Skip to main content
aboutsummaryrefslogtreecommitdiffstats
AgeCommit message (Collapse)AuthorFilesLines
2020-11-25v4940r23v4940r23kS4_18_0_RC1I20201125-1800Eclipse Releng Bot1-1/+1
2020-11-25Bug 565969 - [Big Sur] Table selection drawn over text/checkbox whenLakshmi Shanmugam5-2/+38
SWT.EraseItem is hooked This fixes the incorrect un-focused selection background color seen in Eclipse Trees and Table in dark theme. alternateSelectedControlColor and secondarySelectedControlColor are deprecated in 10.14. The patch uses their respective replacements - selectedContentBackgroundColor and unemphasizedSelectedContentBackgroundColor Change-Id: Icdd0487cd28e92dae29540e51fe69e129f36d003 Signed-off-by: Lakshmi Shanmugam <lshanmug@in.ibm.com>
2020-11-25v4940r23v4940r23jI20201125-0600Eclipse Releng Bot1-1/+1
2020-11-25Bug 567616 - win32 tree with SWT.FULL_SELECTION hard to see on darkNiraj Modi1-5/+1
theme Change-Id: Ic3de470e1d2eb08844eafcefe93d5b9fbf5efe35 Signed-off-by: Niraj Modi <niraj.modi@in.ibm.com>
2020-11-24v4940r23v4940r23iY20201124-1200I20201124-1800Eclipse Releng Bot1-1/+1
2020-11-24Bug 564885 - [Dark] Toolbar has a white lineSravan Kumar Lakkimsetti1-1/+1
Added error correction to fillGradientRectangle method to fill the background completely Change-Id: If749204d426d808eef10df24ab4510bc3d8608f7 Signed-off-by: Sravan Kumar Lakkimsetti <sravankumarl@in.ibm.com>
2020-11-23v4940r23v4940r23hI20201124-0600I20201123-1800I20201123-0600Eclipse Releng Bot1-1/+1
2020-11-23Bug 567826 - [GTK] JVM crash on Display.dispose() due to ClipboardAlexandr Miloslavskiy3-3/+86
If 'SWT.Dispose' handler is a bit slow, user might be able to switch to other application and Ctrl+C there. When this happens, Linux calls 'ClipboardProxy.clearFunc()' with the following stack: Native: clipboard_unset() at gtk/gtkclipboard.c:782 Native: selection_clear_event_cb() at gtk/gtkclipboard.c:424 Native: <emit signal for 'GDK_SELECTION_CLEAR'> Native: gtk_widget_event_internal() at gtk/gtkwidget.c:7808 Native: gtk_main_do_event() at gtk/gtkmain.c:1860 Java: org.eclipse.swt.internal.gtk.GTK._gtk_main_do_event() Java: org.eclipse.swt.widgets.Display.eventProc() Java: org.eclipse.swt.internal.gtk.OS._g_main_context_iteration() Java: org.eclipse.swt.widgets.Display.readAndDispatch() Java: org.eclipse.swt.widgets.Display.release() If 'Callback' for 'clearFunc' is already disposed, JVM crashes. The solution is to dispose callback after 'Display.readAndDispatch()' in 'Display.release()'. This patch has a curious side effect: if 'SWT.Dispose' handler is slow, Shell is now visible (and eventually shows "not responding") for its duration. I understand that this is because 'gtk_clipboard_store()' also does some unexpected event processing. I consider this side effect as not a regression because the same behavior is seen when clipboard is not owned by SWT by the time of 'Display.dispose()'. Change-Id: I2cc11600f06030243e55d70f24b881dcb9b67503 Signed-off-by: Alexandr Miloslavskiy <alexandr.miloslavskiy@syntevo.com>
2020-11-20v4940r23v4940r23gY20201120-1200I20201122-1800I20201122-0600I20201121-1800I20201121-0600I20201120-1800Eclipse Releng Bot1-1/+1
2020-11-20Bug 568590 - [win32] Numbers in Spinner control are not vertically centredAlexandr Miloslavskiy2-2/+82
Adjust Spinner's text position to match that of Text. This patch also removes '+ border' workaround that is no longer needed since Bug 565679 which removed incorrect internal borders. Change-Id: I0361ce0f2b8d358db8fa5827cb16e61580c64b9d Signed-off-by: Alexandr Miloslavskiy <alexandr.miloslavskiy@syntevo.com>
2020-11-18v4940r23v4940r23fS4_18_0_M3I20201120-0320I20201119-0730I20201118-1800Eclipse Releng Bot1-1/+1
2020-11-18Bug 565969: [Big Sur] Table selection drawn over text/checkbox when ↵Lakshmi Shanmugam2-11/+36
