(re-adding list) On Mon, Jul 7, 2014 at 1:25 AM, Samuel Marshall <[email protected]> wrote: > From what I saw, the data isn’t being updated in sync with the window sizes. > This leads me to believe this is an issue with how quickly the window > properties are being updated vs how quickly the window is actually being > resized. I think that most of these issues would go away if you changed from > size_t over to a signed integer type of the same size. Memory corruption was > a thought I had but i found it was actually different expectations of some of > your global state variables instead.
I found with valgrind that info was accessing data that had been freed in a signal handler. For more explanation see e.g. https://www.gnu.org/software/libc/manual/html_node/Why-Block.html#Why-Block. I found it could even start the signal handler for resizing while it was in the middle of handling the last resize. There were a couple of other places as well where very small displays weren't handled well. I've got something that seems to work now: it might need looking again to check it is blocking signals everywhere it needs to. You can try getting the latest version from SVN (http://savannah.gnu.org/projects/texinfo) to see if it works.
