It depends on #4833, right?

Yes. Only with draw_buf for decoded image, we can get info of decoded image.
Could you please review if the method works? Why the area to draw could be larger than image itself?

4393793#diff-242f305692436f289db7b3cdda6298157b9ab35835506b88b28a86b1524051abR266-R274

Can you send a code snippet in which the are to draw is really larger than the image area?

The barcode could be used to reproduce it. Do change LV_BIN_DECODER_RAM_LOAD to 0 to test it.

Comment out this line 274 4393793#diff-242f305692436f289db7b3cdda6298157b9ab35835506b88b28a86b1524051abR274
and you will see logs to report area not ready.

void test_barcode_normal(void)
{
    lv_obj_t * barcode = lv_barcode_create(lv_scr_act());
    lv_obj_center(barcode);

    lv_color_t dark_color = lv_color_black();
    lv_color_t light_color = lv_color_white();
    uint16_t scale = 2;

    lv_barcode_set_dark_color(barcode, dark_color);
    lv_barcode_set_light_color(barcode, light_color);
    lv_barcode_set_scale(barcode, scale);

    lv_result_t res;
    lv_image_dsc_t * image_dsc;
    lv_barcode_set_direction(barcode, LV_DIR_HOR);
    res = lv_barcode_update(barcode, "https://lvgl.io");
    image_dsc = lv_canvas_get_image(barcode);

    lv_obj_set_size(barcode, image_dsc->header.w, 50);

    lv_barcode_set_direction(barcode, LV_DIR_VER);
    res = lv_barcode_update(barcode, "https://lvgl.io");

    image_dsc = lv_canvas_get_image(barcode);

    lv_obj_set_size(barcode, 50, image_dsc->header.h);

}

I think there is an other issue here as well.

The barcode with LV_COLOR_DEPTH 32 and LV_BIN_DECODER_RAM_LOAD 0 looks like this:

The barcode with LV_COLOR_DEPTH 32 and LV_BIN_DECODER_RAM_LOAD 0 looks like this:

Could you try with LV_BIN_DECODER_RAM_LOAD to 1 and check out code XuNeo:fix-bar-code-get-area-cb? I'll resolve the conflict later.
The issue is fixed on my local and will push once the draw buffer things merged. And then we can enable it on CI.

It works well on my end too.

Comment out this line 274 4393793#diff-242f305692436f289db7b3cdda6298157b9ab35835506b88b28a86b1524051abR274
and you will see logs to report area not ready.

Hi Gabor,
This PR has made it to work by clipping the area to the whole image area and then compare if it's inside decoded image area.
You need to comment out above code to see the error.

I've commented out

Enabled LV_BIN_DECODER_RAM_LOAD 1 but I can't see any error.

The areas look good, and clipped when I move the cursor on the bar code.

I can't see and logged errors either.

Exactly for what issue shall I look for?

Enabled LV_BIN_DECODER_RAM_LOAD 1 but I can't see any error.

That's really interesting. I'm not sure which commit has fixed it.

However, now with online latest code + this PR, I can now reproduce the issue steadily with LV_BIN_DECODER_RAM_LOAD configured to 0.

[Warn]  (0.020, +20)     lv_image_set_src: failed to get image info: 0x60e000000a50 lv_image.c:174
[User]  (0.021, +1)      img_decode_and_draw: area not ready lv_draw_sw_img.c:285
[User]  (0.021, +0)      img_decode_and_draw: area not ready lv_draw_sw_img.c:285
[User]  (0.021, +0)      img_decode_and_draw: area not ready lv_draw_sw_img.c:285
[User]  (0.021, +0)      img_decode_and_draw: area not ready lv_draw_sw_img.c:285
[User]  (0.021, +0)      img_decode_and_draw: area not ready lv_draw_sw_img.c:285

I've tested it an I think this is what happens:

Let's we have a 100x1 image, which is tiled to get a 100x20 image. For the sake of simplicity lets say the image is at 0;0.

Now we invalidate the (50;10)(60;15) area. So we need to decode x=50..60 and draw it 6 times (y=10..15). In the first run it's correctly detected that the image is not decoded yet, so that 11 pixel (x=50..60) are decoded. However in the next run here it's assumed that the whole image was decoded which is not true.

So we need to keep relative_decoded_area accross calls of img_decode_and_draw. Maybe saving it in decoder_dsc?

We need some feedback on this pull request.

Now we mark this as "stale" because there was no activity here for 14 days.

Remove the "stale" label or comment else this will be closed in 7 days.

Is it already resolved by some other PRs?

It's still ongoing. Will go back too it in following days.

We need some feedback on this pull request.

Now we mark this as "stale" because there was no activity here for 14 days.

Remove the "stale" label or comment else this will be closed in 7 days.

Shall we pin this issue?

No need, I shall dig into it right now.

Hi @kisvegabor
The code is working. I store the decoded image area between decoder open and close. Please help to review it.

Cool.
The next thing is to add option LV_BIN_DECODER_RAM_LOAD = 0in CI to make sure it always works.
Could you do that?

I've opened #5396 to discuss what kind of test configs are required.