The dotfile says that setting the major-mode leaders to nil will disable
the functionality, but there were a couple of places where that option
was not being respected.
post value for package :step property is both useless and considered
harmful. It would hijack the determinism of the loading order.
'pre value is used very very rarely and I even think it is not even
used since most of the time a package that must be loaded in pre step
is a bootstrap package. But we keep 'pre since it can be useful to
have it for _exceptional_ cases.
While doing the tests I encountered an edge case where a package owned
by the dotfile could be passed either a 'private or 'local location
which is ambiguous (moreover where do we put packages with 'local
location ?).
We remove the ambiguity by removing the 'private value and push the
path ~/.emacs.d/private/local/pkg/ when the owner of a local
package PKG is the dotfile.
When the owner of a local package is a layer then the load path is
in the "local" subdirectory of the layer directory. It adds no value
to use the old 'private location in this case.
This is how evil-leader gets `SPC m` to work for mode specific bindings,
and we need to mirror this step for `M-m`. This is big change, but in
order for `M-m` to work properly we need to remap it to the mode-map
whenever possible.
When the package already exists, use configuration-layer/make-package
to copy information into the existing package object.
This allows overwriting location, step and excluded, but NOT owner,
pre-layers or post-layers.
Should fix#2849. The new package-delete tries to save changes to the
selected package variable during the update process and this can cause
conflicts with how spacemacs wants to update packages (and even lead to
the .spacemacs file being deleted. This is also unnecessary, because we
are only deleting the packages temporarily to be installed at the next
boot, so there should be no need to update this variable.
This change only affects Emacs 25 users.
Instead of error, show a warning that the unavailable package(s) were
skipped. This should handle situations were a package is temporarily
unavailable from MELPA (or other sources).
If spacemacs encounters errors on the initial load of the user
.spacemacs file, now defaults to loading the default template.
Additionally, temporarily suspends pruning of orphan packages to make it
easier for users to recover after they fix their dotfile.
Previously, errors would cause the spacemacs boot process to halt,
leaving users with an almost stock emacs configuration. Now, users will
have a default spacemacs available to them to fix the error.
adds layer preservation.
add prompt for preferred editing style on recovery
Remove advice after first run.
add dotspacemacs/safe-load as function.
moves style prompt into dotspacemacs/init advice
add hybrid as valid editing style.
move conditional into dotspacemacs/load-file.
This sets the default distribution back to spacemacs. This is for people
who for whatever reason don't set `dotspacemacs-distribution` in their
dotfile.
Helm seems to treat "!" specially in pattern matching, so having a ! in
the pattern string when traversing directories is problematic. This
change fixes#2737, because as far as I can tell "+" has no special
meaning in a helm pattern.
Of course, we can choose a different character, but I'm fond of "+" as
representing "more layers here".
Now the user should use `SPC f e h` helm-spacemacs.
This is due to the fact that spacemacs layer is now a regular layer
stored in the "layers" directory.
- Rename contrib directory to layers
- Add new variable dotspacemacs-distribution
- Move spacemacs layer to layers/!distribution
- New layer spacemacs-core in layers/!distribution
- User can now set dotspacemacs-distribution to spacemacs or
spacemacs-core (default spacemacs)
spacemacs-core is very lightweight layer sufficient to build upon
spacemacs.
instead of the downloaded package archive. It fixes a case where
a package maintainer removes a dependency D of a package P and spacemacs
detects it which has the effect of deleting the obsolete package D, but
package-alist is not yet updated, because package-alist is updated only
when P is reinstalled (or upgraded).
Also it is not need to pass a list, example:
```
(setq dotspacemacs-configuration-layers
'((auto-completion :disabled-for org git)))
```
Will disable auto-completion for org and git layers.
Example:
(setq dotspacemacs-configuration-layers
'((auto-completion :disabled-for (org))))
Will not run any org/pre-init-xxx or org/post-init-xxx functions where
xxx is a package owned by the auto-completion layer.
Limitation:
If ownership of an auto-completion package P is stolen by another layer L
then :disabled-for must also be declared with L, otherwise the P will
be configured for org.
(setq dotspacemacs-configuration-layers
'((auto-completion :disabled-for (org))
(L :disabled-for (org)))
The user should not change the default owner of a layer unless really
required (for instance to use his own fork).
The test results buffer is displayed only if there are errors.
Make dotspacemacs/test-dotfile return non-nil when all the tests passed.
dotspacemacs/test-dotfile takes an optional parameter to hide the
result buffer.
Set as default unless user has dotspacemacs-use-ido t.
Adds comments to .spacemacs.template informing users of new behavior
tied to dotspacemacs-use-ido.
- configuration-layer/declare-layers
- configuration-layer/declare-layer
These functions can be used in the config.el file of a layer to add
required layers.
