Write the documentation of lv_led on littlev.hu

Hi. I want to change input device event handling model from polling-event to post-event type.
I don't know whether the change has any advantage or not.
As for electric consumption, it looks that post-event model is better than polling-model.
is it possible for lvgl to prepare the framework for it ?

First of all I am working with the latest beta and can see many many improvements. Two immediate changes are that fonts are clearer and the code seems smaller so thank you for that.

Would it be an idea to allow lv_btn_set_styles to use NULL for button styles which are not to be changed and just supply the ones you want to alter? Or is there already a function to do this?

FYI..
I'm using gcc-arm-none-eabi-6-2017-q2-update/bin/arm-none-eabi-gcc.
It causes "sorry, unimplemented: Thumb-1 hard-float VFP ABI" message.

i am desperate to get this to the Arduino IDE, but i get a huge list of errors when compiling and i can't figure it out due to lack of knowing lvgl well enough... perhaps port it to arduino as well?

Looking in the source I can see that color16_t is using uint16_t as opposed to uint8_t. Is there a reason for this?

Looking for some advice. I have a populated lv_list and would like the button to permanently change colour if it is selected and switch back to the "normal" colour if another button is then selected (with the new button changing colour permanently).

I can get the colour to change by attaching a function to the release event and, in that function, doing an lv_btn_get_style on the passed lv_obj_t and then updating the ccolor and mcolor. I then store the passed lv_obj_t to change the colour back on the next button press. The first press works but the colour change on the next button doesn't reset the first button. Is there a way to do this?

Is there any chance of 3D functions to the engine?
Like lights, polygons, meshes, vertexes.
(There are some repos on GitHub for MCU's that have some of these features)

how to support muti language,such as Chinese Japanese and Korean

I have a question. If I would like to adjust the position of text in lv_ta, which vpad and hpad should I take?

• Style of lv_ta
• Style of scrollbar style_sb from lv_page_get_style_sb(ta)

I made some try. I use ptask to change vpad periodically.
I set vpad on style of lv_ta, it worked on the first time call lv_obj_set_style(ta, &ta_style), but it didn't work on the second time or set style to the other new lv_style_t.
About style_sb, It's no effect by changing its vpad.

Thanks!

Using the following code:

panelId = lv_list_create(sscrn, NULL);
lv_obj_set_size_us(panelId, LCD_GetXSize()-20, LCD_GetYSize() - 135);
lv_obj_set_pos_us(panelId, 10, 10);

I find that if the lv_list is empty then the panel size is a single line, i.e. it ignores the set_size. Is that deliberate?

LittlevGL is nice.
I figured it using some object drawing interface, but when I wana just draw a simple line or rectangle or just for testing the hardware, the Lib just stop in lv_refr_areas(), seems like I should enbale the area firstly?

When LV_MEM_CUSTOM is set to 1, ent_trunc is undefined.

I’m learning the source code，and have ported to the stm32412g-disco.
That's amazing.
The screen is 1.54 inches, a little, suitable for wearable design.

img_add.c reporting error of a missing file
#include "img_conf.h"

I think there is no template file img_conf_tmpl.h which can tell what goes in img_conf.h

so i set the LV_INDEV_POINT_MARKER to 5 to see my presses(?) in the lvgl config file

buuuut.... my IDE throws compiling errors.

the complete error log:

C:\Users\Ramon\Documents\Arduino\libraries\lvgl\src\lv_obj\lv_indev.c: In function 'indev_proc_point':

C:\Users\Ramon\Documents\Arduino\libraries\lvgl\src\lv_obj\lv_indev.c:286:9: error: unknown type name 'area_t'

         area_t area;

         ^

C:\Users\Ramon\Documents\Arduino\libraries\lvgl\src\lv_obj\lv_indev.c:287:13: error: request for member 'x1' in something not a structure or union

         area.x1 = (indev->act_point.x >> LV_ANTIALIAS) - (LV_INDEV_POINT_MARKER >> 1);

             ^

C:\Users\Ramon\Documents\Arduino\libraries\lvgl\src\lv_obj\lv_indev.c:288:13: error: request for member 'y1' in something not a structure or union

         area.y1 = (indev->act_point.y >> LV_ANTIALIAS) - (LV_INDEV_POINT_MARKER >> 1);

