Skip to content

chore(deps): bump the opentui group with 2 updates - #23

Merged
ozandogrultan merged 1 commit into
mainfrom
dependabot/bun/opentui-818d1cb307
Nov 17, 2026
Merged

ozandogrultan merged 1 commit into
mainfrom
dependabot/bun/opentui-818d1cb307

Conversation

@dependabot

@dependabot dependabot Bot commented on behalf of github Nov 16, 2026

Copy link
Copy Markdown
Contributor

Bumps the opentui group with 2 updates: @opentui/core and @opentui/solid.

Updates @opentui/core from 0.8.2 to 0.8.4

Release notes

Sourced from @​opentui/core's releases.

Release v0.8.4

A chat log often puts many CodeRenderable blocks in a ScrollBoxRenderable with viewportCulling: true. With streaming: true and drawUnstyledText: false, a block that gets new content keeps showing its old highlighted text until the new highlight is ready, so the reader never sees a flash of plain text. But a block starts its highlight from its update hook, and a frame calls no hooks on a block that the scroll box culls. In 0.8.3, a culled block therefore kept its old text in its buffer, and layout measured that old text. The scroll box reported a stale scrollHeight until the block scrolled back into view.

In 0.8.4, a block that gets new content first asks the scene whether the hooks of the next frame reach it. If they do, the block keeps its old highlight, as before. If they do not, as for a culled, hidden, or detached block, the block loads the new plain text at once, so layout measures the new content. It does not paint that text until the highlight is ready. The highlight starts in the first frame that reaches the block again. C hosts can ask the same question with ot_scene_reaches_hooks. It answers with the branch selection that the next frame uses.

Added

  • native: ot_scene_reaches_hooks tells if the hooks of a frame reach a node. It selects culled branches as the next frame does. See C.

Fixed

  • core: A culled streaming CodeRenderable with drawUnstyledText: false now measures new content immediately. It does not paint that text until its highlight is ready. See Code.

API changes and documentation: https://opentui.com/docs/releases/0.8.4

Full Changelog: anomalyco/opentui@v0.8.3...v0.8.4

Release v0.8.3

Over ssh, sshd drops COLORTERM. A terminal that draws 24-bit color but sends TERM=xterm-256color without COLORTERM therefore got every color mapped to the 256-color palette. In WezTerm and xterm over ssh, a navy #23344b turned into a gray #3a3a3a. In 0.8.2, only a kitty XTVERSION reply turned on 24-bit color. In 0.8.3, when COLORTERM does not settle the question, the startup queries also ask XTGETTCAP RGB and Tc. A positive reply turns on 24-bit color. So does an XTVERSION reply from Ghostty, WezTerm, foot, XTerm, iTerm2, or tmux, because tmux converts 24-bit colors for its client itself. The replies come back through ssh, so this works without COLORTERM. Inside GNU Screen, OpenTUI sends neither query, because Screen drops 24-bit colors and prints the query bodies as text. When nothing answers, the old TERM rule still applies. Alacritty and mosh answer none of these queries, so they still need COLORTERM=truecolor.

The width method decides how many cells an emoji sequence takes. When OpenTUI picks a different method than the terminal uses, the line overflows or leaves gaps. Before, OpenTUI kept the unicode method in terminals that do not report mode 2027, even though many of them draw a sequence one code point at a time. tmux 3.6 and later draws a sequence as one cluster but got wcwidth. An inherited TERM_PROGRAM=ghostty also forced unicode-wide in other terminals. Now the method comes from one place, with the strongest evidence first:

  1. OPENTUI_FORCE_*, also over SSH.
  2. The multiplexer.
  3. The DECRPM 2027 reply. An unrecognized or permanently reset mode means wcwidth, except in kitty and iTerm2.
  4. XTVERSION.
  5. The environment, where TERM wins over an inherited TERM_PROGRAM.
Terminal Before After
Alacritty, xterm, urxvt, GNOME Terminal, VS Code, Zellij, foot (wcswidth) unicode wcwidth
tmux 3.6+ wcwidth unicode
kitty started from Ghostty unicode-wide unicode
Ghostty over SSH, libghostty unicode unicode-wide

No method matches tmux when variation-selector-always-wide is off. Unicode is the closest. If the new choice is wrong for your terminal, set OPENTUI_FORCE_UNICODE or OPENTUI_FORCE_WCWIDTH. These settings still win.

