This fixes bug #61962.
* lisp/subr.el (safe-copy-tree): New function.
* lisp/emacs-lisp/bytecomp.el (byte-compile-initial-macro-environment): Amend
the entry for eval-and-compile to use safe-copy-tree and
byte-run-strip-symbol-positions for the eval part.
* doc/lispref/lists.texi (Building Lists): Document safe-copy-tree.
* etc/NEWS: Note the new function safe-copy-tree.
To understand the regression consider this recipe where the 'fo'
pattern is typed by the user in the last step.
emacs -Q
C-x b foo RET
C-x b afoo RET
C-x b *scratch* RET
M-x fido-mode RET
C-x b fo
This used to offer both 'foo' and 'afoo' as candidates but now only
offered 'foo'. This is because the pattern 'fo' matches 'foo', but
not 'afoo' with 'basic' completion style.
Fido mode, however, prioritizes 'flex' completion style, and that is
not happening here as it used to.
This was introduced in this commit
commit bf81df86e5
Author: João Távora <joaotavora@gmail.com>
Date: Wed Dec 7 10:43:59 2022 +0000
Don't override completion-category-defaults in fido-mode
I took away the nil setting of 'completion-category-defaults; in Fido
mode's minibuffer. It seemed generally the correct thing to do, and
was done mainly because Eglot added its style preferences to that
variable instead of completion-category-overrides directly, which is a
nono. So, to be able use the Fido UI with Eglot successfully,
'completion-category-defaults' should stay untouched. Or so I
thought.
However, I failed to notice that, for most categories, the default
value of 'completion-category-defaults' prioritizes the 'basic'
completion style.
For example, in the 'buffer' category, the default value has the
styles list '(basic substring)'. This means that if a pattern matches
accoring to the 'basic' style, 'substring' will not be tried. And
neither will 'completion-styles' which in Fido mode's case happens to
be 'flex'.
The solution in this commit is to craft a value for completion
category defaults that is just like the default one, but prioritizes
'flex' completion for every category.
* lisp/icomplete.el (icomplete--fido-ccd): New helper.
(icomplete--fido-mode-setup): Use it.
This reverts commit de4277af00.
It breaks WebP support at least on my system.
There's no reason to require libwebpdecoder library to be
installed, since we don't use the functions from it, at least
not in libwebp 1.2.1.
* lisp/emacs-lisp/cconv.el (cconv-convert): When form is an
interactive lambda form, don't destructively modify it, as it might be
a constant literal. Instead, create a new list with the relevant
place(s) changed.
* test/lisp/emacs-lisp/cconv-tests.el
(cconv-tests-interactive-form-modify-bug60974): New test.
Mentioned in bug#61893.
* lisp/progmodes/c-ts-mode.el (c-ts-mode--anchor-prev-sibling): Handle
"#elif" and "#else".
* test/lisp/progmodes/c-ts-mode-resources/indent-preproc.erts: Add an
"#elif" to the test.
Mentioned in bug#61893, although not the subject of it.
* lisp/progmodes/c-ts-mode.el:
(c-ts-mode--anchor-prev-sibling): Fix the child index for preproc_else
and preproc_elif case.
* test/lisp/progmodes/c-ts-mode-resources/indent-preproc.erts:
New test.
Not the subject of it, but mentioned in bug#61893.
* lisp/progmodes/c-ts-mode.el (c-ts-mode--anchor-prev-sibling): Skip
the sibling if it doesn't start on it's own line.
* test/lisp/progmodes/c-ts-mode-resources/indent.erts: New test.
836044f329 Fix c-ts-mode preproc directive indentation
64980a59b6 ; * lisp/files.el (hack-local-variables): Fix typo in the...
a7cd125d49 More robustly unspoof HOME in Eglot tests (bug#61637)
6c66dbd02c Turn on Eglot inlay hints by default
246f5b541c Update ts modes missed in 4c16fd3a51 to use column-0
0bfba49ca7 Robustify Eglot for "transient" projects
ea5fd375bb Fix documentation of 'normal-mode' in buffers that don't ...
4c16fd3a51 Change tree-sitter indent anchor 'point-min' to 'column-0'
f47b393015 Fix go-ts-mode multi-line string indentation (bug#61923)
e0bf2da3db ; More accurate doc strings for 'window-at' and 'window-a...
Mentioned in bug#61893, although not the subject of that report. This
change fixes indentation for nested directives. For example, when the
directive involves elif and the like, the elif is nested in the if
directive, so simply using grand-parent and great-grand-parent for
anchor is insufficient, because the nesting can grow arbitrarily.
The test added also covers the last preproc fix.
* lisp/progmodes/c-ts-mode.el:
(c-ts-mode--standalone-parent-skip-preproc): New function.
(c-ts-mode--indent-styles): New rules.