For instance the react layer needs both html and javascript layers.
salt layer needs the yaml layer (see the following commits)
It is now possible to override the recipe in a custom layer.
For instance to override the recipe R for package P in layer L
In a private layer L' put at the end of dotspacemacs-configuration-layers
and listing P with an init function for P, L' becomes the ower of
P and the recipe R' defined in L' overrides the recipe R.
The `package-delete` function refuses to delete a package that is a
depency by default, which prevents the spacemacs update function from
working most of the time.
This commit sets the flag to force deletion for Emacs 25.
Now it is possible to have unicode symbols in GUI and not in
terminal using emacsclient
Do to so, dotspacemacs-mode-line-unicode-symbols must be set to
the symbol 'display-graphic-p which will be lazily evaluated
via the macro dotspacemacs|symbol-value
This is especially useful for users having troubles with unicode
characters in their terminal but wants to use them in GUI client.
Allowed properties to be overridden are: location, step and excluded
This mechanism allow the user to alter the location of a given
package, for instance if she wants to use her own fork instead of the
version in ELPA.
Note that for now there is no easy way to override the package recipes.
Once all extensions were moved from "extensions" to "local",
spacemacs was no longer able to find the solarized themes.
Thanks to @TheBB for helping narrow this down!
Closes#2628
Unfortunately spacemacs is not designed for installation
at use-package call time.
For use-package to be called a package must already be installed,
so I will continue on the initial plan, that is providing the
quelpa recipe at the <layer>-packages list level.
This is no bigdeal, this is basically moving :quelpa from
use-package to <layer>-packages variables.
Moreover it makes more sense to define package property at
declaration time instead of initialization (at least in spacemacs
world where installation is decoupled from configuration).
This has a benefit of not assuming that the user .emacs.d/ is in the
user home directory. Should continue to work as expected when this is
the case, but you could also start a fresh Emacs session like
so (assumes OSX):
open -a Emacs.app -n --args -q -l /path/to/emacs.d/init.el
So you don't have to muddle with symlinking your ~/.emacs.d or replacing
it with another just to try Spacemacs (or any other config). Note, that
this won't work with `after-init-hook` which doesn't appear to run when
Emacs is run with -q flag. As a result the `dotspacemacs/config` in your
.spacemacs won't run.
`tooltip-use-echo-area' is obsolete since 24.1; disabling `tooltip-mode'
achieves similar effect, and Tooltip mode has already been disabled in the
current code. Since Spacemacs supports Emacs 24.3 and 24.4,
`tooltip-use-echo-area' usage can be removed safely.
I have seen many "I have a problem" discussions in the Gitter chat which
starts with a barrage of questions "Which OS? Which Emacs version?",
etc., so I thought it may be useful to have one function that will
generate the info to be copy-pasted into the Gitter chat and hence both
the user and others helping in the Gitter chat can jump directly to
solving the problem instead of the support volley to figure out the
setup.
Example output:
ELISP> (spacemacs/system-info)
"OS: darwin Emacs: 24.5.1 Spacemacs: 0.103.0 Spacemacs branch: develop
Layers: ((auto-completion :variables auto-completion-enable-help-tooltip
t) better-defaults emacs-lisp git markdown org (shell :variables
shell-default-height 30 shell-default-position (quote bottom))
syntax-checking version-control c-c++ clojure dash github html osx
python semantic sql)"
References:
From
https://github.com/syl20bnr/spacemacs/issues/2033#issuecomment-113861451 :
> Also what is your emacs version and OS ?
From
https://github.com/syl20bnr/spacemacs/issues/2042#issuecomment-113861501 :
> What's your Emacs version ? I presume it comes from the semantic
layer, can you test without the semantic layer ?
Instead of opening the file for the bookmark, use the bookmark-jump
function to properly jump to the file and location in the file. Also
show the bookmark name and the filename in the list.
Fixessyl20bnr/spacemacs#2431
So if called using the universal argument or with a parameter yes is
assumed to the question to update packages. Useful for calling Emacs
in batch mode to update Spacemacs.
Works with magit-next for now.
Tests to update and evilify functions robustness need to be improved.
Does not work 100% with magit-mode-map because it is created with
`make-keymap` and not `make-sparse-keymap` and `map-keymap` does not
seem to work with `make-keymap`.
Currently, pressing TAB does not move to next widget. Similarly,
pressing RET does not run the button at point. Use `kbd` function to
properly convert to internal key representation in Emacs that is usable
in both GUI and terminal.
Also move the key bindings to spacemacs/keybindings.el since it's a more
suitable place.
- Re-enable
- Move all autosave files to spacemacs-cache-directory
- Allow Auto-saving directly on current buffer.
