Ratatui-Image Terminal Screenshots
Screenshots from various terminal emulators
Last updated: ${timestamp}
EOF
# Add screenshots to index using local images
for terminal in $terminal_list; do
echo "
" >> pages/index.html
echo "
$terminal
" >> pages/index.html
# Check if image exists and add to HTML
if [ -f "pages/images/screenshot-${terminal}.png" ]; then
echo "
" >> pages/index.html
else
echo "
No screenshot available
" >> pages/index.html
fi
echo "
" >> pages/index.html
done
echo "
" >> pages/index.html
- name: Deploy to GitHub Pages
uses: peaceiris/actions-gh-pages@v3
with:
personal_token: ${{ secrets.PAT }}
external_repository: benjajaja/ratatui-image-screenshots
publish_dir: ./pages
publish_branch: gh-pages
ratatui-image-8.1.1/.github/workflows/fast-forward.yml 0000644 0000000 0000000 00000001302 10461020230 0021077 0 ustar 0000000 0000000 name: fast-forward
on:
issue_comment:
types: [created, edited]
jobs:
fast-forward:
if: ${{ contains(github.event.comment.body, '/merge')
&& github.event.issue.pull_request }}
runs-on: ubuntu-latest
permissions:
contents: write
pull-requests: write
issues: write
steps:
- name: Fast forwarding
uses: sequoia-pgp/fast-forward@v1
with:
merge: true
# To reduce the workflow's verbosity, use 'on-error'
# to only post a comment when an error occurs, or 'never' to
# never post a comment. (In all cases the information is
# still available in the step's summary.)
comment: never
ratatui-image-8.1.1/.github/workflows/pull_request.yml 0000644 0000000 0000000 00000001412 10461020230 0021226 0 ustar 0000000 0000000 name: pull-request
on:
pull_request:
types: [opened, reopened, synchronize]
jobs:
check-fast-forward:
runs-on: ubuntu-latest
permissions:
contents: read
# We appear to need write permission for both pull-requests and
# issues in order to post a comment to a pull request.
pull-requests: write
issues: write
steps:
- name: Checking if fast forwarding is possible
uses: sequoia-pgp/fast-forward@v1
with:
merge: false
# To reduce the workflow's verbosity, use 'on-error'
# to only post a comment when an error occurs, or 'never' to
# never post a comment. (In all cases the information is
# still available in the step's summary.)
comment: never
ratatui-image-8.1.1/.github/workflows/release.yml 0000644 0000000 0000000 00000001044 10461020230 0020123 0 ustar 0000000 0000000 on:
push:
tags:
- "v*.*.*"
workflow_dispatch:
name: Upload Release Assets
jobs:
build:
uses: ./.github/workflows/ci.yaml
post-build:
name: Release
needs: nix-build
runs-on: ubuntu-latest
steps:
- name: Create Github Release
id: create_release
uses: actions/create-release@v1
env:
GITHUB_TOKEN: ${{ secrets.PAT }}
with:
tag_name: ${{ github.ref }}
release_name: Release ${{ github.ref }}
draft: false
prerelease: false
ratatui-image-8.1.1/.gitignore 0000644 0000000 0000000 00000000033 10461020230 0014350 0 ustar 0000000 0000000 /target
log
.direnv
result
ratatui-image-8.1.1/CHANGELOG.md 0000644 0000000 0000000 00000016071 10461020230 0014202 0 ustar 0000000 0000000 # Changelog
All notable changes to this project will be documented in this file.
## [Unreleased]
## [8.1.1] - 2025-12-18
* [Chafa](https://hpjansson.org/chafa/) support
The Halfblocks protocol now supports chafa, if the feature is enabled.
This "ASCII Art"-like renderer is a million times better than the basic Halfblocks implementation.
The library is loaded dynamically, if the feature is enabled, at runtime.
That means no linking acrobacies are required, it should "just work" if the user has libchafa installed on their system.
It could still be "wrapped", see how the flake.nix handles `LD_LIBRARY_PATH` to magically include this in the `ratatui-image` binary.
