New functions/methods + some fixes #85

lubieowoce · 2018-10-18T17:44:45Z

Edit: now that I look at it, this is a huge PR... if you'd like me to break it up into smaller ones, please let me know.

Fixes:

Fixed swapped parameters to Vec4 in multiple places (1cf1e86)
Fixed a division by 0 crash in integrations/glfw (173d8bb)

Additions:

Drawlists:

DrawList:
    add_line
    add_rect
    channels_split
    channels_merge
    channels_set_current

get_overlay_draw_list

Styling:

get_style

GuiStyle:
    from_ref
    color

Cursor functions:

get_cursor_pos
set_cursor_pos
set_cursor_screen_pos

Mouse functions:

is_mouse_down
is_mouse_clicked
is_mouse_released
get_mouse_pos

Misc:

plot_lines
get_time

In addition, I added some aliases. ImGui always shortens "position" to "pos", but we have both get_window_position() and get_cursor_pos(). It would be nice to settle on one and use it consistently. In the meantime, I added aliases from *_pos to *_position. (998d03e, b5b165a)

coveralls · 2018-10-18T18:34:00Z

Coverage decreased (-0.4%) to 46.49% when pulling 88f76c0 on lubieowoce:patch-1 into 43938a1 on swistakm:master.

lubieowoce · 2018-10-18T19:27:37Z

Weird, looks like travis crashed with some virtualenv error before running anything on Python 2.7.

swistakm · 2018-10-22T08:21:34Z

Yes it fails on python 2.7 but it's completely unrelated. It's old issue and I'm fixing it right now.

swistakm · 2018-10-22T09:28:11Z

I have added fix for Python 2.7 builds so this should be resolved now. It would be nice to rebase this PR.

Good point with this pos/position inconsistency. I would prefer to have "position" as default name for these Python bindings. Maybe it would be nice to have "pos" aliases marked somehow with warning that this is an alias but it isn't that critical.

There is still ongoing problem with builds on Windows. They are hanging in queued state without any reason. I have submitted issue to Appveyor but if we won't get any fix we may try to use Windows builds on Travis as they are now available. That would greatly simplify the build pipeline.

ImGui consistently uses ImVec4's fields in the `x, y, z, w` order. Using `x, y, w, z` instead caused bugs – for example, converting an "rgba" Vec4 to ImGui's internal color format would swap the blue and alpha components.

On Windows, when a window is minimized, GLFW reports a window size `w, h = (0, 0)`. Here, we're dividing by w and h, so upon minimization the app crashed from a division by zero. NOTE: I'm not sure what `io.display_fb_scale` should be when the window is minimized, but since nothing is rendered, maybe we can leave it as it was before minimizing?

Also reformatted `add_rect_filled`

…, channels_merge]

- get_cursor_pos - set_cursor_pos - set_cursor_screen_pos

ImGui always shortens "position" to "pos", but we have both `get_window_position` and `get_cursor_pos`. It would be nice to settle on one and use it consistently. In the meantime, I added aliases from `*_pos` to `*_position`.

- is_mouse_down - is_mouse_clicked - is_mouse_released - get_mouse_pos + an alias `get_mouse_position` (see previous commit)

lubieowoce · 2018-10-22T13:52:23Z

I rebased the branch. I prefer "position" as well, but I wanted to avoid messing up the PR with renames – I feel it's big enough as is.

lubieowoce · 2018-10-23T14:29:29Z

@swistakm Btw, how is coverage measured? I didn't really see any tests, but I don't know how you'd test it anyway.

(I guess you could test some properties like "When you call imgui.set_next_window_size(w, h) and create a window, imgui.get_window_size() == (w, h)" but that'd cover only a small subset of the desired properties. One way of testing the graphical stuff would be to write "the same program" using imgui (C++) and pyimgui (python) and then check if they result in the same frame-by-frame output, but you'd have to simulate user input somehow.)

swistakm · 2018-10-23T16:16:20Z

We don’t have any explicit tests yet but we have some usage examples in docstrings that are used to generate visual examples in the dosctrings. So there is a small pytest plugin that uses autodoc feature of Sphinx to generate tests out of these examples. That’s how we measure coverage. It’s rather doc-coverage than real test-coverage.

