The 2 attached files demonstrate the problem. To reproduce: tkdiff q.c.old q.c.new; touch View; touch Ignore White Spaces.
A tk_messageBox appears displaying Error 1. The log is attached.
Before accepting this report, I need to confirm the problem was actually encountered against the version it was reported against(V4.3.5). It was a legitimate error in the V4.3.1 timeframe, but was fixed in V4.3.2 and all releases since that time. Unfortunately, the error log (being nothing but a basic stacktrace) does not identify the software version being run, although there is a telltale line offset number for the failing statement (11) which does correlate with this being from a vintage earlier than the V4.3.5 indentified in the report.
For the record, the error had been caused when trying to repaint the intra-line difference highlighting among linepairs for any "change"-type hunks. It had failed to recognize that despite the existance of a legitimate difference at a "hunk" level, it was entirely possible that no differences might be present among the initial pairs of lines (ie. left .vs. right) for that hunk. Actually, this issue derives from an oddity in the method by which Diff chooses the lines being reported, with regard to its not always being a minimal specification. The fix was simply allowing for this "empty set" of highlight data.
Could you please verify the conditions of this report ?
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Before accepting this report, I need to confirm the problem was actually encountered against the version it was reported against(V4.3.5). It was a legitimate error in the V4.3.1 timeframe, but was fixed in V4.3.2 and all releases since that time. Unfortunately, the error log (being nothing but a basic stacktrace) does not identify the software version being run, although there is a telltale line offset number for the failing statement (11) which does correlate with this being from a vintage earlier than the V4.3.5 indentified in the report.
For the record, the error had been caused when trying to repaint the intra-line difference highlighting among linepairs for any "change"-type hunks. It had failed to recognize that despite the existance of a legitimate difference at a "hunk" level, it was entirely possible that no differences might be present among the initial pairs of lines (ie. left .vs. right) for that hunk. Actually, this issue derives from an oddity in the method by which Diff chooses the lines being reported, with regard to its not always being a minimal specification. The fix was simply allowing for this "empty set" of highlight data.
Could you please verify the conditions of this report ?
Hi Michael you are correct. It was not 4-3-5: I just tried that and the problem is gone.
Sorry for the noise. Cheers ... Duncan.
Just cleaning up unneeded ticket per conversation above