Before this release, a host that started an unmodified OpenTUI app could not easily make it load a different native build. setRenderLibPath() needs code, and OTUI_ASSET_ROOT needs the whole asset tree under one root. OPENTUI_LIB_PATH names one library file, under both Bun and Node:

OPENTUI_LIB_PATH=../opentui/packages/native/lib/x86_64-linux/libopentui.so bun run src/index.ts

A relative path resolves against the working directory. setRenderLibPath() takes precedence over the variable, and the variable takes precedence over OTUI_ASSET_ROOT and the platform package. If the file is missing or does not load, the error names the variable. Core now also checks the context ABI version when it loads a library, so a library built for another context ABI fails at load and not at a later call.

A host that embeds a terminal often has to know where the terminal's view is. For example, it can pass an unused wheel step to an outer scroll view or draw a scrollbar. Counting scroll steps does not give the answer. Output and resizes keep the view on its history rows, a scrollback clear jumps to the live screen, and the alternate screen has no history. EmbeddedTerminalRenderable.getScroll() reads the current view from the emulator, including the moves that the emulator makes on its own:

function scrollStatus(terminal: EmbeddedTerminalRenderable): string {
</tr></table> 

... (truncated)

Commits
  • d4a2af2 Release v0.8.4
  • f7e5763 native: select a culled branch as the next frame will
  • b6ab797 native: report whether a frame's hooks reach a node
  • feecf2c core: keep Code highlights only where the scene runs its hooks
  • 07697c5 core: measure culled Code content without waiting for a highlight
  • b9e9d46 Release v0.8.3
  • 6adfa20 native: ask the terminal for 24-bit color before falling back (#1941)
  • e4ad586 native: export the audio engine in opentui.h (#1935)
  • cd23e75 native: report the embedded terminal viewport (#1934)
  • d8e5461 core: load the native library from OPENTUI_LIB_PATH (#1933)
  • Additional commits viewable in compare view

Updates @opentui/solid from 0.8.2 to 0.8.4

Release notes

Sourced from @​opentui/solid's releases.

Release v0.8.4

A chat log often puts many CodeRenderable blocks in a ScrollBoxRenderable with viewportCulling: true. With streaming: true and drawUnstyledText: false, a block that gets new content keeps showing its old highlighted text until the new highlight is ready, so the reader never sees a flash of plain text. But a block starts its highlight from its update hook, and a frame calls no hooks on a block that the scroll box culls. In 0.8.3, a culled block therefore kept its old text in its buffer, and layout measured that old text. The scroll box reported a stale scrollHeight until the block scrolled back into view.

In 0.8.4, a block that gets new content first asks the scene whether the hooks of the next frame reach it. If they do, the block keeps its old highlight, as before. If they do not, as for a culled, hidden, or detached block, the block loads the new plain text at once, so layout measures the new content. It does not paint that text until the highlight is ready. The highlight starts in the first frame that reaches the block again. C hosts can ask the same question with ot_scene_reaches_hooks. It answers with the branch selection that the next frame uses.

Added

  • native: ot_scene_reaches_hooks tells if the hooks of a frame reach a node. It selects culled branches as the next frame does. See C.

Fixed

  • core: A culled streaming CodeRenderable with drawUnstyledText: false now measures new content immediately. It does not paint that text until its highlight is ready. See Code.

API changes and documentation: https://opentui.com/docs/releases/0.8.4

Full Changelog: anomalyco/opentui@v0.8.3...v0.8.4

Release v0.8.3

Over ssh, sshd drops COLORTERM. A terminal that draws 24-bit color but sends TERM=xterm-256color without COLORTERM therefore got every color mapped to the 256-color palette. In WezTerm and xterm over ssh, a navy #23344b turned into a gray #3a3a3a. In 0.8.2, only a kitty XTVERSION reply turned on 24-bit color. In 0.8.3, when COLORTERM does not settle the question, the startup queries also ask XTGETTCAP RGB and Tc. A positive reply turns on 24-bit color. So does an XTVERSION reply from Ghostty, WezTerm, foot, XTerm, iTerm2, or tmux, because tmux converts 24-bit colors for its client itself. The replies come back through ssh, so this works without COLORTERM. Inside GNU Screen, OpenTUI sends neither query, because Screen drops 24-bit colors and prints the query bodies as text. When nothing answers, the old TERM rule still applies. Alacritty and mosh answer none of these queries, so they still need COLORTERM=truecolor.

