[vim/vim] Send DECRQM sequences based on version response (PR #19938)

14 views
Skip to first unread message

Foxe Chen

unread,
Apr 8, 2026, 8:19:42 PM (8 days ago) Apr 8
to vim/vim, Subscribed

Should fix #19852

You can view, comment on, or merge this pull request online at:

  https://github.com/vim/vim/pull/19938

Commit Summary

  • f0e5daf send decrqm sequences based on version response

File Changes

(4 files)

Patch Links:

—
Reply to this email directly, view it on GitHub.
You are receiving this because you are subscribed to this thread.Message ID: <vim/vim/pull/19938@github.com>

h_east

unread,
Apr 9, 2026, 4:32:39 AM (8 days ago) Apr 9
to vim/vim, Subscribed
h-east left a comment (vim/vim#19938)

Thank you for your PR. Please confirm two points.
(This review was created by Claude Opus 4.6 with 1M context)

1. tpr_set_by_termresponse should be TRUE

File: src/term.c (init_term_props)

    term_props[TPR_DECRQM].tpr_set_by_termresponse = FALSE;

TPR_DECRQM is set to TPR_YES inside handle_version_response() (i.e., based on the DA2 response), so tpr_set_by_termresponse should be TRUE, not FALSE.

When tpr_set_by_termresponse is FALSE, init_term_props(FALSE) (called during terminal re-initialization) will NOT reset TPR_DECRQM, which means a stale value from a previous terminal could persist.

For reference, TPR_CURSOR_STYLE, TPR_CURSOR_BLINK, TPR_UNDERLINE_RGB, and TPR_MOUSE — all of which are set in handle_version_response() — have tpr_set_by_termresponse = TRUE.

2. GNU screen consideration

File: src/term.c (handle_version_response)

The xterm fallback block (where term_props[TPR_MOUSE].tpr_status is TPR_UNKNOWN) sets TPR_DECRQM = TPR_YES unconditionally. GNU screen (arg[0] == 83) matches an earlier branch and doesn't reach this fallback, so it won't get TPR_YES — which is probably correct since GNU screen likely doesn't support DECRQM. Just want to confirm this is intentional.

—
Reply to this email directly, view it on GitHub, or unsubscribe.
You are receiving this because you are subscribed to this thread.Message ID: <vim/vim/pull/19938/c4212757322@github.com>

h_east

unread,
Apr 9, 2026, 4:33:53 AM (8 days ago) Apr 9
to vim/vim, Subscribed

@h-east commented on this pull request.

In runtime/doc/builtin.txt:

> @@ -11989,6 +11989,7 @@ terminalprops()						*terminalprops()*
 		   underline_rgb	whether |t_8u| works **
 		   mouse		mouse type supported
 		   kitty		whether Kitty terminal was detected
+		   decrqm		whether sending DECRQM sequences work.
⬇️ Suggested change 
-		   decrqm		whether sending DECRQM sequences work.
+		   decrqm		whether sending DECRQM sequences work

—
Reply to this email directly, view it on GitHub, or unsubscribe.
You are receiving this because you are subscribed to this thread.Message ID: <vim/vim/pull/19938/review/4080849917@github.com>

Foxe Chen

unread,
Apr 10, 2026, 9:03:47 AM (7 days ago) Apr 10
to vim/vim, Push

@64-bitman pushed 1 commit.

  • 02900c9 send decrqm sequences based on version response

—
View it on GitHub or unsubscribe.
You are receiving this because you are subscribed to this thread.Message ID: <vim/vim/pull/19938/before/5c1ba4ed51fcb2ad965d9f72549ccd9799fb918b/after/02900c9a0b71a26198f1b86c0903858edeed5f98@github.com>

Foxe Chen

unread,
Apr 10, 2026, 9:05:44 AM (7 days ago) Apr 10
to vim/vim, Push

@64-bitman pushed 1 commit.

  • a627976 send decrqm sequences based on version response

—
View it on GitHub or unsubscribe.

You are receiving this because you are subscribed to this thread.Message ID: <vim/vim/pull/19938/before/02900c9a0b71a26198f1b86c0903858edeed5f98/after/a627976a2992e5b9e2d7a163030503ff41e90bc4@github.com>

Foxe Chen

unread,
Apr 10, 2026, 9:31:44 AM (7 days ago) Apr 10
to vim/vim, Push

@64-bitman pushed 1 commit.

  • e008d34 send decrqm sequences based on version response

—
View it on GitHub or unsubscribe.

You are receiving this because you are subscribed to this thread.Message ID: <vim/vim/pull/19938/before/a627976a2992e5b9e2d7a163030503ff41e90bc4/after/e008d34bd1bc4ec8a90978d6148f69840ec83a9d@github.com>

h_east

unread,
Apr 15, 2026, 11:48:43 AM (2 days ago) Apr 15
to vim/vim, Subscribed
h-east left a comment (vim/vim#19938)

Here are additional review comments.
(Supported by Claude Opus 4.6 with 1M context)

1. Missing TPR_DECRQM for PuTTY / SecureCRT

File: src/term.c (handle_version_response, version == 136 branch)

The PuTTY branch (arg[0] == 0 && version == 136) and SecureCRT branch (arg[0] == 1 && version == 136) don't set TPR_DECRQM. PuTTY supports DECRQM, so it should probably get TPR_YES. Without it, TPR_DECRQM stays TPR_UNKNOWN and DECRQM sequences won't be sent.

2. Missing TPR_DECRQM for libvterm / Konsole

File: src/term.c (handle_version_response, version == 100 || version == 115 branch)

The libvterm/Konsole branch doesn't set TPR_DECRQM either. Konsole supports DECRQM. For libvterm (used inside Vim's :terminal), it may not be needed, but the two share the same branch — might need to distinguish them or add a comment explaining why it's left as unknown.