I know that’s kinda kludgy but it works. I was even thinking about making some reference images so we could compare these “tests results” with some reference images but I didn’t researched yet how we could do off-screen rendering on Travis/Appveyor.

swistakm · 2018-10-23T16:23:33Z

I’m merging it. There are still some issues on Appveyor builds but these should be now fine as I have merged some CI fixes just recently.

lubieowoce · 2018-10-23T20:32:28Z

how we could do off-screen rendering on Travis/Appveyor

This should be doable in OpenGL – you just draw to a buffer, and you're free to copy it and save to disk or something. I don't know what the environment is like in Travis/AppVeyor though – if it's all headless and there's there's no GUI shell/WM present, getting an OpenGL rendering context could be hard/impossible. (but I might be wrong about that – I don't have a lot of experience with headless servers)

lubieowoce · 2018-10-23T21:15:07Z

This Stackoverflow answer talks about testing a GUI app on Travis.

The linked repo's travis.yml sets up a virtual screen with a DISPLAY env var. It also references this link: Travis CI: GUI and headless browsers which sounds relevant.

swistakm · 2018-10-23T21:23:18Z

Yeah I know. This is already done that way for documentation purposes with FBOs although I create hidden window with glfw because it’s the easiest way to create OpenGL context. Besides context we need actual OpenGL implementation. I doubt Travis workers have any graphics hardware. I did already some research in past and it seems that this could be done with software driver like one available in Mesa 3D.

swistakm · 2018-10-23T21:25:32Z

Oh didn’t see your earlier post. I will look into these links.

swistakm · 2019-05-19T20:33:17Z

Released as 1.0.0

lubieowoce changed the title ~~Additions + a few fixes~~ New functions/methods + some fixes Oct 18, 2018

lubieowoce force-pushed the patch-1 branch from a511c8e to ed7a9f7 Compare October 18, 2018 18:15

lubieowoce added 13 commits October 22, 2018 15:43

Fix swapped parameters to Vec4

64b4cab

ImGui consistently uses ImVec4's fields in the `x, y, z, w` order. Using `x, y, w, z` instead caused bugs – for example, converting an "rgba" Vec4 to ImGui's internal color format would swap the blue and alpha components.

Removed redundant attribute accesses in OpenGL renderers

b52a4a4

Added DrawList.[add_line, add_rect]

321962e

Also reformatted `add_rect_filled`

Added channel methods: DrawList.[channels_split, channels_set_current…

baacd69

…, channels_merge]

Added default argument 0 to is_mouse_clicked

75a882d

Implemented get_style and added a way to access a GuiStyle's colors

6b77097

Added cursor functions

a3a8771

- get_cursor_pos - set_cursor_pos - set_cursor_screen_pos

Added pos -> position aliases

870f9e2

ImGui always shortens "position" to "pos", but we have both `get_window_position` and `get_cursor_pos`. It would be nice to settle on one and use it consistently. In the meantime, I added aliases from `*_pos` to `*_position`.

Added mouse functions

5a94281

- is_mouse_down - is_mouse_clicked - is_mouse_released - get_mouse_pos + an alias `get_mouse_position` (see previous commit)

Added get_time

58668c5

Added get_overlay_draw_list

6b6b213

Added plot_lines

88f76c0

lubieowoce force-pushed the patch-1 branch from ed7a9f7 to 88f76c0 Compare October 22, 2018 13:46

swistakm closed this Oct 23, 2018

swistakm reopened this Oct 23, 2018

swistakm merged commit 41112c5 into pyimgui:master Oct 23, 2018

swistakm added the release pending Merged but still needs official release label Oct 23, 2018

swistakm mentioned this pull request Nov 2, 2018

how to use PlotLine in pyimgui? #75

Closed

swistakm added discussion and removed release pending Merged but still needs official release labels May 19, 2019

New functions/methods + some fixes #85

New functions/methods + some fixes #85

Uh oh!

Conversation

lubieowoce commented Oct 18, 2018 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Fixes:

Additions:

Uh oh!

coveralls commented Oct 18, 2018 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

lubieowoce commented Oct 18, 2018

Uh oh!

swistakm commented Oct 22, 2018

Uh oh!

swistakm commented Oct 22, 2018

Uh oh!

lubieowoce commented Oct 22, 2018 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

lubieowoce commented Oct 23, 2018 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

swistakm commented Oct 23, 2018

Uh oh!

swistakm commented Oct 23, 2018

Uh oh!

lubieowoce commented Oct 23, 2018 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

lubieowoce commented Oct 23, 2018

Uh oh!

swistakm commented Oct 23, 2018

Uh oh!

swistakm commented Oct 23, 2018

Uh oh!

swistakm commented May 19, 2019

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

lubieowoce commented Oct 18, 2018 •

edited

Loading

coveralls commented Oct 18, 2018 •

edited

Loading

lubieowoce commented Oct 22, 2018 •

edited

Loading

lubieowoce commented Oct 23, 2018 •

edited

Loading

lubieowoce commented Oct 23, 2018 •

edited

Loading