Add support for 9-slice/9-patch -> merge with latest upstream version + remove build.py related code #304

lautriva · 2021-05-31T18:55:42Z

It seems that the PR #69 is abandoned, but the 9patch feature would be nice to have in Nuklear and as @Hejsil suggests, it look like very close to validation 🙂
So I've done the merge to latest master commit (at the time of writing) + removed changes related to build.py

No changes have been made on @Michael-Kelley's code 🙂

I think that Nuklear could be used also as an end-user GUI and not only as developer's one and I think this PR is a nice step toward it
(And I'm more convinced when I saw the last demo screenshot with the army-fps-ish options menu)

…cremented version to 4.01.6.

…er again

…thon 2.x on Windows and added setting to .gitattributes to ensure line endings are consistent amongst contributers

…draw functions.

dumblob · 2021-05-31T19:24:52Z

@FrostKiwi could you maybe find some time for this? There seems to be quite some demand to prioritize this one and I'm still out of time for this project 😢.

FrostKiwi · 2021-06-06T21:22:33Z

@FrostKiwi could you maybe find some time for this? There seems to be quite some demand to prioritize this one and I'm still out of time for this project .

Will dive into it :]

Hejsil

Just took a look over it again. Haven't tested it, but the code looks good 👍 I've left some nitpick comments :)

Hejsil · 2021-06-18T13:57:05Z

src/nuklear_9slice.c

+{
+    struct nk_9slice s;
+    nk_zero(&s, sizeof(s));
+    struct nk_image *i = (struct nk_image*)&s;


Why the cast when we could just do struct nk_image *i = &s.img; ?

Can we get this changed, @dav64 ? It makes the code less clear, as it is, IMO.
Other that this, the rest of the change looks good.
It would be good to have a usage example in the demos -- @FrostKiwi, are you still up for writing a new skinning example?

After code investigating, I didn't found if there were specific reason for the original author (Michael) to not using something like @Hejsil solution as it also look cleaner to me
Maybe that only was Nuklear casting fashion at the time of writting 🤔
I'll update that

Updated, let me know if you need anything else 🙂

Hejsil · 2021-06-18T13:58:39Z

src/nuklear_9slice.c

+{
+    struct nk_9slice s;
+    nk_zero(&s, sizeof(s));
+    struct nk_image *i = (struct nk_image*)&s;


Hejsil · 2021-06-18T13:58:46Z

src/nuklear_9slice.c

+{
+    struct nk_9slice s;
+    nk_zero(&s, sizeof(s));
+    struct nk_image *i = (struct nk_image*)&s;


Hejsil · 2021-06-18T13:58:52Z

src/nuklear_9slice.c

+{
+    struct nk_9slice s;
+    nk_zero(&s, sizeof(s));
+    struct nk_image *i = (struct nk_image*)&s;


Hejsil · 2021-06-18T13:58:58Z

src/nuklear_9slice.c

+{
+    struct nk_9slice s;
+    nk_zero(&s, sizeof(s));
+    struct nk_image *i = (struct nk_image*)&s;


Hejsil · 2021-06-18T13:59:38Z

src/nuklear_9slice.c

+{
+    struct nk_9slice s;
+    nk_zero(&s, sizeof(s));
+    struct nk_image *i = (struct nk_image*)&s;


It's everywhere! 😱

Hejsil · 2021-06-18T14:02:26Z

src/nuklear_button.c

+    switch(background->type) {
+        case NK_STYLE_ITEM_IMAGE:
+            nk_draw_image(out, *bounds, &background->data.image, nk_white);
+            break;
+        case NK_STYLE_ITEM_9SLICE:
+            nk_draw_9slice(out, *bounds, &background->data.slice, nk_white);
+            break;
+        case NK_STYLE_ITEM_COLOR:
+            nk_fill_rect(out, *bounds, style->rounding, background->data.color);
+            nk_stroke_rect(out, *bounds, style->rounding, style->border, style->border_color);
+            break;
    }


Seems we have this switch in quite a few places. I wonder if an nk_draw_background function would be something nice to have 🤔

FrostKiwi · 2021-06-28T18:32:22Z

Learning to properly use skinning in this new context took me forever and properly exporting skinning designs from Photoshop has also pitfalls.
Next PR after this should be an updated documentation :S
Have an exam coming up this week, will give my review with some designs to test on next week.

lockie · 2021-08-05T04:57:50Z

Is there a chance this would be merged anytime soon?

dumblob · 2021-08-07T09:40:13Z

@lockie we strongly prefer review of more than one person but here only @Hejsil found his precious time to review it so far (I'm too busy for this 😢). If you could review it, we'll immediately merge this.