This is still gated behind a feature, because calling chafa uses `unsafe`.
* Improved Halfblocks renderer
Picks upper block, lower block, or space, depending on pixel colors.
This gives some vague feedback when the output is rendered without colors, e.g. in insta testing snapshots.
# [8.0.2] - 2025-09-03
* Performance improvements in kitty and iterm2 protocols
* Use base64-simd
* Picker::from_query_stdio returns a halfblocks picker when: no capabilities, stdio query
timeout, or no font-size received.
* Add screenshot test suite to CI
* Add tokio support to ThreadProtocol
* Updated to rust 2024
* Add support for terminals: BobCat, warp
# [8.0.1] - 2025-05-30
Fix halfblocks not having the same size as other protocols.
# [8.0.0] - 2025-05-18
Fix `TextSizingProtocol` detection (was incorrectly detecting support on Foot terminal).
Separate "stdio query response" from "capability", as they do not match one-to-one.
Moved `Capability` into the `picker` module, as the interpretation is done there.
# [7.0.0] - 2025-05-18
`Picker` has a new field / method `capabilities`, which return the precise capabilities detected by `Picker::from_query_stdio`.
Since the struct now holds a `Vec`, `Picker` is no longer derives `Copy` (but still derives `Clone`).
This uncovered some unnecessary copying on a method that previously consumed self for no good reason.
Add `Picker::from_query_stdio_with_options` to specifically query for [Text Sizing Protocol](https://sw.kovidgoyal.net/kitty/text-sizing-protocol/#detecting-if-the-terminal-supports-this-protocol).
Not used in this library, but rather an optional convencience feature so that stdio must not be queried twice by some programs.
The capability, if detected, is `Capability::TextSizingProtocol` and can be read from `Picker::capabilities()`.
# [6.0.0] - 2025-05-17
`Image::new` uses a non-mutable borrow.
The mutability was only required for the Kitty protocol, necessary to track wether the image has already been transmitted or only needs placement.
Overall this change should remove the need to use locks when using threads or tokio.
# [5.0.0] - 2025-03-01
Add `StatefulProtocol::size_for`, that can be used to get the size that an image will be rendered to.
This allows positioning the image before it has been rendered, for example centering the image widget with the usual ratatui layout options.
ThreadImage and ThreadProtocol work with `ResizeRequest` and `ResizeResponse` instead of some tuples.
They internally track an ID so that a response for a stale area is discarded correctly.
Huge internal refactor that removes duplicated code across image protocols.
- `Errors` variant case names have been fixed.
- `StatefulProtocol` becomes a struct (was enum).
- `StatefulProtocol::area` which returned the last rendered area has been removed, use `size_for` for accurate results.
- `StatefulProtocol` methods has some parameters removed.
- All protocols except kitty lose their `Stateful...` struct implementation, as one struct can share both protocol implementations.
# [4.2.0] - 2024-12-31 đ
Fix Sixel and iTerm2 not working with tmux.
# [4.1.2] - 2024-12-30
Add a release job to the CI that makes a github release when a `v*.*.*` is pushed.
The tag itself and pushing to crates.io is done locally.
# [4.1.0] - 2024-12-23 đ
### Transparency support for Kitty and iTerm2
The image data gains an alpha channel, as well as the background color, which is now transparent by default when the protocol is not Sixel.
The area behind the image is cleared with the control sequences `ECH`, `CUD`, and `CUU` repeatedly for each row and column.
This is not particularly efficient, but it works in most terminals.
`DECERA`, which should erase the entire rectangle with only one sequence, is not implemented correctly (or not at all) in some major terminals.
Sixels could also support transparency, as the spec directly supports it.
A palette color could be set to transparent, and `icy_sixel` in fact supports this, however this is after the fact that the image has been encoded from an API perspective.
In other words, we could set a palette color index to be transparent, but we don't control that this color index would match any transparency of the input image.
This is something that would have to be added to `icy_sixel`.
### Resize: Scale
This feature is brought to you by