRebuilder.Rebuild (CanvasUpdate executing)
UnityEngine.UI.CanvasUpdateRegistry.PerformUpdate ()
UnityEngine.Canvas:SendWillRenderCanvases()

I've just downloaded the files as a .zip and dragged them into my unity project. Upon trying to add the TextTyper.cs script to my TextMeshPro object, I receive a pop up window with the error stated above.

Style tags are not considered the "Built-in" type, so they end up getting included in the calculations for visible text, even though they get stripped. This causes text to be truncated and index out of bounds errors.

To fix we should include "Style" as part of the Unity tags in TextTagParser.

See Pull Request #14

It's a simple fix, really, but the challenge is how to structure it so that user's aren't getting errors if they don't have TMP in the project.

The Readme could use some help for how you expect people to use TextTyper. I think more could be done to make it "scripting free" but right now even the scripting could use a little API overview.

Some suggestions for how to achieve it:

• How do you install this thing? There's no package linked. If you want people to just download the code, at least mention that.
• Snippet of all code needed to get something Typing
• A list of "other methods" we've exposed that people can hook into - don't need to go into detail - this isn't an API Reference exactly

As I user of a tool like this, I want to get up and running as fast as possible. I might not want to use it once I see it in action so let me get to that point as fast as I can. I'll dive in for more features later. I definitely don't want to email the owner and ask "How do you intend for people to use this tool?"

It would be great if I could use "Anim=Fullshake" tags as part of the Style attributes in a TextMeshPro style sheet.

Rich Text Tags support quotes around some parameters (ie. <color="red">).

As a user, I want to be able to optionally use quotes around my parameter, (ie <anim="shake">).

I'm using this component for tutorial dialogue. There is a case where I use it in a pause screen where Time.timeScale = 0. This component doesn't work there.
I tried changing this line:
yield return new WaitForSeconds(this.characterPrintDelays[currPrintedChars - 1]);
to
yield return new WaitForSecondsRealtime(this.characterPrintDelays[currPrintedChars - 1]);
but it leads to an exception related to the array characterPrintDelays.
Can you please add the option? Like a bool field to toggle using scaled time or not

"OutstandingTags" aren't cleared when you type to a TextTyper, so if you print to a TextTyper while it is in the middle of printing text inside a tag, it will keep an unpaired tag in the OutstandingTags list forever.

Right now it's hard to share animations between fonts of different sizes. This is because a shake that is extreme on a small font is minimal on a bigger font.

I think using a relative strength might be most correct by default, but alternatively we could add a flag to each animation which determines if it will use absolute positions or positions relative to the size of the character.

It would be helpful to know if text is actively being typed. For example I use a queue system for successive messages but don't want to auto-dequeue if typing is in progress.

I personally just added public bool IsTyping { get; private set; } to the TextTyper and toggle it from appropriate locations.

Anyway, thanks for this lib. I was looking for something that handles colors and more critically, wrapping. So far this works well!

When restarting a dialogue queue, the Skip() function seems to jump dialogue lines and fails after the first run through. Using the TextTyperTester example, I changed Start() to a function StartTyping() and have a button to call that function. After the dialogue finishes when using the "skip or advance" button, replaying the dialogue causes a bug on that button.

The text generator currently moves an alpha tag to change what characters are visible, which is not optimal for garbage collection. A solution to this is to use the following code in place of the function that changes the location of the alpha tag, but some changes would likely need to be made since this is mostly unmodified from the system that I use.
// textField is the text object that the player sees, and is also where the unformatted text starts // I don't remember what tempString does, I just know that you need it // I also don't remember what start or end index do , but again, they are very important public void InitializeText() { textField.color = Color.clear; textField.text = text; Canvas.ForceUpdateCanvases(); for (int i = 0; i < textField.cachedTextGenerator.lines.Count; i++) { int startIndex = textField.cachedTextGenerator.lines[i].startCharIdx; int endIndex = (i == textField.cachedTextGenerator.lines.Count - 1) ? textField.text.Length : textField.cachedTextGenerator.lines[i + 1].startCharIdx; int length = endIndex - startIndex; tempString += textField.text.Substring(startIndex, length) + "\n"; } textField.text = string.Empty; textField.color = Color.white; // Replace with whatever color you need text = tempString; }
I'm sorry I don't know as much as I should about a script that I wrote, but I wrote this a very long time ago, and I never used to put comments in my code (I have since learned my lesson), so I hope you can make at least some sense of this. Despite all of these problems with it, I hope you get at least some use out of this script even though it was very poorly done, and i will probably need to redo it at some point

When having a tag, the script assumes it requires a closing and adds it to the final text that's being printed. Also for some reason the sprite appears as soon as the typing appears instead of being "typed".

We could break the symbol list into two lists - one for visible and one for hidden text. Instead of doing operations on text, we should do operations on the symbol lists.

For example: RemoveFirstOccurance(string, text) uses a string operation to take out tags. We could instead just remove it from the symbol list.

So we build the two strings entirely from symbol lists, then just do a ToString() on each SymbolList.

SymbolList should probably be made into a class.

This is all just to make the code a bit cleaner.

Sometimes the text being typed is larger than the target area. Even when controlling for this in the content, translation/localization passes may have greatly extended text lengths.

It would be ideal for unity-text-typer to support the paged overflow mode in TMP, with something like a "Page Complete" event and a way to resume typing the next page.

https://docs.unity3d.com/Packages/[email protected]/api/TMPro.TMP_PageInfo.html

Hi, can someone build an asmdef for the project? Thanks.

We need to strip all native tags when parsing the text, or else it will calculate the wrong number of characters to print and results in an ArgumentException.

Repro:

1. Add a <style> tag to text
2. Print it
3. Observe ArgumentException

To fix, we either need to add all missing "Unity" tags to the parser, or calculate characters to print based on the visible text. For now the easiest fix is to include all missing tags. Ex:

style
link

Another issue that could come up is you can use color like "<#FFFFFF>" instead of <color=#FFFF>. If that results in this same issue, we may want to look at solution that examines the visible text.

Users shouldn't have to modify core code to change the framerate of the animations.

We should either expose the framerate as a setting, or make it a setting per Animation. I'd vote the latter if it's possible.

We need to go through all issues and verify they are still issues with the new TextTyper.

When there is no Shake or Curve Library wired up on the TextTyper, it will not print anything.

Repro steps:

1. Add a TMPRo component with a TextTyper component
2. Clear out any animation libraries
3. Print text that has an Anim tag.
4. Observe nothing is printed

I expect it to error, and print the text as normal without the animation.

Documentation is a bit dated. We need to go through the existing documentation to see what could be improved or what is no longer relevant.

When having centered and bolded text, while it's being typed it will slightly move left each time a character appears.

If we add arguments passed out from events, it will break everyone's code. Mostly an easy fix, but still annoying.

To fix this we can use EventArgs https://docs.microsoft.com/en-us/dotnet/standard/events/

Delay as a fixed amount is a bit problematic if users change default print speed. This tag could support a multiplier like <delay=1.5x> which would multiply their default print speed.

The text typer doesn't reset the text speed on a new line. The bug occurs when you skip dialogue between delay tags. I fixed this temporarily by prefacing each new line with its own delay:

scripts.Enqueue("<delay=0></delay>" + conversationText);