Some Chinese symbols cannot be displayed correctly

[jie326513988]

Hi, thanks for the library. It's great.
When using non isoline fonts, Chinese punctuation marks cannot be displayed correctly.
Are you still updating.
^_^

[k-omura]

Thank you for posting.
I'm doing maintenance to make all TTFs visible.
I don't fully understand non-isoline fonts. Can you tell me the name of the font that is not working? Also, if possible, could you post a picture of the failure? Thank you.

[jie326513988]

I have solved this problem by replacing other fonts. I hope that the font rendering speed can be improved in the future.
On the left is abnormal display, and on the right is normal display.
On the left is 文泉驿等宽正黑.ttf On the right is 华康金刚黑极细.ttf.

[k-omura]

thank you for teaching me. It was nice to be able to work with other fonts for the time being.
I will try 文泉驿等宽正黑.ttf. (although it will be slow)
I recognize that it is a problem that it takes time. It takes a lot of time to fill in the inside of the characters, but I don't have any ideas for effective countermeasures. I will continue to challenge myself.

This item will be closed when the 文泉驿等宽正黑.ttf error is resolved.

[jie326513988]

Thank you, cheer for you.
I also found that the newline character (0x0a) will cause the device to crash (esp8266).

I modified the code here, and it will work normally.

[jie326513988]

Hello, I optimize some places.

1. Fix the memory leak.
2. Fix the problem that ascii symbols will cause system crash.
3. Add routine linked with GXEPD2, which is applicable to ink screen and esp8266.
   However, there are still some Chinese fonts that will cause the system crash. The reason is unknown.
   This is the optimized address https://github.com/jie326513988/truetype_Arduino

[k-omura]

Hello.
thank you. I was aware of the memory leak problem, but I could not fix it.
I'm also aware that some font files have some issues, but I'm working on a way to work around that.
thank you for your cooperation

[jie326513988]

There is a BUG in the this->fillGlyph() function.
The reason is that the variable type of _x_min is incorrect. When the incoming parameter is a negative number, the execution time of the for loop will be very long. It should be changed to int16_t type, and the parameters of the for loop should also be modified together.

[k-omura]

Thank you for pointing out the bug. that's right.
I will fix this problem.
(I'm sorry I haven't gotten around to the original issue on this board yet.)

[jie326513988]

Regarding the problem that result.leftSideBearing value is too large, temporarily use the modified method to solve it.

In the same ttf file, the values of advanceWidth and leftSideBearing parsed by some characters will be extremely large, which are obviously abnormal values. If these two values ​​can be correctly parsed, perhaps the problem of most fonts not being displayed or displaying in abnormal positions can be solved.

[jie326513988]

When parsing the advanceWidth value and leftSideBearing value abnormally, you can use another value instead.
advanceWidth = advanceWidthMax, that is, using the advanceWidth value of special glyph index 0.
leftSideBearing = glyph.xMin.

[jie326513988]

I found that not the first glyph of each font will provide the correct advanceWidth value, so the advanceWidth value of the first glyph cannot be used as the advanceWidthMAX value. Instead, the advanceWidth value of the English space glyph is used, and multiplied by 2 as the advanceWidthMAX value.
Together with the following modifications, most fonts can correctly display Chinese characters and punctuation marks.

[k-omura]

OK. Thanks for the analysis. I am currently checking locally using the TrueType code for STM32 (porting of TrueType_Arduino).
After that, it will be reflected in TrueType_Arduino as well.
Or, if you'd like, you can pullrequest your fix. Your contribution to this library is significant!