             ^

C:\Users\Ramon\Documents\Arduino\libraries\lvgl\src\lv_obj\lv_indev.c:289:13: error: request for member 'x2' in something not a structure or union

         area.x2 = (indev->act_point.x >> LV_ANTIALIAS) + ((LV_INDEV_POINT_MARKER >> 1) | 0x1);

             ^

C:\Users\Ramon\Documents\Arduino\libraries\lvgl\src\lv_obj\lv_indev.c:290:13: error: request for member 'y2' in something not a structure or union

         area.y2 = (indev->act_point.y >> LV_ANTIALIAS) + ((LV_INDEV_POINT_MARKER >> 1) | 0x1);

             ^

C:\Users\Ramon\Documents\Arduino\libraries\lvgl\src\lv_obj\lv_indev.c:291:18: warning: passing argument 1 of 'lv_rfill' from incompatible pointer type [-Wincompatible-pointer-types]

         lv_rfill(&area, NULL, LV_COLOR_MAKE(0xFF, 0, 0), LV_OPA_COVER);

                  ^

In file included from C:\Users\Ramon\Documents\Arduino\libraries\lvgl\src\lv_obj\lv_indev.c:16:0:

c:\users\ramon\documents\arduino\libraries\lvgl\src\lv_draw\lv_draw_rbasic.h:41:6: note: expected 'const lv_area_t * {aka const struct <anonymous> *}' but argument is of type 'int *'

 void lv_rfill(const lv_area_t * cords_p, const lv_area_t * mask_p,

      ^

If I disable file system interface
USE_FSINT = 0

I am not able to build at all. So its not modular as such. Don't you think functions like:
lv_img_create_file should be protected with flags like

#if USE_FSINT && USE_UFS

There looks to be an issue with a password text area field. The following:

lv_obj_t *pin_inp = lv_ta_create(lv_scr_act(), NULL);
 lv_ta_set_one_line(pin_inp, true);
 lv_ta_set_cursor_show(pin_inp, true);
 lv_ta_set_pwd_mode(pin_inp, true);
 lv_ta_set_text(pin_inp, "");
 lv_ta_set_cursor_pos(pin_inp, LV_TA_CUR_LAST);
 lv_ta_add_char(pin_inp, '1');
 lv_ta_set_cursor_pos(pin_inp, LV_TA_CUR_LAST);
 lv_ta_del(pin_inp);
 lv_ta_set_cursor_pos(pin_inp, LV_TA_CUR_LAST);
 lv_ta_add_char(pin_inp, '2');
 char buff[3];
 strcpy(buff, lv_ta_get_txt(pin_inp));

Will leave "21" in buff instead of "2"

Hi,

I would like to use the library on a system with just a clickable rotary encoder as an input device. How can I accomplish that since the library doesn't seem to have the active widget state?

i.e. rotate the encoder changes the active button (or list item or whatever) and clicking on it sends a "click" to the active widget.

At present I'm using the free_num in the returned object to add an index on listview items but envisage that there will be more than 1000 items. Would it be possible to add a free_word or is there a better way to store the index values?

1. LV_MEM_SIZE
   This is for LVGL working needed, there are lots of dynamic ram to be allocated.

2. LV_VDB_SIZE, LV_VDB_ADR
   I use RGB interface in my stm32f4, so this memory is my display area. The screen will change immediately as I change something in this area.

3. LV_VDB_DOUBLE, LV_VDB2_ADR
   The same as VDB, another parellel memory for double buffering. When I operate VDB1, LCD controller will fetch from VDB2, and swap in next cycle.

Is that right?

It is a small suggestion to keep library clean and independent. Is it possible to move lv_app/lv_appx out of lvgl source as they are not part of library as such. May it can be moved to a new repo just like hal/misc etc. This way lvgl as a source will have no code that may or may not be usable to others. One can always close the lvapp repo if he/she needs.

