Tk Source Code

scrollbar-10.[12] fail on macOS as follows:

==== scrollbar-10.1 <MouseWheel> event on scrollbar FAILED
==== Contents of test case:

    pack [text .t -yscrollcommand {.s set}] -side left
    for {set i 1} {$i < 100} {incr i} {.t insert end "Line $i\n"}
    pack [scrollbar .s -command {.t yview}] -fill y -expand 1 -side left
    update
    focus -force .s
    event generate .s <MouseWheel> -delta -120
    after 200 {set eventprocessed 1} ; vwait eventprocessed
    .t index @0,0

---- Result was:
77.0
---- Result should have been (exact matching):
5.0
==== scrollbar-10.1 FAILED



==== scrollbar-10.2 <MouseWheel> event on scrollbar FAILED
==== Contents of test case:

    pack [text .t -xscrollcommand {.s set} -wrap none] -side top
    for {set i 1} {$i < 100} {incr i} {.t insert end "Char $i "}
    pack [scrollbar .s -command {.t xview} -orient horizontal] -fill x -expand 1 -side top
    update
    focus -force .s
    event generate .s <Shift-MouseWheel> -delta -120
    after 200 {set eventprocessed 1} ; vwait eventprocessed
    .t index @0,0

---- Result was:
1.120
---- Result should have been (exact matching):
1.4
==== scrollbar-10.2 FAILED

Seems I forgot to close this one.

Merged to core-8-6-branch and trunk on 2018-10-18.

At this point scrollbar.test runs 100% PASS on all 3 plaftorms (Win Vista, Linux Debian 8 and macOS 10.12).

Bugfix branch: bug-874dca4873

Fixed by [bb66444f7e].

Tests scollbar-10.1 and scrollbar-10.2 have to be split in the notAqua (== win|unix case) and the aqua case because on aqua the %D substitution for MouseWheel is a multiple of +/-1 whereas on Win/Linux it is a multiple of +/-120.

I believe I see what's going on here.

These two tests were added in [2d95721f] to check non regression of the fix for [1927212fff]. They are run on macOS because -constraints {win|unix} is true.