This issue seems to only happen on NetBSD - you only get a few undosets and then the program crashes. If you adjust the allocation you can get more undos but then you'll start seeing control characters and other random characters showing up all over the place.
This is likely because the memory that gets allocated for the undo is overlapping the memory for the buffer.
This issue seems to only happen on NetBSD - you only get a few `undoset`s and then the program crashes. If you adjust the allocation you can get more undos but then you'll start seeing control characters and other random characters showing up all over the place.
This is likely because the memory that gets allocated for the undo is overlapping the memory for the buffer.
So, I was testing out using an undo/redo trail instead of consecutive memory for undo/redo and it appears to have worked. Commit 91b91fedbd has the fix which is merged into version 1.5.
I've seen in other editors that instead of saving the buffer or a section of buffer that they save the action and have an undo of each action. That is a far better solution to this problem but the way it currently is written appears to be working so I can't complain.
So, I was testing out using an undo/redo _trail_ instead of consecutive memory for undo/redo and it appears to have worked. Commit 91b91fedbdffef34e73c7f1528ae5505b828e739 has the fix which is merged into version 1.5.
I've seen in other editors that instead of saving the buffer or a section of buffer that they save the _action_ and have an undo of each action. That is a far better solution to this problem but the way it currently is written appears to be working so I can't complain.
Kevin "The Nuclear" Bloom
This issue seems to only happen on NetBSD - you only get a few
undosets and then the program crashes. If you adjust the allocation you can get more undos but then you'll start seeing control characters and other random characters showing up all over the place.This is likely because the memory that gets allocated for the undo is overlapping the memory for the buffer.
So, I was testing out using an undo/redo trail instead of consecutive memory for undo/redo and it appears to have worked. Commit
91b91fedbdhas the fix which is merged into version 1.5.I've seen in other editors that instead of saving the buffer or a section of buffer that they save the action and have an undo of each action. That is a far better solution to this problem but the way it currently is written appears to be working so I can't complain.