Some LCD display with capacitive touch have multi touch capability.

like for STM32, commonly typedef struct used is like this :

typedef struct
{
uint8_t touchDetected; /*!< Total number of active touches detected at last scan /
uint16_t touchX[TS_MAX_NB_TOUCH]; /!< Touch X[0], X[1] coordinates on 12 bits /
uint16_t touchY[TS_MAX_NB_TOUCH]; /!< Touch Y[0], Y[1] coordinates on 12 bits */

#if (TS_MULTI_TOUCH_SUPPORTED == 1)
uint8_t touchWeight[TS_MAX_NB_TOUCH]; /*!< Touch_Weight[0], Touch_Weight[1] : weight property of touches /
uint8_t touchEventId[TS_MAX_NB_TOUCH]; /!< Touch_EventId[0], Touch_EventId[1] : take value of type @ref TS_TouchEventTypeDef /
uint8_t touchArea[TS_MAX_NB_TOUCH]; /!< Touch_Area[0], Touch_Area[1] : touch area of each touch /
uint32_t gestureId; /!< type of gesture detected : take value of type @ref TS_GestureIdTypeDef /
#endif / TS_MULTI_TOUCH_SUPPORTED == 1 */

} TS_StateTypeDef;

So it's good if multi touch capability can be implemented in lvgl.

Would it be possible to add additional cursor style?

for example:

• block style
  
• block-outline style
  
• underline style
  

Thank you!

1. set a non-empty string without newline to lv_ta.
2. enable lv_ta_set_one_line
3. take lv_ta_cursor_right to move cursor from left to right .
4. cursor will disappear once it reach the right-most character.

I want it to play video files, so i want to Make a mediaplayer by littlevgl .

The following code should leave "1234" in buff but actually leaves "***4":

lv_obj_t *pin_inp = lv_ta_create(lv_scr_act(), NULL);
lv_ta_set_one_line(pin_inp, true);
lv_ta_set_pwd_mode(pin_inp, true);
lv_ta_set_text(pin_inp, "");
lv_ta_set_cursor_pos(pin_inp, LV_TA_CUR_LAST);
lv_ta_add_char(pin_inp, '1');
lv_ta_set_cursor_pos(pin_inp, LV_TA_CUR_LAST);
lv_ta_add_char(pin_inp, '2');
lv_ta_set_cursor_pos(pin_inp, LV_TA_CUR_LAST);
lv_ta_add_char(pin_inp, '3');
lv_ta_set_cursor_pos(pin_inp, LV_TA_CUR_LAST);
lv_ta_add_char(pin_inp, '5');
lv_ta_set_cursor_pos(pin_inp, LV_TA_CUR_LAST);
lv_ta_del(pin_inp);
lv_ta_set_cursor_pos(pin_inp, LV_TA_CUR_LAST);
lv_ta_add_char(pin_inp, '4');
char buff[5];
strcpy(buff, lv_ta_get_txt(pin_inp));

Is it possible to send signal with a NULL display input?

I would like to sync the a physical button with a lv_btn, so I try to send signals to it.

lv_btn_signal(btn, LV_SIGNAL_RELEASED, NULL)

In current implementation, to set display input to NULL, it will cause the following function error when button is released.
https://github.com/littlevgl/lvgl/blob/c96b1d4db37a2ab3156f7a6f97f6eed2052dfaca/lv_obj/lv_dispi.c#L113

I've optimized the font storage to use less memory.
@ajaybhargav please generate the symbol fonts again with the new python script: https://github.com/littlevgl/utils/blob/beta/font_create/fnt2c.py

Usage:
python fnt2c.py -f <font_name> [-o <output file> -s <start unicode> -e <end unicide>]

Or see python fnt2c.py -h for a detailed explanation.

If you have no time right now to deal with it just send me the fnt and png files and I will do this.

Thank you in advance!

First of all, great lib!

On most microcontrollers, copying the data from the memory to the display can be handled by a DMA controller or timers. If such method is used, the CPU is not utilized during data transfer to the display. So the CPU can perform other tasks, like rendering the next part of the screen.

In virtual drawing mode this can speedup the GUI a lot. Instead of one buffer, use two buffers. While the first buffer is being transferred to the screen by the DMA controller, the second buffer is filled by the CPU.

