| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
| |
Somebody wanted a very old MLP file I posted years ago, restores the
public URL.
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
|
|
|
|
| |
Some of these I don't use at all anymore, and others I use too
infequently to be built all the time.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
New features:
* A new :reader mode has been added which Tridactyl binds can run on.
If you prefer the old mode, it is accessible with :reader --old.
We've left gr bound to the old mode for now but you can update it
with :bind gr reader
* With :set tabsort mru, tab IDs within Tridactyl are now determined
by recency of use almost everywhere. This means, e.g. that you can
use 2gt to switch to the second most recently used tab (so identical
to <C-^>), 3gt to the third most recently used tab, etc.
* yq bind added to display current URL as a QR code, with q bind in
visual mode. See :text2qr for details
* :set keyboardlayoutforce true will now force Tridactyl to use, by
default, the US keyboard layout for keybinds, even if you're using a
different layout. Useful if you use multiple layouts but don't want
Tridactyl binds to move, or if you use a layout such as a Cyrillic
one where the default binds are difficult to use. Change the forced
layout with :set keyboardlayoutbase
* Experimental support for opening arbitrary pages in the sidebar has
been added. See :help sidebartoggle and :help sidebaropen for more
details
* :autocmd and related commands now have completions
* :open and related commands now have completions for searchurls and
previous searches
* Quality of life improvements for tab groups
* -x flag added to :hint to exclude CSS selectors from hints
* With the command line open, <C-o>t opens a new tab in the background
for the selected completion - especially useful with :back
completions
* :hint -C [selectors] added for custom hint modes while including the
default selectors
* ;gd rapid hint mode with discarded tabs added
* :undo now has fuzzy matching
* :source can now read from the command line and the clipboard with
--strings and --clipboard
* :tab{grab,push} now respect tabopenpos
Bug fixes:
* :find will now find start from the current viewport
* Comments now work in multiline commands in RC files
* :editor now works on Gradio apps (please spare me when AI takes over
the world)
* :winmerge completions are more readable with many windows
* :urlincrement now operates on the decoded URI so e.g. %20 doesn't
become %21
* ;m and ;M for Google reverse-image search now work again
* The midnight theme no longer has a white background in FF102+
* Container colours fixed for FF108+
* gi no longer goes to the start of the search box on Google
* modeindicator border is now configurable
* Unbindable default binds should no longer be possible
* Config should no longer reset when interacting with help page
Miscellaneous:
* :jsua command added to preserve "user action" intent when using
browser binds, needed for triggering certain Web Extension APIs such
as browser.sidebarAction.open(). See :help jsua for more information
* tri.hinting_content.hintElements function added for advanced usage
in :js - see the source for documentation
* Groundwork for multi-browser support added
* All tests now run using GitHub actions
* Docs are now built and hosted on our website
* Native messenger now supports Arm-based Macs with :native version at
least 0.4.0
* Native messenger 0.4.1 checks %XDG_CONFIG_HOME% on Windows
* when_feature.sh [feature] helper script added to repo to find the
earliest possible version a feature appeared in Tridactyl
Signed-off-by: aszlig <aszlig@nix.build>
|
|
|
|
|
|
|
|
|
|
|
|
| |
A lot has changed between 1.5.38, but mostly around the integrated
editor (which I don't use) and circumvention of content security
policies on certain sites.
I might want to drop this plugin at some point, since it does a whole
lot more than what I personally want (which is simply: use a different
CSS for a *very* small and selected amount of sites).
Signed-off-by: aszlig <aszlig@nix.build>
|
|
|
|
|
|
|
|
|
|
|
|
| |
Most changes are about installing the native messaging app and packaging
on various distributions and OS, additionally it also tries to support
more browsers.
In addition to that, there is now a way to set mpv arguments via the
plugin's new options page. We however want to set this via Nix, so I'm
completely ignoring that feature and didn't include it in our ff2mpv.py.
Signed-off-by: aszlig <aszlig@nix.build>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This was introduced in Firefox 120.0 and essentially implements the
following specification:
https://privacycg.github.io/gpc-spec/
Similar to Do-Not-Track, it implements a header (Sec-GPC) that expresses
the persons preference to the site of whether the site may sell/share
the browsers interaction.
Since Do-Not-Track pretty much is ignored on most sites that formerly
adhered to it, I expect a similar fate for GPC, but maybe this time
things work out differently (crossing fingers).
