This reverts commit feac531415.
This patch has deactivated 253 test cases without a sufficient reasoning.
Instead it speaks about a shotgun in its commite message.
The patch is reverted because
- It hasn't been discussed on emacs-devel. It should have, because it is
a serious change in our infrastructure (new official tag :nobatch). Any
documentation of this change, for example in test/README, is missing.
- The proper way to deactivate such tests would have been
(skip-when noninteractive)
Even better to skip for the respective reasons.
- There is no fault report. There is no information about how these tests
have failed. Since it hasn't been a problem so far for us, nobody will
work on a fix forever.
There were a bunch of tests that were breaking make check and should
never be run in batch mode, because they do things like assuming there
is a controlling tty or assuming we can access network services when
we can't (e/g. in a CI/CD environment). I have shotgunned this
problem by tagging all the failing tests with :nobatch and then
changing the default and expensive selectors so make check won't barf
all over its shoes.
As many of these :nobatch should be individually removed as possible, after
upgrading the test harness to mock the environmental stuff they need.
Investigate these failures with "make check-nobatch".
* test/lisp/progmodes/eglot-tests.el (eglot--call-with-fixture):
Normalise 'temporary-file-directory' to stave off problems that
occur when it contains symlinks, which is common on MacOS.
Emacs 28 started fontifying several 'composite' (non-builtin) GNU M4
macros which, while listed under '(m4) Macro index', are not defined
by GNU M4, and are included in its manual for illustrative purposes.
These macro keywords range from clearly misleading (e.g., 'example',
which the GNU M4 manual explicitly describes as nonexistent),
to common in Autoconf's M4sugar layer (e.g., 'm4_quote'),
to variations thereof (e.g., 'foreachq').
It is arguably too late to revert all of these additions, so this
patch selects a handful to remove, and categorizes the rest; this
should make it easier to hide some categories behind a font lock
level or user option in the future, if desired.
This patch reverts the Emacs 28 addition of 'example', and removes
the older 'file', 'line', and 'gnu' macros. GNU M4 renamed 'gnu' to
'__gnu__' already in 1990; and I couldn't find references to the
other two which exist as '__file__' and '__line__', respectively.
The remaining macros are partitioned into three sets: built-in,
M4sugar-like, and other (bug#80412). For discussion, see also:
https://lists.gnu.org/r/emacs-devel/2025-10/msg00357.html
* lisp/progmodes/m4-mode.el (m4-program-options): Remove stale
commentary.
(m4--macro-list): Remove, replacing with...
(m4--builtin, m4--autoconf, m4--composite): ...these new rx
definitions. All uses changed. Remove 'example', 'file', 'gnu',
and 'line' as keywords.
(m4-font-lock-keywords): For consistency with real #-comments,
fontify dnl macro with font-lock-comment-delimiter-face, and only
the rest of its line with font-lock-comment-face. Fontify argument
references with font-lock-variable-use-face rather than
font-lock-variable-name-face. Use font lock faces rather than their
eponymous but obsolete variables. Remove redundant entries for $@
and $*. Prefer shy regexp groups where applicable.
(m4-mode-syntax-table): Quote syntactically special characters.
(m4-m4-buffer): DRY using m4-m4-region.
(m4-m4-region): Take region bounds as optional arguments, as
recommended under '(elisp) The Mark'. Quote shell command.
Support non-contiguous regions.
(m4-current-defun-name): Recognize m4 backtick in addition to
Autoconf bracket. Prefer shy regexp groups where applicable.
(m4-mode): Simplify font-lock-defaults.
* test/lisp/progmodes/m4-mode-resources/font-lock.m4:
* test/lisp/progmodes/m4-mode-tests.el: New test files.
The custom Pdb class enables native completion in pdb by
wrapping the pdb's native completer. It also makes necessary
function definitions like __PYTHON_EL_* available between pdb
frames, and enables non-native completion/ffap/eldoc
functionalities when debugging inside python modules.
* lisp/progmodes/python.el (python-shell-send-setup-code): Fix
the separator between python-shell-setup-codes.
(python-shell-completion-native-setup): Move common completion
setup code ...
(python-shell-completion-setup-code): ... here.
(python-shell-completion-at-point): Enable native completion for
pdb and respect the delimiter of pdb completer.
(python-shell-pdb-setup-code): New variable.
(python-shell-comint-watch-for-first-prompt-output-filter): Send
setup codes only once.
(python-ffap-module-path, python-eldoc--get-doc-at-point): Stop
sending setup code in every function call.
* test/lisp/progmodes/python-tests.el (python-tests--pdb-1)
(python-shell-pdb-1): New test.
