Commit graph not using minimal number of columns

This is something that annoys me a lot whenever I'm maintaining a git repo with two branches - a main development one and a "fork" or alteration of the main one, which periodically merges changes from the main one while at the same time maintaining some additional feature.

When I display the commit graph with `git log --graph` or `gitk`, this type of graph shows up really nice, with two parallel "tracks" corresponding to the two branches with periodic merges between them. However, if for some reason I merge an early version of the main branch when there's already a newer commit in that branch, then the entire graph becomes an ugly mess that uses a lot of horizontal space, and I lose the visual representation of "two parallel tracks". This seems to be due to git relying on commit dates to decide which commit goes first, rather than relying solely on topological order.

Example: ``` mkdir TEST_git cd TEST_git git init -b main

touch A git add A git commit -m 'First commit'

git switch -c fork touch Z git add Z git commit -m 'Introduced feature Z'

for i in B C D E F; do
    git switch main
    touch "$i"
    git add "$i"
    git commit -m "Add $i"
    git switch fork
    git merge --no-edit main
done

git log --all --decorate --oneline --graph # ^ displays a pretty "two-track" graph

git switch main
for i in G H; do
    touch "$i"
    git add "$i"
    git commit -m "Add $i"
done
git switch fork
git merge --no-edit main~1

git log --all --decorate --oneline --graph # ^ displays a complete mess that doesn't resemble two tracks ```

This seems to happen because commit H was created after G but before merging G into `fork`, and can be mitigated by updating the committer date of H so that it is newer than the merge of G into `fork`, but that is not always an option. It also seems to go away after merging H into `fork`, but that won't work if H is part of a temporary/abandoned branch I don't plan to merge yet. And I found that calling `git log main --all ...` somehow fixes the graph as well. In these cases, the option `--date-order` actually seems to yield better graphs than the default (`--topo-order`), but that option is not ideal, since it may interleave commits from two separate branches, whereas the default will group commits "semantically".

What do you think? Do you agree that this sort of thing looks annoying? Do you think there might be a way to solve this kind of issue by improving the graph generation algorithm, so that it minimizes the number of columns in the graph (for example, preferring to display merge commits as late as possible)? Or is this actually intentional, because the developers consider that in this sort of situation it would be preferable to group all the merges together?