SWT.EraseItem is hooked In Table/Tree, highlightSelectionInClipRect() call from Cocoa is ignored when SWT.EraseItem is hooked and selection is drawn in drawInteriorWithFrame_inView. On BigSur, calling highlightSelectionInClipRect in drawInteriorWithFrame_inView draws over the full row and not just the cellRect. This causes drawing over other cells content. Workaround is to draw the highlight background ourselves and not call highlightSelectionInClipRect to draw it. Change-Id: I9f2e5e93730a0d63221eb628807d0fd7567deba7
2020-11-18Disabled serialized uid warnings for snippets.Alexander Kurtakov1-1/+9
Change-Id: I18f29de664a17b5f90e25d64e42d959c5b686b53 Signed-off-by: Alexander Kurtakov <akurtako@redhat.com>
2020-11-18Remove unnecessary semicolon.Alexander Kurtakov1-1/+1
Change-Id: If9a282df5d22d557610e491bf679fd3704edf4fb Signed-off-by: Alexander Kurtakov <akurtako@redhat.com>
2020-11-18Bug 568771 - Disabled icon not being rendering properlyPaul D'Pong1-5/+21
- Added back the automatic graying of icons when setEnabled(false) is set - The way this is done is by having the default disabledImage to be a grayed version of the original image set for the tool item. This way we maintain the HiDPI fixes while allowing for the disable behavior. Change-Id: If0bfe963ffb806fa1b35797baa66ff10889048cb Signed-off-by: Paul D'Pong <sdamrong@redhat.com>
2020-11-17v4940r23v4940r23eY20201117-1200I20201118-0600I20201117-1800I20201117-0600Eclipse Releng Bot1-1/+1
2020-11-17Bug 568740 - [Win32] TextLayout renders underscore, strikeout and borderPaul Pazderski4-54/+157
only on last line Each of the three styles underscore, strikeout and border are rendered separately in there own method. TextLayout implementation for windows tries to optimize rendering of those styles by drawing adjacent styles in one a single call. (for border it also looks much better) It will test if two adjacent style parts (separate for underscore, strikeout and border) are 'adherent' which seem to mean they look the same. This can happen for once if the user supplies suboptimal styles, i.e. the same or an equal style instance for adjacent ranges instead of one style for the whole range at once. But even if adjacent styles are different it will still try to optimize rendering if a specific substyle is equal, e.g. two adjacent styles got different fonts but both use underline with the same color. In this case it will draw the underline in one call for both ranges. Now for the failing case. If a style expands over more than one line TextLayout will internally split those styles on line delimiter because it renders per line. This will include a separate StyleItem for the line delimiter to mark it as lineBreak which inherits the style of the range before. An example. The two lines content "abc\n123" with an underline over the hole range produce internally three StyleItems all with same style. While drawing the first line it will see underline for the "abc" part but postpone rendering because the adjacent style for "\n" has the same underline style. But line breaks are rendered differently so the postponed underline drawing will never happen. Change-Id: Ic19f13a559c79d148f62fc7944101dd59ae17dde Signed-off-by: Paul Pazderski <paul-eclipse@ppazderski.de>
2020-11-16v4940r23v4940r23dI20201116-1800Eclipse Releng Bot1-1/+1
2020-11-16Bug 568861 - [GTK] Gtk-CRITICAL in Label.setFontDescription()Alexandr Miloslavskiy1-7/+9
Change-Id: I6b4611a9c56091f72bff1fec87325873f7f37c64 Signed-off-by: Alexandr Miloslavskiy <alexandr.miloslavskiy@syntevo.com>
2020-11-13v4940r23v4940r23cY20201113-1200I20201116-0600I20201115-1800I20201115-0600I20201114-1800I20201114-0740I20201114-0600I20201113-1800Eclipse Releng Bot1-1/+1
2020-11-13Bug 568777 - `GC.drawText(.., transparent = true)` uses wrong font forSravan Kumar Lakkimsetti2-0/+74
