| Commit message (Collapse) | Author | Age |
|---|
| |
|
|
| |
See https://github.com/clojure-emacs/clj-refactor.el/pull/392#issuecomment-352399191
|
| | |
|
| | |
|
| |
|
|
|
|
| |
Previously in such cases we'd get a `do` expression without a body, which isn't
be nice. Now we generate the config expression only after checking it needs to
generated to begin with.
|
| |
|
|
|
|
|
| |
It evaluates the current top-level form up to the point. You can think of this
as a poor man's contextual evaluation.
The command is bound by default to `C-c C-v (C-)z` in `cider-mode`.
|
| |
|
|
|
|
|
|
|
|
|
|
| |
(#2156)
The `cider-tramp-prefix` does not return only tramp prefix.
It used the `tramp-make-tramp-file` in the `with-parsed-tramp-file-name` which force it to make a full tramp file path, not only prefix. I placed the `cider-make-tramp-prefix` for making prefix only. It is modified a little bit from `tramp-make-tramp-file`.
The `cider--client-tramp-filename` does not remove `file: scheme` for tramp filename.
It should return the string with tramp prefix and remote source path. But previous version leaves `file: scheme` in a remote source path. I just added a line to remove this scheme in this function.
The `cider-find-file` leaves a useless list buffer per opening remote archived source like zip, jar.
When continuously tracking, it'll will generate a lot of trash and make us tired to remove it. Now it'll be removed automatically.
|
| |
|
|
| |
See https://blog.readthedocs.com/securing-subdomains/ for details.
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
Cljc buffers send their evals to both clj and cljs repls if available due to
`cider-map-connections`. Toggling a current buffer's connection involves hiding
the other connection.
Previously, when toggling _again_, the original list was not consulted and only
the truncated list, preventing the other connection from being found. This
allows for the full list to be searched for the other buffer.
In addition, a prefix dictates that the local connection list is discarded in
favor of the full list, restoring the evaluation in both clj and cljs
buffers (if both are present).
|
| | |
|
| | |
|
| |
|
|
|
|
|
|
| |
Fixes #2133
In Emacs 26, the print representations of structs changed from `[cl-struct ...]`
to `#s(...)`. This broke tests that compare structs against print
representations. Building structs dynamically resolves this.
|
| |
|
|
| |
See: https://emacs.stackexchange.com/a/10058/10269
|
| |
|
|
|
|
|
|
|
|
| |
Fix #2130.
Emacs 26 obsoletes `if-let` and `when-let`, replacing them with `if-let*` and
`when-let*`. This raises byte-compilation warnings (treated as errors) when
testing against Emacs 26.
See: http://git.savannah.gnu.org/cgit/emacs.git/tree/etc/NEWS?h=emacs-26#n1278
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| |
|
|
|
| |
- new function `cider-project-connections-types`
- `cider-connection-type-for-buffer` returns "multi" in cljc an cljx
|
| |
|
| |
Basically we simply apply the ANSI color and discard it, so it would interfere with the `clojure-mode` font-locking we're doing.
|
| |
|
|
|
|
|
| |
when filtering stackframes, there was only a "remove" type of filter. If we wish
to filter to just our project, we need a way to ensure that project frames
persist even if another tag says to remove it. This adds the "positive" filter
mechanics.
|
| |
|
|
|
|
|
|
|
|
| |
In order to add "positive" filters, we must only show or hide when we're
actually on a stackframe line. This method takes a very naive way of starting at
the top of the buffer and considering whether to hide all lines. The other lines
just don't have any 'flags properties so they never come up as needing to be
hidden in the `(seq-intersection filters flags)` part, so hide is always false
on them. But if we want to show _only_ those lines with a particular flag this
hides the cause line and any other lines that don't have flags at all.
|
| | |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
These tests pass in their current form, but they do not pass on a build of emacs
from source tracking master. The tests use cl-every and cl-some and the tests
assert that the results are either `:equal-to t` or `:not :equal-to t`. Right
now, cl-some returns t but on my build of emacs from master it returns the truth
value that it finds, causing buttercup to fail in its comparison with:
FAILED: Expected cider-reader-conditional-face to be `eq' to t
So rather than checking against some value for truth, we use the buttercup
assertion :to-be-truthy.
Further, we remove the reliance on the cl library and switch over to seq.el.
Also some line length formatting and rename test predicate helpers.
|
| | |
|
| |
|
| |
This also fixes some issues checkdoc was currently reporting.
|
| | |
|
| | |
|
| |
|
|
|
| |
This will allow users to customize options that are not related to tasks and passed directly to
their build tool of choice. For instance: `-o` in `lein -o ....` for offline mode.
|
| | |
|
| | |
|
| |
|
|
|
|
| |
Use 'cider-repl-pretty-print-width' to control the width that the pretty
printer uses on the repl. If nil, default to fill-column, and finally to
80 if fill-column can be not set.
|
| |
|
|
|
| |
The functions in `cider-compat.el` were actually added in Emacs 25.1,
not Emacs 24.4.
|
| | |
|
| |
|
|
|
|
|
|
|
|
| |
Add an option to define exclusions for injected dependecies. Add
`org.cloure/clojure` exclusion to `org.clojure/tools.nrepl` to mitigate
problem with tools.nrepl's clojure dependency.
Note about boot: the latest stable release does not support defining
dependency exclusions on the command line. However, this feature is
available in the upcoming 2.7.x release. This feature will be added
for boot when 2.7.x is released.
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