- Add option to customize.
- Update .spacemacs template.
bookmark-all-names returns a list of names, but
spacemacs-buffer//insert-file-list expect a list of filenames.
Shortcut definitions seems to be a mix of two approaches/commits.
Removed the unused bits.
Refactor spacemacs/open-file to spacemacs/view-org-file to correctly
reflect its functionality. In Emacs, view means read-only. Make the
anchor-text optional so we can omit when not needed. Also add
expand-scope that can be either 'subtree or 'all to open a tree at point
or expand everything. 'subtree is used for the Change Log while 'all is
used for all Spacemacs documents.
Currently helm-spacemacs//documentation-candidates grabs everything it
can in the doc directory. In the future we might put some images there
so better just choose the correct file type to include.
- Use `~` marker to represent key bindings. The `~` and `=` markers are
equivalent in Emacs: `=` uses `org-verbatim verbatim` faces, while `~`
uses `org-code verbatim)`, but both `org-verbatim` and `org-code` is
inherited from `shadow` face. So we can use `~` marker for this purpose.
`~` marker is also displayed properly on Github, since it's part of
standard Emacs, so no web functionality is broken.
- Add org-kbd face for displaying key bindings.
- Update CHANGELOG.org: use the proper `~` marker for key bindings.
- Update VIMUSERS.org: use the proper `~` marker for key bindings.
It's important to make templating variables something unique-looking so
that you don't accidentally write template variables where you don't
mean to. For example, in extensions.template, "package names go here"
was being replaced with "package <name of package>s go here", etc.
Also Search case-insensitively for good measure
Otherwise when the widget list is deleted to insert new widgets, the
separator "widget" is essentially text with properties and can cause
error when trying to delete, which breaks quick help and release note
buttons, making the buttons not function anymore.
Adding packages to this list will install them without
needing them to be wrapped in a layer.
Ideal for niche languages of any package that don't need
any configuration.
- Modularize and add quickhelp button
Split the current hardcoded release note display functions into smaller
reusable functions. Then reuse it for creating quickhelp button.
- fix: bind widget-button-click to left mouse click
widget-button-press is for keyboard and binding it to a left mouse
button is not suitable. The clicks on buttons are often ignored. Use the
proper widget-button-click.
- Since this is develop, point it to 0.102.x
- Beautify org CHANGELOG
Setting org-indent-mode hides the leading stars of all Org headers except
the last one just enough to indicate indentation, which make it easier
to read. Also, put it in read-only-mode so user don't accidentially mess
up his reading material.
Also put the CHANGELOG in view-mode for reading and navigating read-only
content.
- Add Evil and Emacs tutorial buttons to quickhelp
So it is even easier for first time users.
Also refactor spacemacs-buffer//insert-release-note:
- Change it to spacemacs-buffer//insert-note
- Just insert content. Widgets are optional
- Insert different widget from the two functions
spacemacs-buffer//insert-quickhelp-widget and spacemacs-buffer//insert-release-note-widget
- Put the cursor on the quickhelp button
So that users can start using it immediately. New users only knows RET
to press a button and learn the rest there.
- Add m to jump back to top menu
Also update the quickhelp content.
highlight-parentheses highlights the sexp scope the cursor is in and
3 upper sexp levels. We don't have to actually place the cursor on a
parenthesis just to see where it ends or begins. This package takes of this
issue regardless where we are in the sexp.
Currently, `r` and `p` moves to different locations when the change log
text is folded/unfolded. Make it always jumps to the correct location
regardless of content of home buffer by actually search for it.
Also set keys to make home buffer usable with holy-mode.
Conflicts:
core/core-spacemacs-buffer.el
Now in addition to the <layer>/init-<pkg> function there are
- <layer>/pre-init-<pkg> (executed before <layer>/init-<pkg>)
- <layer>/post-init-<pkg> (executed after <layer>/init-<pkg>)
The init function is mandatory, if it is not present then the
package is ignored and not installed.
This mechanism allows soft (implicit) cross layers dependencies
between packages (see company for more info).
It is now possible to remove flycheck from spacemacs layer and
move it to its own syntax-checking layer.
- backup the packages to be updated
- then delete them
- the user restart emacs and spacemacs will install the last version
Fixes some update errors related to byte-compilation like the one
which affected the powerline (void variable left)
This commit also adds some page break for clarity
Standard Doge 1) uses two-word phrases and 2) violates the selectional
restrictions of the modifiers 'such', 'so', 'very', 'many', and 'much'.
(see
http://the-toast.net/2014/02/06/linguist-explains-grammar-doge-wow/#BjOVLZxGdAHsebDq.99)
These are my suggestions for banner phrases that more accurately match the usual
doge grammar. Oh god what am i doing with my life