short texts in 2020-12 M2 - Second implementation without reverting the drawText optimization Change-Id: I0e7826d352fbc9ec85253956844a7a16d1cb5126
2020-11-13Bug 568033 - IllegalArgumentException in StyledText.replaceTextRange()Alexandr Miloslavskiy3-28/+164
Fixed to no longer throw exception if \r\n were on different lines already. Even though the test snippet reproduces exactly the stack as reported by users, I do feel that users might be having a different problem. This feeling mostly comes from the fact that we only received reports on Windows in combination with 'WM_IME_COMPOSITION', while the snippet reproduces the problem on all 3 platforms. For this reason, I also changed 'DefaultContent.isValidReplace()' to throw 3 different exceptions with additional info to help future debugging if this patch doesn't solve the problem. I didn't find any outside uses of 'DefaultContent.isValidReplace()', so I think that it's OK to remove 'protected' and change signature. Change-Id: I6d0234414e2cad637e029f557e4c9ead213f1db3 Signed-off-by: Alexandr Miloslavskiy <alexandr.miloslavskiy@syntevo.com>
2020-11-13Bug 568550 - Java editor inactive after refactor to extract methodAndrey Loskutov2-0/+113
This partly reverts commit 6371f380811160e02b570e6441e42938de3db6da for bug 563555, because it caused a regression in JDT UI. The revert only will restore the previous behavior for shells with SWT.ON_TOP flag set. Change-Id: I3167f57fe1e7c14914b1a43c7e731240ba1a9110 Signed-off-by: Andrey Loskutov <loskutov@gmx.de> Also-by: Alexandr Miloslavskiy <alexandr.miloslavskiy@syntevo.com>
2020-11-13Bug 565899 - [HIDPI][MAC][RETINA] Gutter with linenumbers on scrambledSravan Kumar Lakkimsetti1-85/+15
in (Java) editor Refactored getRepresentation code to make it more generic. Change-Id: Icfb2f98b171513dae5855dce981c6040d2c7ecc5 Signed-off-by: Sravan Kumar Lakkimsetti <sravankumarl@in.ibm.com>
2020-11-13Bug 565899 - [HIDPI][MAC][RETINA] Gutter with linenumbers on scrambledSravan Kumar Lakkimsetti1-12/+4
in (Java) editor - corrected getrepresetation_100 method itself. rather than error checking the in internal_new_gc - This contains the fix. Refactoring part will be added as separate commit Change-Id: I8dac337ae14ed60f077294ea9350d69a6087a95c
2020-11-12v4940r23v4940r23bI20201112-1800Eclipse Releng Bot1-1/+1
2020-11-12Bug 568407 - [Linux] Toolbar icons are double size on hi-res (200%Paul D'Pong1-2/+10
scaling) - enter/leave notify events - For GTK3, changed gtk_image_set_from_pixbuf to gtk_image_set_from_surface in the enter and leave notify event handlers Change-Id: I59671ffb22a431f39fbba18c4b478c13001faf24 Signed-off-by: Paul D'Pong <sdamrong@redhat.com>
2020-11-12Bug 568475 - NPE problems with comparing colorsJeeeyul Lee4-0/+16
Change-Id: I20090e16b1b3eea24f4f62f83017aa2341e90396 Signed-off-by: Jeeeyul Lee<jeeeyul@gmail.com>
2020-11-11v4940r23v4940r23aI20201111-1800Eclipse Releng Bot1-1/+1
2020-11-11Enforce UTF-8 in ant javac call in buildFragment.xml.Alexander Kurtakov1-1/+1
Ensures that we don't get compile errors like "error: unmappable character (0xC3) for encoding US-ASCII" on systems that don't default to UTF. Change-Id: I2a2762a8a5eff12efc31a9b5bb028671f00a6a1e Signed-off-by: Alexander Kurtakov <akurtako@redhat.com>
2020-11-11Bug 568073 - [GTK4] Port Cursor snippetsPaul D'Pong1-29/+27
- Put guard on gtk_window_get_position & gtk_window_move as they no longer are available in GTK4. - Removed GTK4 guard in gtk_button_press_event in Shell since it is only called in GTK3 Change-Id: Ica94dc081ad92b3a4f5ea2610cf1665ec77d7eeb Signed-off-by: Paul D'Pong <sdamrong@redhat.com>
2020-11-10v4940r23v4940r23Y20201111-0210Y20201110-1200I20201110-1800Eclipse Releng Bot1-1/+1
2020-11-10v4940r23Eclipse Releng Bot3-4/+4
2020-11-10Bug 568401 - [GTK4] Port SWT Button snippetsPaul D'Pong5-9/+39
