Fix memory leak, prevent corrupt database, fix typos in qml

[zjeffer]

I added an option to CMakeLists.txt to enable address sanitizer (first commit). I ran it and noticed the following memory leak when closing the main window:

=================================================================
==143229==ERROR: LeakSanitizer: detected memory leaks

Direct leak of 24 byte(s) in 1 object(s) allocated from:
    #0 0x56a6832241b2 in operator new(unsigned long) (/home/zjeffer/git/notes/build/notes+0x2521b2) (BuildId: f3febbdce0342b087d8f64ae39d11f58ce6cd95c)
    #1 0x56a6833aadfb in MainWindow::MainWindow(QWidget*) /home/zjeffer/git/notes/src/mainwindow.cpp:132:20
    #2 0x56a6833a8184 in main /home/zjeffer/git/notes/src/main.cpp:96:16
    #3 0x7a8a82a34e07 in __libc_start_call_main /usr/src/debug/glibc/glibc/csu/../sysdeps/nptl/libc_start_call_main.h:58:16
    #4 0x7a8a82a34ecb in __libc_start_main /usr/src/debug/glibc/glibc/csu/../csu/libc-start.c:360:3
    #5 0x56a6830e5584 in _start (/home/zjeffer/git/notes/build/notes+0x113584) (BuildId: f3febbdce0342b087d8f64ae39d11f58ce6cd95c)

SUMMARY: AddressSanitizer: 24 byte(s) leaked in 1 allocation(s).

It's not much but it should be fixed either way, so I did that in the second commit. We have lots more (and much bigger) memory leaks btw, will take a look at the others too.

In the third commit, I just fixed some typos in qml files.

In the fourth commit, I ensured that #504 can never happen again, by simply returning if the user tries to move a note into the same folder that it's already in. I followed @guihkx steps and could not reproduce the issue after this commit.

Closes #504
Closes #234

[zjeffer]

There's an issue with clang-format, apparently it wants to format

emit itemDelegate()->sizeHintChanged(index);

into

emit itemDelegate() -> sizeHintChanged(index);

Investigating 🤔

EDIT: fixed locally by updating clang-format based on these instructions: https://bugreports.qt.io/browse/QTCREATORBUG-31858, but the action still fails, likely because the Github Action uses clang-format v18, where this bug occurs. I changed the ci/cd to always run all jobs, even if linting fails. @guihkx can you take a look at those changes?

[guihkx]

I changed the ci/cd to always run all jobs, even if linting fails

Not a big fan of that tbh... Can you try downgrading the runner used in the "Code Linting" job to ubuntu-22.04 and see if the issue goes away?

As for the other changes, I'm not on my Linux machine right now to test this properly, but I will do that as soon as I can.

[zjeffer]

@guihkx Done, it worked!

There's a problem with the snap build, seems like an authentication issue. @nuttyartist can you take a look?

[guihkx]

Done, it worked!

Nice.

There's a problem with the snap build, seems like an authentication issue

Hmm, I actually think that error comes from sg, not snap...

Do you wanna try to also use ubuntu-22.04 for the snap job?

It looks like the recent upgrade of ubuntu-latest to Ubuntu 24.04 will require a few changes on our side, but we don't have to do it right now, because the ubuntu-22.04 runner will not be deprecated anytime soon.

Note to self: actions/runner-images#9932

[nuttyartist]

Thanks, @zjeffer! I want to take some time to review this, hopefully today.

[guihkx]

The changes to this file are fine, I guess, but I can still reproduce the database corruption of #504 by following these reproduction steps.

[zjeffer]

I can't reproduce it anymore with those steps. Step 4 logs onMoveNodeRequested Can't move a node into the same parent in the console, indicating it won't try to move it into itself (which caused the bug).

Did you test it on a clean database?

[guihkx]

Did you test it on a clean database?

Yup. I'm not sure why you can't reproduce it, but here's a video if it helps:

broken.mp4

I also didn't do all steps because my terminal was already being spammed with qt.sql.qsqlquery: QSqlQuery::value: not positioned on a valid record at only step 2, and by then the db is already broken.

[zjeffer]

Ah, we're seeing 2 different bugs. The bug I fixed in this PR occured when moving a non-empty node into the folder it's already in. The bug you're triggering in your video happens when moving an empty note into a folder. It does indeed also trigger the same problem where folders then can't be deleted.
I'll look into that bug in another PR.