Signed-off-by: aszlig <aszlig@nix.build>
|
|
|
|
|
|
|
| |
We're managing extensions via Nix, so there is no need to check whether
extension updates are available.
Signed-off-by: aszlig <aszlig@nix.build>
|
|
|
|
| |
Just seems more detailed than neofetch.
|
|
|
|
| |
Finally, wifi works.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This is mainly because of the following build failure that seems to have
been fixed upstream:
widgets/emojiregistry.cpp: In member function 'EmojiRegistry::Category EmojiRegistry::startCategory(QStringRef) const':
widgets/emojiregistry.cpp:40:10: error: 'uint32_t' is not a member of 'std'; did you mean 'wint_t'?
40 | std::uint32_t ucs;
| ^~~~~~~~
| wint_t
widgets/emojiregistry.cpp:44:9: error: 'ucs' was not declared in this scope
44 | ucs = QChar::surrogateToUcs4(in[0], in[1]);
| ^~~
widgets/emojiregistry.cpp:46:9: error: 'ucs' was not declared in this scope
46 | ucs = in[0].unicode();
| ^~~
widgets/emojiregistry.cpp:51:9: error: 'ucs' was not declared in this scope
51 | if (ucs == 0x200d)
| ^~~
widgets/emojiregistry.cpp:53:9: error: 'ucs' was not declared in this scope
53 | if (ucs == 0xfe0f)
| ^~~
widgets/emojiregistry.cpp:55:9: error: 'ucs' was not declared in this scope
55 | if (ucs <= 0x39 && in.length() > 2 && (ucs >= 0x30 || ucs == 0x2a || ucs == 0x23) && in[1].unicode() == 0xfe0f
| ^~~
widgets/emojiregistry.cpp:61:38: error: 'ucs' was not declared in this scope
61 | auto lb = ranges_.lower_bound(ucs);
| ^~~
Of course, the second reason is that the current version seems to be way
too stable and I want to have a fresh round of new bugs.
Signed-off-by: aszlig <aszlig@nix.build>
|
|
|
|
|
|
|
|
|
|
| |
This fixes the following evaluation error:
error: The option `krb5.kerberos' can no longer be used since it's
been removed. The option `krb5.kerberos' has been moved to
`security.krb5.package'.
Signed-off-by: aszlig <aszlig@nix.build>
|
| |
|
| |
|
|
|
|
|
|
|
|
| |
Not only were there a lot of smaller fixes in the meantime, but latest
master now properly supports the (not so) new (anymore) path
interpolation syntax.
Signed-off-by: aszlig <aszlig@nix.build>
|
|
|
|
| |
Cos wifi dongles will always dongle.
|
|
|
|
| |
Without it, you can't view the trash in pcmanfm
|
|
|
|
| |
Let's not talk about why.
|
|
|
|
|
| |
Recommended by aszlig, due to the infrequency this machine actually gets
updated with.
|
| |
|
| |
|
|
|
|
| |
Seems like this changed from 301 to 302 now.
|
|
|
|
|
|
|
|
| |
The patch doesn't apply cleanly anymore because a new quirk flag was
added (NVME_QUIRK_FORCE_NO_SIMPLE_SUSPEND), so I rebased the patch
accordingly.
Signed-off-by: aszlig <aszlig@nix.build>
|
| |
|
|
|
|
| |
The former is unmaintained.
|
|
|
|
|
|
|
|
|
|
|
| |
While those .orig files might be useful for post-mortem analysis of a
rebase or merge, I never actually use them and rather resort to things
such as reflog to find the old tree/file.
Having those .orig files around just adds unnecessary noise, so let's
avoid that.
Signed-off-by: aszlig <aszlig@nix.build>
|
|
|
|
|
| |
Not sure if this will fix it, but at the moment gunnr isn't booting
after the update to 23.11.
|
| |
|
|
|
|
|
|
|
| |
This release just includes fixes for "Firefox Accounts" rebranding, no
real feature changes.