- Fixed Button snippet crashes due to previous GTK3.99.3 update (GtkCheckButton no longer inhertied from GtkToggleButton) - Replaced gtk_widget_get_can_default with gtk_widget_get_receives_default to allow for Snippet345 to work Change-Id: If916948019427310cf14d25f496110069d538021 Signed-off-by: Paul D'Pong <sdamrong@redhat.com>
2020-11-10Bug 567485 - [GTK4] Cleanup old porting code - Cleaned up GTKPaul D'Pong7-366/+494
Change-Id: I1d1fd0981183e05f2f236e8dc6b7d1ebb8594430 Signed-off-by: Paul D'Pong <sdamrong@redhat.com>
2020-11-09v4940r22v4940r22I20201109-1800Eclipse Releng Bot1-1/+1
2020-11-09v4940r22Eclipse Releng Bot3-4/+4
2020-11-09Bug 568648 - [GTK4] Update to GTK3.99.4Paul D'Pong2-5/+5
Change-Id: I8451b9260724adf08dece22b9c33839fdcdb281d Signed-off-by: Paul D'Pong <sdamrong@redhat.com>
2020-11-09Bug 565899 - [HIDPI][MAC][RETINA] Gutter with linenumbers on scrambledSravan Kumar Lakkimsetti1-0/+8
in (Java) editor - The problem is getRepresentation_100() can return 200% image representation of there are no 100% image representations available. We need to find out whether we are getting the proper image representation when we are creating the imageContext. Change-Id: I5a129cd849688b2dd5089fbc907868e34c299255 Signed-off-by: Sravan Kumar Lakkimsetti <sravankumarl@in.ibm.com>
2020-11-06v4940r21v4940r21Y20201109-1200Y20201109-0220Y20201108-1200Y20201107-1200Y20201106-1200I20201108-1800I20201107-1800I20201107-0750I20201106-1800I20201106-0710Eclipse Releng Bot1-1/+1
2020-11-06v4940r21Eclipse Releng Bot3-4/+4
2020-11-06Bug 567485 - [GTK4] Cleanup old porting code - GdkDevice/GdkDisplayPaul D'Pong9-78/+151
- Organized the GdkDevice functions - Removed gdk_device/gdk_display related implicit declaration warnings when building SWT with GTK4 - Removed the os.h guards for these functions and replaced them with dynamic javadoc flags Change-Id: I973a5557b61f267bef8d2e704b45405b821e61dc Signed-off-by: Paul D'Pong <sdamrong@redhat.com>
2020-11-05v4940r20v4940r20I20201105-1800Eclipse Releng Bot1-1/+1
2020-11-05v4940r20Eclipse Releng Bot3-4/+4
2020-11-05Bug 568407 - [Linux] Toolbar icons are double size on hi-res (200%Paul D'Pong11-90/+57
scaling) - Added GTK4 guard to gtk_image_set_from_pixbuf which doesn't seem to correctly do Hi-DPI scaling. - Reverted back to gtk_image_set_from_surface for GTK3 Change-Id: I3a2aea2e3a5df55bd34b9d09468ab2e7d2be4be1 Signed-off-by: Paul D'Pong <sdamrong@redhat.com>
2020-11-05v4940r19Y20201105-1200Eclipse Releng Bot1-1/+1
2020-11-05v4940r19Eclipse Releng Bot3-4/+4
2020-11-05Bug 567371 - [GTK4] Removal of blocking dialog functions - LicensePaul D'Pong1-0/+13
Change-Id: I66eea58e62a4378a9fbfb616c04387d650dae7d4 Signed-off-by: Paul D'Pong <sdamrong@redhat.com>
2020-11-05Bug 567371 - [GTK4] Removal of blocking dialog functionsPaul D'Pong11-196/+275
- Added SyncDialogUtil to be used to transform nonblocking dialog calls to blocking ones - Cleanup FontDialog's open function - Added NO_ guards in os.h Change-Id: Ifb87b8d6fa421c783dfb71b9131e119e666c263b Signed-off-by: Paul D'Pong <sdamrong@redhat.com>
2020-11-05Bug 565305 - [GTK] SWT List.getItemHeight inconsistent heights dependingPaul D'Pong2-25/+29
on interface font size Add gtk_widget_create_pango_layout to get a test layout to find the pango_layout_get_pixel_size. Then determine the padding of the item with respect to the cell renderer & BASE_ITEM_PADDING. BASE_ITEM_PADDING is a number found through testing and represent the base padding that the cell renderer adds. This value is true event if gtk_cell_renderer_set_padding set padding to 0. Added test case that checks that having an empty vs filled list does not change the value returned by getItemHeight. Change-Id: Ie3bee546c555e8d7d907cec74121e30a389b0338 Signed-off-by: Paul D'Pong <sdamrong@redhat.com>

Back to the top