Stefan Dotterweich <stefandotterweich@xxxxxx> wrote: > When using filters, the commit list shows not only commits matching > the filter criteria, but also boundary commits. When going through a > list of say, all commits changing the variable `foo`, often half of > the displayed commits are boundary commits. In this case the boundary > commits are of little interest. > > However, there is no way to hide them or quickly distinguish them from > the actual commits. Boundary commits can be identified by the white > color inside the circle, but that is not easily recognisable. On each > line you have to look at the circle color to identify the commit > type. This makes it hard to just quickly skim a list of commits, > especially when looking at dates and authors which are further to the > right. > > Therefore, to make boundary commits easier to recognise, display their > text in a different color. I would like to go one step further and not show the boundary commits at all. Why do we need them again? (I think this has been discussed before, but I can't find it right now. The only reference I could find is this thread: <https://public-inbox.org/git/571F6852.1070405@xxxxx/T/#u>, which doesn't explain *why* gitk shows the boundary commits in the first place.) In my opinion, when saying "gitk --author=foo", the list of commits in the top pane should look the same as the ouput of "git log --oneline --author=foo". -- Stefan Haller Berlin, Germany http://www.haller-berlin.de/
