Woops, posted too soon (damn mouse) I guess I'll be more specific. Qt 3.3.8b on the main GUI guitone is linked against Qt 4.6.2_2d3d3e5 Yes other apps work happily when 'mixed' like this :-)
Qt 3.3.8b :-)
I could reproduce this issue with Qt 4.7.3 on Mac OS X. As minimal use case just clone the mtn repository and execute $ ./guitone exec path/to/monotone.mtn showFileDiff 'doc/monotone.texi' 5bfdedba0571477ec41755bc60db55402bb5d5f0 76e40d0c7a53b8934a32f2bee6b20f6e9bba59d2 From what I have seen this is something inside Qt (again), because the diff parsing and applying is through from my side and guitone is just waiting for Qt to render the model appropriately. I have to rebuild with a debug version of Qt at first to see what is really going on: #0 0x0000000100b2fec8 in QAlgorithmsPrivate::qReverse<int*> () #1 0x0000000100b300c2 in QAlgorithmsPrivate::qMerge<int*, int const, qLess<int> > () #2 0x0000000100b300de in QAlgorithmsPrivate::qMerge<int*, int const, qLess<int> > () #3 0x0000000100b30196 in QAlgorithmsPrivate::qStableSortHelper<int*, int, qLess<int> > () #4 0x0000000100b30196 in QAlgorithmsPrivate::qStableSortHelper<int*, int, qLess<int> > () #5 0x0000000100b301a7 in QAlgorithmsPrivate::qStableSortHelper<int*, int, qLess<int> > () #6 0x0000000100b30196 in QAlgorithmsPrivate::qStableSortHelper<int*, int, qLess<int> > () #7 0x0000000100b30196 in QAlgorithmsPrivate::qStableSortHelper<int*, int, qLess<int> > () #8 0x0000000100b30196 in QAlgorithmsPrivate::qStableSortHelper<int*, int, qLess<int> > () #9 0x0000000100b301a7 in QAlgorithmsPrivate::qStableSortHelper<int*, int, qLess<int> > () #10 0x0000000100b30196 in QAlgorithmsPrivate::qStableSortHelper<int*, int, qLess<int> > () #11 0x0000000100b30196 in QAlgorithmsPrivate::qStableSortHelper<int*, int, qLess<int> > () #12 0x0000000100b28f19 in QSortFilterProxyModelPrivate::sort_source_rows () #13 0x0000000100b29139 in QSortFilterProxyModelPrivate::create_mapping () #14 0x0000000100b297d8 in QSortFilterProxyModel::hasChildren () #15 0x0000000100adc5c2 in QTreeViewPrivate::hasVisibleChildren () #16 0x0000000100ae5eb4 in QTreeViewPrivate::layout () #17 0x0000000100ae9acd in QTreeView::doItemsLayout () #18 0x0000000100ae3a99 in QTreeView::verticalOffset () #19 0x0000000100aa82f0 in QAbstractItemViewPrivate::contentsOffset () #20 0x0000000100608b7c in QWidgetPrivate::paintBackground ()Status: Accepted
Browse the revision 76e40d0c7a53b8934a32f2bee6b20f6e9bba59d2 of the monotone repository in guitone. (web: https://code.monotone.ca/p/monotone/source/commit/76e40d0c7a53b8934a3 2f2bee6b20f6e9bba59d2/) Richard Hopkins reported for that: * as soon as i double click monotone.texi to open the diff window, the PRivate Bytes counter shoots up and continues going to 1.5 GB, then closing the diff window the Private Bytes stays at 1.5GB * then double cliking the file below it (std_hooks.lua), Private Bytes started going to down from 1.5GB to 20MB (normal), then back up again to 1.6 GB, then back down to 20MB then up to 40MB, then the contents appeared in the diff window * then closing that file, and diffing the next one "func/trust_hooks...", Private Bytes goes down to 20MB and shows the contents
Hi AJK! Sorry for the late response - what Qt version are you running? Qt 4.7.x by any chance? Unfortunately there is a bug (either in guitone or in Qt) that is quite hard to track down. Upstream is notified, but hasn't done anything so far: https://bugreports.qt.nokia.com/browse/QTBUG-19261 Feel free to upvote this. Thanks for the report!Status: Accepted
The SSL error most likely origins from missing OpenSSL binaries - not that I forgot to package these. Are libeay32.dll and ssleay32.dll available in the installation directory? The cause of the actual crash itself is unfortunately not shown in the log. Could you please raise the log verbosity to debug (needs a debug build) or alternatively run it under gdb (cygwin needed here probably)? Sorry that I'm not of much help here, the guitone support lacks quite a bit. I have tons of other stuff to do at the moment...
Steps to reproduce the problem: 1. Checkout net.venge.monotone with monotone 2. Modify README in any way, such as add some lines to the end 3. Open net.venge.monotone workspace in guitone Expected result: The Workspace window appears showing all of the files, and README with a status of modified Actual result: Crash System and version information: Windows XP SP3, standard releases of monotone 1, guitone 1rc5
Additional note: Workspace is updated properly, but the segfault kills guitone, a bit annoying to have to reopen guitone after this.
Steps to reproduce the problem: 1.Open a workspace 2.Try to update it 3. Expected result: An updated workspace Actual result: Tail end of strace... clock_gettime(CLOCK_MONOTONIC, {2057138, 382481122}) = 0 poll([{fd=3, events=POLLIN}, {fd=8, events=POLLIN}, {fd=7, events=POLLIN}, {fd=36, events=POLLIN}], 4, 12724) = 1 ([{fd=36, revents=POLLIN}]) read(7, 0x829b028, 4096) = -1 EAGAIN (Resource temporarily unavailable) read(36, "\0"..., 1) = 1 waitpid(27535, [{WIFEXITED(s) && WEXITSTATUS(s) == 0}], WNOHANG) = 27535 ioctl(-1, FIONREAD, [0]) = -1 EBADF (Bad file descriptor) ioctl(-1, FIONREAD, [0]) = -1 EBADF (Bad file descriptor) close(29) = 0 close(37) = 0 close(36) = 0 clock_gettime(CLOCK_MONOTONIC, {2057138, 382750868}) = 0 clock_gettime(CLOCK_MONOTONIC, {2057138, 383164446}) = 0 --- SIGSEGV (Segmentation fault) @ 0 (0) --- +++ killed by SIGSEGV +++ System and version information: Linux 2.6.38.2, SMP PREEMPT i686 AMD Phenom(tm) II X4 945 Processor AuthenticAMD GNU/Linux Slackware 13.37 with QT4 but running KDE 3.5 on QT3 Compiled with: LDFLAGS/CFLAGS/CXXFLAGS ="-O2 -march=i486 -mtune=i686"
Guitone should get basic support for database migrations. Right now the mtn process just terminates if it finds a database that is too old to use. guitone should instead prompt the user to migrate the database and re-load the database / workspace after it has migrated the database,.