What do you think?
What needs to be changed in the lib to support this? At least the blocking interface API functions need to be changed.

Hi I just wanted to know why LV_UPSCALE_MAP is now removed from core?

I'm trying to use a text area which is one_line and pwd_mode both set true. Using lv_ta_set_cursor_pos LV_TA_CUR_LAST and lv_ta_add_char to append characters shows them on the screen before they are replaced by '*'. However, when I use lv_ta_get_txt I get 'Text area' returned and if I use lv_label_get_text I get an empty string. If I also use lv_label_append_text then I am able to get the text back that I would expect. Wouldn't it better to have lv_ta_get_txt return the contents of the text area or have I misunderstood how this works?

I have the following sequence of events: Create list full width of screen -> create smaller list -> create screen 2 which covers the physical screen -> delete full width screen list -> delete screen 2. Now the smaller is list is below the full width list. I can see why this is happening (order based on creation time) but I wondered if it is possible within the existing code to change the order so that the smaller list always stays on top of the full width list?

Is it possible to add a gauge object, similar operation to that or chart/progress bar maybe. Gauges looks great on displays and make it look more intuitive.

It might be good to allow the background colour of a label to be set via mcolor and inherit the parent object's mcolor if it's not supplied.

If a system already provide malloc/free functions then there is no need to use misc/mem module. This will save lot of code space. You can move memory management to HAL instead so people can either write their own dm_* functions or enable USE_DYN* flag if system does not have malloc/free function.

Create a separate dm_mem.h file to define only the function prototypes of needed functions:

dm_init()
dm_alloc()
dm_free()
dm_realloc()

Don't know if I'm doing something wrong but switching FONT_ANTIALIAS and doubling the font size means that the default height of an lv_ta doubles as well, despite the text displaying at half size. Is there a way around this?

I dont know how VDB is linked to objects, it seems display area is recreated on VDB everytime display needs to be updated (right?). I have gone through lv_refr.c and found that vdb is filled there and then flushed on to display row by row. Somehow this makes display update slower and one can see line by line update like raster scan. I was thinking it would be great if there is a way VDB gets filled first and then sent to display routine to flush it at once. Smaller VDB size makes it even worse. I am not taking about VDB=0 as its not good at all.

In the next release (v4.3.0) I'm planning to add new smart phone like object types.
I thought about the following types:

• lv_split a hor. or ver. line for decoration purpose
• lv_valset a label with "+" and "-" buttons
• lv_tabview organise content with tabs (support swipe left/right)
• lv_btngrp display more buttons to choose an option from them
• lv_switch turn on/off by tap (a little slider)
• lv_roller roller to select a value (like on smartphones)

If you have any new idea or suggestion don't hesitate to write a comment!

i using qt_mingw build err,and correct "int main(viod)"to "int winmain(int a,char *arg)" build ok and runing ok!

Would it be possible to add a sort function to the lv_list_add function or create a new lv_list_add_sorted function?

It would be good to have way to delete a list item.

Using default style for container and buttons (each button has its own object):

Before pressing "1":

After pressing "1":

Note corner "decorations" which appear after pressing

The next release will be a major release therefore it is possible to update API.
The question is: is it good to abbreviate in API function's name or not?
For example:

The not abbreviated names are more descriptive but takes more time to write and read.
What's worth more? What is you opinion?

https://www.youtube.com/watch?v=DcJdK137WKM
That is amazing!

stm32 is a powerful mcu,can you port to it ?
stm32f429-disco

I have some question about lv_ta on website's document

1. text style

   http://www.gl.littlev.hu/object-types/text-area-lv_ta

   Style usage

   • ...
   • Text style is inherited from the parent page.

   Dose the "parent page" mean lv_ta's parent must be a lv_page object?
   But it seems that the text style can be accessed from label member in lv_ta.

2. cursor style

   Style usage

   • ...
   • Cursor width and color comes from the parent's ccolor and line_width style properties.

   I try to create a lv_con as the parent of lv_ta, then change its ccolor and line_width, but it doesn't apply to cursor.

Thank you!

Thanks for your work. Please tell me how I can use fonts with UTF-8 encoding support