And if you'll review few other PRs, I'll give you full rights to this repo afterwards as we need more of interested people.

lockie · 2021-08-07T12:35:35Z

@dumblob thank you for your trust! I'll have a deeper look into the code of this PR a bit later.

lockie · 2021-08-14T10:25:39Z

I've took the time to review this patch, and besides the possible issue with naming and language bindings, it is pretty straightforward and LGTM 👍

lockie · 2021-08-14T10:23:47Z

nuklear.h

@@ -480,7 +480,8 @@ struct nk_rect {float x,y,w,h;};
 struct nk_recti {short x,y,w,h;};
 typedef char nk_glyph[NK_UTF_SIZE];
 typedef union {void *ptr; int id;} nk_handle;
-struct nk_image {nk_handle handle;unsigned short w,h;unsigned short region[4];};
+struct nk_image {nk_handle handle; nk_ushort w, h; nk_ushort region[4];};
+struct nk_9slice {struct nk_image img; nk_ushort l, t, r, b;};


The naming here might be problematic for other language bindings. In those the nk_ prefix is usually removed and moved to e.g. package name (or whatever target language provides), so in those bindings we got something like nk.image instead of nk_image. I'm author of Common Lisp bindings, and there probably wouldn't be the problem there - CL allows for names like 9slice, but other languages might not.

Maybe rename to scale9 or grid9 or slice9?

Something like nine_slice might do the trick as well.

Renamed 9slice to nine_slice
Let me know if you need anything else 🙂

Prevents breaking other languages bindings when nk_ is replaced as a namespace and identifiers cannot start with a digit

lockie

Looks like it is good to merge 🙂

Hejsil · 2021-08-17T20:34:58Z

Nice!

Add support for 9-slice/9-patch (Immediate-Mode-UI#304)

FrostKiwi · 2023-02-27T17:31:09Z

@Michael-Kelley @dav64
Anyone know what nk_ushort l, nk_ushort t, nk_ushort r, nk_ushort b stands for? It's a public API, but not documented as to how you are supposed to use it. Anytime you interact with anything 9-slice you have to specify it. I figured it's the square corner size of the 9 slice and have treated it as such, however, that doesn't actually seem to be quite true.

What does l t r b stand for? Would love to properly document it.

lautriva · 2023-02-27T17:51:12Z

If I remember correctly, it stands for Left, Top, Right and Bottom

But I cannot remember exactly the specifics/inner working as its been a while since I didn't used Nuklear

Hope this helps

Ashe · 2023-03-07T15:33:33Z

Funnily enough I found myself here also needing help with this.

The letters are indeed top-left-right-bottom, but to my surprise the entire image is used in each segment of the nineslice.

I was EXPECTING it to say 'the top t pixels of the image is used for the top 3 segments of the nine-slice', but actually it's 'the entire texture is stretched as each part of the nineslice'

It looks very bizarre, maybe I'm missing something but I don't know how this is supposed to be used other than

                nk_nine_slice slice;
                slice.img = *skin->Get();
                slice.b = 50;
                slice.t = 50;
                slice.l = 50;
                slice.r = 50;
                m_context->style.window.fixed_background = nk_style_item_nine_slice(slice);

This results in a copy of the image being in the top-left corner as a 50x50 image, then a stretched image of 50 px in height at the top, then another 50x50 in the top right.. Same for all the other sides.

revolucas · 2023-03-08T13:56:18Z

Seemed to work when I tested it. I couldn't get it to look right with smaller resolution images I created but had it working with the example image from here: https://devforum.roblox.com/t/9-slice-editor-studio-beta/1534739

Are you setting the image width, height and region property in your 'Get'?

Ashe · 2023-03-08T14:10:52Z

Get is just returning my nk_image, which I create in my object like so:

image = std::make_unique<struct nk_image>(nk_image_id(*textureID));

So yeah, you're right I didn't set the width/height/region, whoops. Crazy that images render fine without them though!.

What is 'region'? Could you give me a snippet of code for where those values come from?

EDIT: Thanks for your reply by the way!

EDIT EDIT: I got it working. Here's what they represent for future readers:

region[0]: x-coordinate of the top-left corner of the sub-region
region[1]: y-coordinate of the top-left corner of the sub-region
region[2]: width of the sub-region
region[3]: height of the sub-region

The region array in nk_image represents the sub-region of the image that should be displayed, if you don't specify the region values, they default to zero, which means that the entire image is displayed.