* test/lisp/progmodes/c-ts-mode-resources/indent-preproc.erts: New test.
A fair number of LSP servers allow user-local installations instead of
system wide installations. Emacs's technique of spoofing the HOME env
var to some non-existent or non-home directory during 'make check'
breaks these tests. That's because the executables are still found by
executable-find, but their invocation will rarely be successful as
HOME isn't what they expect it to be.
Eglot tests already had a technique for dealing with this,
"unspoffing" HOME just for the invocations of LSP server but it
stopped working a while back. So make it more robust.
Eventually, we'll want to decide wether these local servers should be
considered in 'make check' runs at all, or whether there is a way to
use them with a spoofed HOME.
* test/lisp/progmodes/eglot-tests.el (eglot--call-with-fixture):
More robustly unspoof HOME.
This is like any other server-provided feature, and may be turned off
client-side by setting eglot-ignored-server-capabilities like
(add-to-list 'eglot-ignored-server-capabilities :inlayHintProvider)
* lisp/progmodes/eglot.el (eglot--maybe-activate-editing-mode):
Activate eglot-inlay-hints-mode.
(eglot-inlay-hints-mode): Instead of warning about missing
:inlayHintProvider, turn off eglot-inlay-hints-mode.
When Eglot needs to synthesize a "transient" project for
default-directory sometimes the value of that variable is set to an
unexpanded value, sometimes not. This can cause simple invocations
like.
Emacs -Q ~/path/to-some-python-file.py -f eglot
to fail, because eglot--current-server will be looking for a project
in the registry called
(transient . "~/path")
where in reality it is stored there as
(transient . "/home/someuser/path")
The fix is to always expand default-directory in eglot--current-project.
* lisp/progmodes/eglot.el (eglot--current-project): Use
expand-file-name.
Originally discovered by Tyler Dodge in his article "Significant Garbage
Collection Improvement For Emacs".
While testing this change on Archlinux system with Intel i5-7200U CPU,
average time of garbage collection gets reduced by ≈25%. Other users
report improvements up to 50%. While monitoring PSS of emacs with and
without customizations loaded before and after the patch, no
statistically significant differences were discovered. So overall, this
change is a win.
* src/alloc.c (BLOCK_ALIGN): increase from 1024 to 32768.
* lisp/emacs-lisp/eldoc.el (eldoc-minibuffer-message): Redirect
eldoc messages to the mode-line when in edebug-mode (bug#56459).
(eldoc-minibuffer--cleanup): New function, used in above.
Adding eglot--maybe-activate-editing-mode to find-file-hook isn't
really necessary, since it is already in
'after-change-major-mode-hook', and that also runs every time we find
a file.
This reduces the number of project.el logic that runs every time the
user visits a file.
* lisp/progmodes/eglot.el (find-file-hook): No need to add
eglot--maybe-activate-editing-mode here.
When `total_length` is 0 there should be no subtree at all, but
`delete_interval` only deletes one interval, so make sure we don't end
up with some stale child of `i`.
The change
commit 3a651773d2
Author: João Távora <joaotavora@gmail.com>
Date: Fri Mar 3 13:13:35 2023 +0000
Eglot: pay better attention to hints' paddingLeft/Right (bug#61924)
Inadvertently added a line to this file, which shouldn't (yet) bet
added.
* lisp/icomplete.el (icomplete--vertical-minibuffer-setup): Remove
truncate-lines.
Reverts 59365f9285 and
56cd810b9d.
Reverting because the use of treesit-ready-p is very much intentional.
* lisp/progmodes/go-ts-mode.el (treesit-ready-p): Revert change.
* doc/misc/erc.texi: Update ERCVER to 5.5.
* lisp/erc/erc.el: Increment main version header to 5.5. Update
Compat version in package-requires header to 29.1.3.4. Update
`customize-package-emacs-version-alist' entry by mapping ERC 5.5 to
Emacs 29.1.
(erc-version): Update value to 5.5.
This resolves an issue with running test/lisp/server-tests.el when
'DISPLAY' is set: the tests would start the Emacs server in a batch
session, and then request a new GUI frame. However, that caused the
tests to terminate with SIGIO, since a batch Emacs doesn't set a SIGIO
handler.
Ref: <https://lists.gnu.org/archive/html/emacs-devel/2023-02/msg00879.html>
* lisp/frame.el (make-frame): Always create a terminal frame in batch
sessions.
* lisp/gnus/message.el (message-setup-1):
* lisp/progmodes/cc-engine.el (c-forward-single-comment):
* lisp/progmodes/ruby-mode.el (ruby-add-log-current-method):
Eliminate uses of condition-case without handlers. These seem to have
been there for quite a long time. This change does not affect the
behaviour of the code and makes some warnings go away.
