@vadz @a-wi I'm just wondering why this commit is needed? because LPtoDP/DPtoLP should work correctly in RTL layout as documented.
Anyway, currently wxMSW is broken in RTL layout:
wxGrid (in griddemo)I tested this PR in RTL layout and it works correctly for me. Although running the RTL tests still fail and I think I should disable them for now untill I (or anyone else) manage to run them successfully in the future!
https://github.com/wxWidgets/wxWidgets/pull/25426
(3 files)
—
Reply to this email directly, view it on GitHub, or unsubscribe.
You are receiving this because you are subscribed to this thread.![]()
@AliKet pushed 1 commit.
—
View it on GitHub or unsubscribe.
You are receiving this because you are subscribed to this thread.![]()
Sorry, I don't have time to look at this right now, but it seems wrong to disable the previously working tests unless we have good reasons to think the tests themselves were incorrect -- do we?
—
Reply to this email directly, view it on GitHub, or unsubscribe.
You are receiving this because you are subscribed to this thread.![]()
Sorry, I don't have time to look at this right now, but it seems wrong to disable the previously working tests unless we have good reasons to think the tests themselves were incorrect -- do we?
I'll try to look into the failing tests later and try to fix them, but RTL in master is definitely broken! see:
drawing_rtl_bad.png (view on web)
grid_rtl_bad.png (view on web)
With this PR:
drawing_rtl_ok.png (view on web)
grid_rtl_ok.png (view on web)
—
Reply to this email directly, view it on GitHub, or unsubscribe.
You are receiving this because you are subscribed to this thread.![]()
Yes, it does look broken, but I'm worried about partially but not completely fixing it in 3.3.0, resulting in a different kind of brokenness there and more fixes needed for 3.3.1.
The problem is really that we plan to release it in 2 days and this is a relatively big change and I (don't know about you?) am far from being 100% confident that it's correct...
—
Reply to this email directly, view it on GitHub, or unsubscribe.
You are receiving this because you are subscribed to this thread.![]()
Yes, it does look broken, but I'm worried about partially but not completely fixing it in 3.3.0, resulting in a different kind of brokenness there and more fixes needed for 3.3.1.
The problem is really that we plan to release it in 2 days and this is a relatively big change and I (don't know about you?) am far from being 100% confident that it's correct...
It's not urgent for me for this to be included in the upcoming release, and I plan to debug this later to try to get rid of commit 928d7eb if possible because it really bothers me. So I suggest adding work neededto this as a reminder, TIA!
—
Reply to this email directly, view it on GitHub, or unsubscribe.
You are receiving this because you are subscribed to this thread.![]()
@AliKet pushed 2 commits.
—
View it on GitHub or unsubscribe.
You are receiving this because you are subscribed to this thread.![]()
drawing sample with and without commit 508460d and a transformation applied to the DC:
With:
drawing_trans_rtl_bad.png (view on web)
Without:
drawing_trans_rtl_ok_2.png (view on web)
—
Reply to this email directly, view it on GitHub, or unsubscribe.
You are receiving this because you are subscribed to this thread.![]()
I'd really like to merge this but I feel that if we do it now, we'll never reenable the tests later, so let's postpone it a bit until someone has time to understand what is broken here.
—
Reply to this email directly, view it on GitHub, or unsubscribe.
You are receiving this because you are subscribed to this thread.![]()
@AliKet pushed 8 commits.
—
View it on GitHub or unsubscribe.
You are receiving this because you are subscribed to this thread.![]()
@AliKet pushed 6 commits.
—
View it on GitHub or unsubscribe.
You are receiving this because you are subscribed to this thread.![]()
@AliKet pushed 1 commit.
—
View it on GitHub or unsubscribe.
You are receiving this because you are subscribed to this thread.![]()
@AliKet pushed 2 commits.
—
View it on GitHub or unsubscribe.
You are receiving this because you are subscribed to this thread.![]()
@vadz @paulcor Is this commit 6e94053 looks correct to you? Because it works correctly for me with wxGTK3 in RTL layout (try moving panes in the auidemo).
Disclaimer: I haven't tested it under wxGTK2 because last time I checked the RTL support wasn't great there!
—
Reply to this email directly, view it on GitHub, or unsubscribe.
You are receiving this because you are subscribed to this thread.![]()
@AliKet pushed 4 commits.
—
View it on GitHub or unsubscribe.
You are receiving this because you are subscribed to this thread.![]()
@vadz @paulcor I think there is a bug in wxDataViewCheckIconTextRenderer() when using the native wxDataViewCtrl control in wxGTK3 (I'm not sure because I'm not familiar with the (native) wxDataViewCtrl drawing code). i.e.: the check+icon+text is always drawn LTR. The generic one (compiled with --disable-nativedvc) works correctly though with these changes.
—
Reply to this email directly, view it on GitHub, or unsubscribe.
You are receiving this because you are subscribed to this thread.![]()
@vadz Sorry this PR is no longer wxMSW port specific!
—
Reply to this email directly, view it on GitHub, or unsubscribe.
You are receiving this because you are subscribed to this thread.![]()
@AliKet pushed 15 commits.
—
View it on GitHub or unsubscribe.
You are receiving this because you are subscribed to this thread.![]()
@AliKet pushed 11 commits.
—
View it on GitHub or unsubscribe.
You are receiving this because you are subscribed to this thread.![]()
@AliKet pushed 4 commits.
—
View it on GitHub or unsubscribe.
You are receiving this because you are subscribed to this thread.![]()
@vadz @paulcor I think there is a bug in
wxDataViewCheckIconTextRenderer()when using the nativewxDataViewCtrlcontrol inwxGTK3(I'm not sure because I'm not familiar with the (native)wxDataViewCtrldrawing code). i.e.: the check+icon+text is always drawn LTR. The generic one (compiled with--disable-nativedvc) works correctly though with these changes.
Fixed in d547521
Before:
dataview_rtl_bad.png (view on web)
After:
dataview_rtl_ok.png (view on web)
—
Reply to this email directly, view it on GitHub, or unsubscribe.
You are receiving this because you are subscribed to this thread.![]()
@AliKet pushed 1 commit.
—
View it on GitHub or unsubscribe.
You are receiving this because you are subscribed to this thread.![]()
@vadz approved this pull request.
This looks very nice, I really appreciate getting rid of the hacks in wxMSW and platform checks in the tests, thanks a lot for your work on this!
I'll merge this soon, if you agree that we should use GetConnectWidget(), we can/will change this later.
In include/wx/gtk/private/event.h:
> {
// origin in the upper right corner
GtkAllocation a;
- gtk_widget_get_allocation(win->m_wxwindow, &a);
+ gtk_widget_get_allocation(win->m_wxwindow ? win->m_wxwindow : win->m_widget, &a);
I wonder if we should do this here:
⬇️ Suggested change- gtk_widget_get_allocation(win->m_wxwindow ? win->m_wxwindow : win->m_widget, &a); + gtk_widget_get_allocation(win->GetConnectWidget(), &a);
?
—
Reply to this email directly, view it on GitHub, or unsubscribe.
You are receiving this because you are subscribed to this thread.![]()
Merged #25426 into master.
—
Reply to this email directly, view it on GitHub, or unsubscribe.
You are receiving this because you are subscribed to this thread.![]()