Signed-off-by: aszlig <aszlig@nix.build>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Changes since 1.5.33 minus bugfixes and CI stuff:
* Added an option to use adoptedStyleSheets API (Chrome/ium 73+,
Firefox 101+), instead of using DOM elements
* Combine style & editor settings
* Add an option to update only enabled styles
* Auto-retry style finder on network failure
* Faster import/undo
* Highlight the Save button in unsaved styles
* Improve the write-style UI a bit
* Show errors when importing the backup
* Show instant inject option in Firefox
* Speed up opening the manager (a bit)
* Speed up opening the multi-sectioned editor with lots of sections
* Support fractional numbers with Alt-key increment
* Use new userstyles.org API & uso-archive url
* Show autosaved draft's css code prior to confirming + auto-remove
the autosaved draft when saving
* Apply eyedropper color immediately
* Open installer for .user.less
* Remember all editor options
* Remember searchMode in manager
The full release notes can be found at:
https://github.com/openstyles/stylus/releases
Signed-off-by: aszlig <aszlig@nix.build>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
A recent nixpkgs change[1] added an explicit requireSigning attribute,
which in turn sets MOZ_REQUIRE_SIGNING to a zero-width string.
The change also changed the check in the Firefox wrapper to check
whether the browser derivation exposes a requireSigning attribute, which
it doesn't.
Since I'm busy with other things and just want my machine configuration
to evaluate, I quickly worked around this by re-adding the
requireSigning (and allowAddonSideload for that matter) to the
derivation attributes on our side.
[1]: https://github.com/NixOS/nixpkgs/commit/c4c81ac8a219e8725dae514f49c
Signed-off-by: aszlig <aszlig@nix.build>
|
|
|
|
|
|
| |
Removed due to a lack of maintenance:
http://github.com/nixos/nixpkgs/commit/022ee4d701878c185785ce72dd436639dd687199
The ELK NixOS test continues to exist.
|
| |
|
| |
|
| |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The "thunderbird-wayland" attribute has been an alias for a long time
and with a recent cleanup[1] of old aliases, the alias got dropped and
we now get an eval error due to a missing attribute.
Since "thunderbird-wayland" already was an alias, it should be safe to
just use "thunderbird".
[1]: https://github.com/NixOS/nixpkgs/commit/0c51bf975b5be87375d73aa5f3ba21542c240635
Signed-off-by: aszlig <aszlig@nix.build>
Cc: @sternenseemann
|
|
|
|
|
|
| |
This reverts commit e08ff60d61bb18dcd14a47f2ee6dec66ad8cbd91.
The CVE fixes have long been upstreamed.
|
|
|
|
|
|
|
|
|
|
| |
The module options have been restructured in a way that the attribute
set for defining the actual sites is now in a dedicated "sites" option
instead of eg. services.mediawiki directly, so the checks of whether we
should include this for testing always return true since
services.mediawiki is *always* a non-empty attribute set.
Signed-off-by: aszlig <aszlig@nix.build>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This one never was in use and it was WIP code to do a more fleshed out
mailserver configuration that should match an Ansible deployment of a
mailserver we had back then at OpenLab.
The machine was never in use (which is apparent from its configuration)
and I even *added* it to Vuizvui from "LaberNix" (Vuizvui's predecessor)
in 915e56fb4453b0701a423b0c96fb145318162ffd probably just for the sake
of completeness.
Signed-off-by: aszlig <aszlig@nix.build>
|
|
|
|
|
|
|
|
|
|
|
| |
In 75ea723caea05056810dad9160bc93e571c77985, I removed a bunch of
modules in a series of fixups but forgot to do a final eval check.
This broke evaluation for all the machines and tests because the
module-list.nix file is not only needed for all the machines but also
for figuring out which tests we need to run for which machine config.
Signed-off-by: aszlig <aszlig@nix.build>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The former no longer exists and the latter has a new hardware and it's
self-managed by the owner now (which is always the best IMHO).
With both machines gone, we no longer need the "managed" profile and the
BFQ module that was used for the old hardware of Brawndo was no longer
used anywhere else. Same with the T100HA module, since Tyree was an ASUS
T100HA and apart from the module being probably very much outdated we
don't have another such hardware in Vuizvui.
Signed-off-by: aszlig <aszlig@nix.build>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This reverts commit b9a7b3f0710ed828c9ec7562a6e641810fadd61b.
Our override is no longer needed and it actually doesn't build anymore
since even current stable Nix (2.17.0) is too new for nixos-option. This
is why in nixpkgs, there is a specific override for nixos-option which
uses Nix 2.15 instead.
For us this essentially means that we can drop the override, since
either someone fixes nixos-option to work with newer Nix versions or we
get a specific override that doesn't interfere with our (as in Vuizvui)
override of the "nix" attribute.
Signed-off-by: aszlig <aszlig@nix.build>
|
| |
|
| |
|