However, if you set the w and h values of the nk_image to something other than zero, it means that the image is expected to have a different size than its actual size. This is useful for cases where you want to scale an image up or down, or when you want to display only a part of a larger image.

Michael-Kelley and others added 10 commits March 2, 2020 10:24

Added support for 9-slice/9-patch textures for various widgets and in…

fe2f2fe

…cremented version to 4.01.6.

Synced with upstream, fixed merged conflicts, and bumped version numb…

4948eb2

…er again

Fixed issue with paq.bat generating inconsistent line endings with Py…

9e08590

…thon 2.x on Windows and added setting to .gitattributes to ensure line endings are consistent amongst contributers

Proper fix for newline issues in build.py

daed6fc

Replaced if-else with switch-case for background rendering in widget …

f7dee86

…draw functions.

Synced with upstream and bumped version number.

561b1c0

Bumped minor version.

4472388

Merge remote-tracking branch 'upstream/master'

292b99f

Remove missing text background

b253d9c

Fix date typo

b150310

Hejsil approved these changes Jun 18, 2021

View reviewed changes

Hejsil mentioned this pull request Jun 21, 2021

Added support for 9-slice/9-patch image for various widgets #69

Closed

David B added 2 commits August 8, 2021 11:38

Merge branch 'master' of https://github.com/Immediate-Mode-UI/Nuklear

e9a9a05

9slice: Replace cast

49f5d8f

lockie reviewed Aug 14, 2021

View reviewed changes

David B added 2 commits August 17, 2021 18:54

Merge remote-tracking branch 'remotes/upstream/master'

22a9ead

Rename 9slice to nine_slice

9bba99e

Prevents breaking other languages bindings when nk_ is replaced as a namespace and identifiers cannot start with a digit

lockie approved these changes Aug 17, 2021

View reviewed changes

Hejsil merged commit 6322f53 into Immediate-Mode-UI:master Aug 17, 2021

sthagen added a commit to sthagen/Immediate-Mode-UI-Nuklear that referenced this pull request Aug 19, 2021

Merge pull request #8 from Immediate-Mode-UI/master

679fc4c

Add support for 9-slice/9-patch (Immediate-Mode-UI#304)

RobLoach mentioned this pull request Dec 22, 2021

9-slice/9-patch support for skinning #68

Closed

FrostKiwi mentioned this pull request Feb 27, 2023

[5.x] Community sourced Nuklear Default skin and Design file #548

Open

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Add support for 9-slice/9-patch -> merge with latest upstream version + remove build.py related code #304

Add support for 9-slice/9-patch -> merge with latest upstream version + remove build.py related code #304

lautriva commented May 31, 2021

dumblob commented May 31, 2021

FrostKiwi commented Jun 6, 2021

Hejsil left a comment

Hejsil Jun 18, 2021

2asoft Aug 8, 2021

lautriva Aug 8, 2021

lautriva Aug 8, 2021

Hejsil Jun 18, 2021

Hejsil Jun 18, 2021

Hejsil Jun 18, 2021

Hejsil Jun 18, 2021

Hejsil Jun 18, 2021

Hejsil Jun 18, 2021

FrostKiwi commented Jun 28, 2021

lockie commented Aug 5, 2021

dumblob commented Aug 7, 2021

lockie commented Aug 7, 2021

lockie commented Aug 14, 2021

lockie Aug 14, 2021

dumblob Aug 17, 2021

lockie Aug 17, 2021

lautriva Aug 17, 2021

lockie left a comment

Hejsil commented Aug 17, 2021

FrostKiwi commented Feb 27, 2023

lautriva commented Feb 27, 2023

Ashe commented Mar 7, 2023

revolucas commented Mar 8, 2023

Ashe commented Mar 8, 2023 •

edited

Loading

Add support for 9-slice/9-patch -> merge with latest upstream version + remove build.py related code #304

Add support for 9-slice/9-patch -> merge with latest upstream version + remove build.py related code #304

Conversation

lautriva commented May 31, 2021

dumblob commented May 31, 2021

FrostKiwi commented Jun 6, 2021

Hejsil left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

FrostKiwi commented Jun 28, 2021

lockie commented Aug 5, 2021

dumblob commented Aug 7, 2021

lockie commented Aug 7, 2021

lockie commented Aug 14, 2021

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

lockie left a comment

Choose a reason for hiding this comment

Hejsil commented Aug 17, 2021

FrostKiwi commented Feb 27, 2023

lautriva commented Feb 27, 2023

Ashe commented Mar 7, 2023

revolucas commented Mar 8, 2023

Ashe commented Mar 8, 2023 • edited Loading

Ashe commented Mar 8, 2023 •

edited

Loading