fix(diagnostic): show backtrace for deprecation warnings

[wookayin]

Problem: On nvim 11.0-dev, deprecation warnings due to an use of hard-deprecated APIs such as:

• vim.diagnostic.disable()
• vim.diagnostic.is_disabled()

etc. are not accompanied by backtrace information.

It makes difficult for users to figure out which lines or which plugins are still using deprecated APIs.

Solution: use backtrace = true in vim.deprecate() call.

[wookayin]

Note: The argument deprecate = false was first introduced in #26618, and I don't see any special reason to suppress backtrace here as well. For other places where vim.deprecate() is used usually, stacktrace information is not suppressed.

EDIT: seems it was intentional? https://github.com/neovim/neovim/pull/26618/files#r1429248469

[gpanders]

EDIT: seems it was intentional? https://github.com/neovim/neovim/pull/26618/files#r1429248469

That particular instance was intentional, because the backtrace there doesn't provide any useful information. It's not a call site of a specific function, it's warning about the use of sign_define for diagnostic signs, which could be happening anywhere in the user's config. The backtrace in this case would just be misleading.

[wookayin]

I see, thanks for the explanation. This might be too niche, but in my specific configs it was a plugin (neo-tree) rather than user config:

vim.diagnostic.is_disabled() is deprecated, use vim.diagnostic.is_enabled() instead. :help deprecated
Feature will be removed in Nvim 0.12
stack traceback:
        ...neovim/runtime/lua/vim/diagnostic.lua:1597: in function 'is_disabled'
        .../neo-tree.nvim/lua/neo-tree/utils/init.lua:244: in function 'get_diagnostic_counts'
        .../neo-tree.nvim/lua/neo-tree/setup/init.lua:45: in function <.../neo-tree.nvim/lua/neo-tree/setup/init.lua:44>
        [C]: in function 'pcall'
        .../neo-tree.nvim/lua/neo-tree/events/queue.lua:65: in function 'fire_event_internal'
        .../neo-tree.nvim/lua/neo-tree/events/queue.lua:100: in function <....neo-tree.nvim/lua/neo-tree/events/queue.lua:99>
        [C]: in function 'pcall'
        .../neo-tree.nvim/lua/neo-tree/utils/init.lua:126: in function 'debounce'
        .../neo-tree.nvim/lua/neo-tree/utils/init.lua:71: in function ''
        vim/_editor.lua: in function ''
        vim/_editor.lua: in function <vim/_editor.lua:0>

So I think backtrace information could be still useful for this particular case as well as for user configs?

[gpanders]

So I think backtrace information could be still useful for this particular case as well as for user configs?

For vim.diagnostic.disable() and vim.diagnostic.is_disabled() I agree. I don't know why the backtrace was omitted from those.

[wookayin]

That particular instance was intentional, because the backtrace there doesn't provide any useful information. It's not a call site of a specific function, it's warning about the use of sign_define for diagnostic signs, which could be happening anywhere in the user's config. The backtrace in this case would just be misleading.

This makes sense -- backtrace should be omitted for this. I've updated the PR accordingly.