I was testing lldb-mi and I found a few issues with the break-insert command.
1. The constructor of CMICmdArgValString used in this command did not provide an option to set m_bHandleDirPaths. So the parsing will fail on command like "26-break-insert --thread-group i1 -f /home/abidh/demos/main.c:22". I added another parameter to the constructor to solve this issue.
2. The parameter -f does not take any argument. The code was thinking that it takes one which was causing an error.
After these changes the break-insert command is working fine for me. Please let me know if you see any problem with these change. Otherwise I will apply them.
break_insert.patch (3.95 KB)
I actually had a bunch of fixes and new features to upstream for lldb-mi, which I plan to do tomorrow.
Could you please hold off committing this patch till then?
Sure. Actually I also have a few other fixes. But I will hold on to them for now.
Are all the changes localized to lldb-mi or do they trickle down into other areas of LLDB? If the latter, it would be nice if there were a way to split them out into separate changelists. Otherwise I suspect it will be a difficult merge, and likely to stomp some other stuff that has gone in since the commit activity has picked up recently.
No, the changes would all be contained within lldb-mi as it just uses the LLDB public API.