Why do these two nearly-identical bindings take different time to execute?

Question

I like to open Vim with many buffers open, and switch between them using plugins. Because I work a lot in C++, I like to have a way to quickly switch between header files and their corresponding source files and vice-versa. With some work, I have created these two bindings which essentially do the same thing: find a buffer with the same name as the currently opened buffer, but with a .c* extension if my current buffer is a header file, and with a .h* extension if my current buffer is a source.

My puzzlement derives from the fact that the binding that opens the other file using a split generally executes instantly, while the one that replaces the current buffer takes a second or so before it terminates. However, running the :buffer command manually is instant. Why does this happen?

This question seems similar to this one, but as far as I can see my bindings don't really shadow any control codes.

Here are the two bindings (the top one opens a split, the second one replaces):

nmap <silent><leader>v :execute 'vert sbuffer' '\<'.expand("%:t:s,\\v\\.h(..)?$,.X123X,:s,\\v\\.c(..)?$,.h,:s,.X123X$,.c,")<CR>
nmap <silent><leader>c :execute 'buffer' '\<'.expand("%:t:s,\\v\\.h(..)?$,.X123X,:s,\\v\\.c(..)?$,.h,:s,.X123X$,.c,")<CR>

Answer 1

Very often the delay are due to the interval ('timeoutlen') that Vim wait before deciding which mapping to trigger when you have overlapping mappings (in your case: <leader>c and <leader>cd)

After hitting <leader>c Vim waits timeoutlen microseconds for an additional d before deciding that your intention was <leader>c.

Very often there are numerous mappings that could conflict. A good way to detect the problem is to use the nmap command that allow you to list all the normal mode mappings that share a given prefix.

In your case:

nmap <leader>c