The width method decides how many cells an emoji sequence takes. When OpenTUI picks a different method than the terminal uses, the line overflows or leaves gaps. Before, OpenTUI kept the unicode method in terminals that do not report mode 2027, even though many of them draw a sequence one code point at a time. tmux 3.6 and later draws a sequence as one cluster but got wcwidth. An inherited TERM_PROGRAM=ghostty also forced unicode-wide in other terminals. Now the method comes from one place, with the strongest evidence first:

  1. OPENTUI_FORCE_*, also over SSH.
  2. The multiplexer.
  3. The DECRPM 2027 reply. An unrecognized or permanently reset mode means wcwidth, except in kitty and iTerm2.
  4. XTVERSION.
  5. The environment, where TERM wins over an inherited TERM_PROGRAM.
Terminal Before After
Alacritty, xterm, urxvt, GNOME Terminal, VS Code, Zellij, foot (wcswidth) unicode wcwidth
tmux 3.6+ wcwidth unicode
kitty started from Ghostty unicode-wide unicode
Ghostty over SSH, libghostty unicode unicode-wide

No method matches tmux when variation-selector-always-wide is off. Unicode is the closest. If the new choice is wrong for your terminal, set OPENTUI_FORCE_UNICODE or OPENTUI_FORCE_WCWIDTH. These settings still win.

Before this release, a host that started an unmodified OpenTUI app could not easily make it load a different native build. setRenderLibPath() needs code, and OTUI_ASSET_ROOT needs the whole asset tree under one root. OPENTUI_LIB_PATH names one library file, under both Bun and Node:

OPENTUI_LIB_PATH=../opentui/packages/native/lib/x86_64-linux/libopentui.so bun run src/index.ts

A relative path resolves against the working directory. setRenderLibPath() takes precedence over the variable, and the variable takes precedence over OTUI_ASSET_ROOT and the platform package. If the file is missing or does not load, the error names the variable. Core now also checks the context ABI version when it loads a library, so a library built for another context ABI fails at load and not at a later call.

A host that embeds a terminal often has to know where the terminal's view is. For example, it can pass an unused wheel step to an outer scroll view or draw a scrollbar. Counting scroll steps does not give the answer. Output and resizes keep the view on its history rows, a scrollback clear jumps to the live screen, and the alternate screen has no history. EmbeddedTerminalRenderable.getScroll() reads the current view from the emulator, including the moves that the emulator makes on its own:

function scrollStatus(terminal: EmbeddedTerminalRenderable): string {
</tr></table> 

... (truncated)

Commits

Dependabot will resolve any conflicts with this PR as long as you don't alter it yourself. You can also trigger a rebase manually by commenting @dependabot rebase.


Dependabot commands and options

You can trigger Dependabot actions by commenting on this PR:

  • @dependabot rebase will rebase this PR
  • @dependabot recreate will recreate this PR, overwriting any edits that have been made to it
  • @dependabot show <dependency name> ignore conditions will show all of the ignore conditions of the specified dependency
  • @dependabot ignore <dependency name> major version will close this group update PR and stop Dependabot creating any more for the specific dependency's major version (unless you unignore this specific dependency's major version or upgrade to it yourself)
  • @dependabot ignore <dependency name> minor version will close this group update PR and stop Dependabot creating any more for the specific dependency's minor version (unless you unignore this specific dependency's minor version or upgrade to it yourself)
  • @dependabot ignore <dependency name> will close this group update PR and stop Dependabot creating any more for the specific dependency (unless you unignore this specific dependency or upgrade to it yourself)
  • @dependabot unignore <dependency name> will remove all of the ignore conditions of the specified dependency
  • @dependabot unignore <dependency name> <ignore condition> will remove the ignore condition of the specified dependency and ignore conditions

Bumps the opentui group with 2 updates: [@opentui/core](https://github.com/anomalyco/opentui/tree/HEAD/packages/core) and [@opentui/solid](https://github.com/anomalyco/opentui/tree/HEAD/packages/solid).


Updates `@opentui/core` from 0.8.2 to 0.8.4
- [Release notes](https://github.com/anomalyco/opentui/releases)
- [Commits](https://github.com/anomalyco/opentui/commits/v0.8.4/packages/core)

Updates `@opentui/solid` from 0.8.2 to 0.8.4
- [Release notes](https://github.com/anomalyco/opentui/releases)
- [Commits](https://github.com/anomalyco/opentui/commits/v0.8.4/packages/solid)

---
updated-dependencies:
- dependency-name: "@opentui/core"
  dependency-version: 0.8.4
  dependency-type: direct:production
  update-type: version-update:semver-patch
  dependency-group: opentui
- dependency-name: "@opentui/solid"
  dependency-version: 0.8.4
  dependency-type: direct:production
  update-type: version-update:semver-patch
  dependency-group: opentui
...

Signed-off-by: dependabot[bot] <support@github.com>
@dependabot dependabot Bot added dependencies Pull requests that update a dependency file javascript Pull requests that update javascript code labels Nov 16, 2026
@ozandogrultan
ozandogrultan merged commit 1c087fd into main Nov 17, 2026
6 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

dependencies Pull requests that update a dependency file javascript Pull requests that update javascript code

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant