https://bugs.kde.org/show_bug.cgi?id=488624
--- Comment #6 from JATothrim <jarmo.tii...@gmail.com> ---
> Going the address this directly sorry for the delay unfortunately I find
> myself as the only one doing work with this code. My attention has been else
> where lately. In particular dealing with the odd error on startup that I
> managed to reproduce. So far this hasn't happened for me but looks like an
> issue with Diff3Line::getString itself. Look at it now.
No need for sorry, I understand. I see you as one of the important people as
best described by xkcd 2347. kdiff3 somehow has ended up as the *only*
usable+sane three-way merge tool with GUI) for me, so I was going to give you
code on this time, rather than just complaining. I wrote an fix-up commit to
correct the bug. Will take a day to set-up to publish a fork with the commit on
gitlab.
Ah, I see you already did a fix on master. I would slightly reconsider this
approach:
(gdb) frame 13
(gdb) p range
$1 = (HistoryRange &) @0x7fffffffcc10: {
start = {lineA = {static invalid = -1, mLineNumber = {m_t = -1}}, lineB =
{static invalid = -1, mLineNumber = {m_t = 61}}, lineC = {static invalid = -1,
mLineNumber = {m_t = -1}}, bAEqC = false, bBEqC = false, bAEqB = false,
bWhiteLineA = true, bWhiteLineB = false, bWhiteLineC = true, pFineAB =
std::shared_ptr<const DiffList> (empty) = {get() = 0x0}, pFineBC =
std::shared_ptr<const DiffList> (empty) = {get() = 0x0}, pFineCA =
std::shared_ptr<const DiffList> (empty) = {get() = 0x0}, mLinesNeededForDisplay
= 1, mSumLinesNeededForDisplay = 0}, end = non-dereferenceable iterator for
std::list, startIdx = 64, endIdx = -1}
My fix was just going to select any of lineA, lineB or lineC (prioritizing
nearest lineC as in original code) after the the exact selection would have
failed. Unless it's possible that all *three* selector are invalid at same,
yours solution would be still required.
--
You are receiving this mail because:
You are watching all bug changes.