3. arg[1] >= 2500 branch covers a wide range of terminals

File: src/term.c (handle_version_response)

if (arg[1] >= 2500)
{
    term_props[TPR_UNDERLINE_RGB].tpr_status = TPR_YES;
    term_props[TPR_DECRQM].tpr_status = TPR_YES;
}

This matches Gnome terminal (1;3801;0, 1;4402;0, 1;2501;0), xfce4-terminal (1;2802;0), and potentially other terminals. Do all of them handle DECRQM properly? The comment says "Assuming any version number over 2500 is not an xterm", which is quite broad.

4. need_flush and DECRQM sending moved outside #ifdef FEAT_TERMRESPONSE

File: src/term.c (handle_version_response)

The original send_decrqm_modes() was called inside #ifdef FEAT_TERMRESPONSE in vim_main2(). In the new code, the DECRQM sending and need_flush / out_flush() are placed outside the #ifdef FEAT_TERMRESPONSE block. If DECRQM sending depends on the termresponse feature, it should stay inside the #ifdef.

5. Guard termcap_active && cur_tmode == TMODE_RAW removed

File: src/term.c

The original send_decrqm_modes() had:

if (termcap_active && cur_tmode == TMODE_RAW)
{
    // send DECRQM ...
}

The new code in handle_version_response() sends DECRQM without this guard. If it's guaranteed that handle_version_response() is always called in raw mode, a brief comment noting that would be helpful.

—
Reply to this email directly, view it on GitHub, or unsubscribe.
You are receiving this because you are subscribed to this thread.Message ID: <vim/vim/pull/19938/c4253473982@github.com>

Foxe Chen

unread,
Apr 15, 2026, 12:07:20 PM (2 days ago) Apr 15
to vim/vim, Push

@64-bitman pushed 1 commit.

  • 2b46723 send decrqm sequences based on version response

—
View it on GitHub or unsubscribe.

You are receiving this because you are subscribed to this thread.Message ID: <vim/vim/pull/19938/before/e008d34bd1bc4ec8a90978d6148f69840ec83a9d/after/2b46723635617986e2178564f807920640e4fa65@github.com>

Christian Brabandt

unread,
Apr 15, 2026, 2:04:14 PM (2 days ago) Apr 15
to vim/vim, Subscribed
chrisbra left a comment (vim/vim#19938)

hm, this causes quite a few test failures. Let me re-trigger CI again.

—
Reply to this email directly, view it on GitHub, or unsubscribe.
You are receiving this because you are subscribed to this thread.Message ID: <vim/vim/pull/19938/c4254340153@github.com>

Foxe Chen

unread,
Apr 15, 2026, 2:26:03 PM (2 days ago) Apr 15
to vim/vim, Push

@64-bitman pushed 1 commit.

  • 59af2b3 Change decrqm property from 'u' to 'y'

—
View it on GitHub or unsubscribe.

You are receiving this because you are subscribed to this thread.Message ID: <vim/vim/pull/19938/before/2b46723635617986e2178564f807920640e4fa65/after/59af2b37c95700841267d4f078ed5f494f0d6432@github.com>

Foxe Chen

unread,
Apr 15, 2026, 2:26:25 PM (2 days ago) Apr 15
to vim/vim, Subscribed
64-bitman left a comment (vim/vim#19938)

hm, this causes quite a few test failures. Let me re-trigger CI again.

Should be fixed

—
Reply to this email directly, view it on GitHub, or unsubscribe.
You are receiving this because you are subscribed to this thread.Message ID: <vim/vim/pull/19938/c4254463142@github.com>

Foxe Chen

unread,
Apr 15, 2026, 3:23:59 PM (2 days ago) Apr 15
to vim/vim, Push

@64-bitman pushed 2 commits.

  • acb5cfb send decrqm sequences based on version response
  • 539d94a Change decrqm property from 'u' to 'y'

—
View it on GitHub or unsubscribe.

You are receiving this because you are subscribed to this thread.Message ID: <vim/vim/pull/19938/before/59af2b37c95700841267d4f078ed5f494f0d6432/after/539d94a6701b52447b8023cf2b769cffc8c75067@github.com>

Christian Brabandt

unread,
Apr 16, 2026, 4:39:29 PM (9 hours ago) Apr 16
to vim/vim, Subscribed
chrisbra left a comment (vim/vim#19938)

there still seem to be some screen dump failures :(

—
Reply to this email directly, view it on GitHub, or unsubscribe.
You are receiving this because you are subscribed to this thread.Message ID: <vim/vim/pull/19938/c4263254772@github.com>

Foxe Chen

unread,
Apr 16, 2026, 4:48:29 PM (9 hours ago) Apr 16
to vim/vim, Subscribed
64-bitman left a comment (vim/vim#19938)

there still seem to be some screen dump failures :(

I'll try fixing them when I can (soon hopefully), it is exam season for me :/

—
Reply to this email directly, view it on GitHub, or unsubscribe.
You are receiving this because you are subscribed to this thread.Message ID: <vim/vim/pull/19938/c4263308246@github.com>

Christian Brabandt

unread,
Apr 16, 2026, 4:57:11 PM (9 hours ago) Apr 16
to vim/vim, Subscribed
chrisbra left a comment (vim/vim#19938)

no worries, take your time and good luck for your exams 🤞

—
Reply to this email directly, view it on GitHub, or unsubscribe.
You are receiving this because you are subscribed to this thread.Message ID: <vim/vim/pull/19938/c4263359373@github.com>

Reply all
Reply to author
Forward
0 new messages