* etc/NEWS: Mention the change. (bug#80182)
We add a symbol property 'elisp-scope-variable-spec', which
holds a specification for the values of a variable with that
property. For example, (put 'foo '(symbol . face)) says
that the value of variable 'foo' is a face name.
This allows elisp-scope.el to analyze forms such as
(setq foo 'bar) and (let ((foo 'bar)) ...) more accurately.
This is also used for analyzing macros that let-bind (one of)
their arguments to a specific special variable, such as
'with-connection-local-application-variables'.
We initially add this new property to some common variables.
* lisp/emacs-lisp/elisp-scope.el: Add
'elisp-scope-variable-spec' property to
'coding-system-for-read/write' and 'major-mode'.
(elisp-scope--variable-spec): New defsubst.
(elisp-scope--let-1, elisp-scope-let*, elisp-scope-setq):
Use it.
* test/lisp/progmodes/elisp-mode-resources/semantic-highlighting.el:
Add test form.
* test/lisp/progmodes/eglot-tests.el
(eglot-test-project-wide-diagnostics-rust-analyzer):
Bind project-vc-non-essential-cache-timeout to 0 so that the
result of project-current is refreshed (bug#80387).
* lisp/progmodes/json-ts-mode.el (json-ts--get-path-at-node)
(json-ts--path-to-jq, json-ts--path-to-python): New functions.
(json-ts-jq-path-at-point): New command for getting JSON path at point.
* test/lisp/progmodes/json-ts-mode-tests.el: New file.
Add tests for the utility command.
* etc/NEWS: Announce new command 'json-ts-jq-path-at-point' (bug#80190).
* lisp/emacs-lisp/cl-generic.el (cl--generic-make-function):
Preserve advertised-calling-convention info.
* test/lisp/emacs-lisp/pcase-tests.el (pcase-tests-quote-optimization):
Require `byte-opt` to fix the test when the compiler is not loaded yet.
* lisp/progmodes/elisp-mode.el: Fix some >80column problems.
(elisp--xref-format-extra)
(elisp--xref-format): Make them constant, now that we don't have
the purespace. Also, use %S since some of the elements
don't necessarily have names and even if they do, we'd want to
escape any funny characters in them to avoid ambiguities.
(elisp--xref-find-definitions): Fix uses of
`elisp--xref-format-extra` accordingly.
Improve heuristic to distinguish proper `cl-defgeneric` from
implicit ones.
(elisp-eldoc-docstring-length-limit)
(elisp-eldoc-funcall-with-docstring-length): Remove redundant `:group`.
* lisp/cedet/mode-local.el (xref-mode-local-overload): Pass the
override symbol rather than its name through `elisp--xref-format-extra`.
* test/lisp/progmodes/elisp-mode-tests.el (find-defs-constructor):
Adjust test to new text.
* lisp/progmodes/eglot.el (eglot--semtok-request)
(eglot--semtok-after-send-changes): Unbreak for 26.3.
* test/lisp/progmodes/eglot-tests.el (eglot--tests-connect): Use
split-string.
(eglot-test-rust-completion-exit-function): Use skip-unless.
* test/lisp/progmodes/eglot-tests.el
(eglot--wait-for): Fix thinkos and improve.
(eglot--tests-connect): Take TIMEOUT and SERVER kwargs.
(eglot-test-eclipse-connect)
(eglot-test-slow-sync-connection-wait)
(eglot-test-slow-sync-connection-intime): Update
eglot--tests-connect call.
Discussed in
https://lists.gnu.org/archive/html/emacs-devel/2025-12/msg00903.html.
* lisp/progmodes/project.el (project-files) <vc>: Look up the
value of 'project-vc-ignores' inside DIR rather than project root.
This allows this variable to be applied in external dirs as well.
(project-ignores) <vc>: Do likewise.
(project--value-in-dir): Ensure that DIR has a trailing slash.
The caller might pass the value without slash if produced
differently.
* test/lisp/progmodes/project-tests.el
(project-vc-ignores-in-external-directory): New test.
(project-vc-supports-files-in-subdirectory): Update the test,
considering the subtle change in behavior: the directory DIR
directly includes a .dir-locals.el which sets project-vc-ignores,
which wasn't previously applied because it wasn't set in the root.
Some of the fixes are to continue to use American
rather than British spelling.
* doc/misc/modus-themes.org (my-modus-themes-engraved-faces):
Fix misspelled ‘:foreground’s.
* etc/themes/modus-themes.el (modus-themes-faces):
Fix misspelled ‘modus-themes-bold’.
* lisp/emacs-lisp/rx.el (rx--normalize-char-pattern):
Rename from rx--normalise-char-pattern.
(rx--optimize-or-args): Rename from rx--optimise-or-args.
* lisp/frame.el (frame--special-parameters):
Fix misspelled "right-divider-width".
* lisp/net/tramp.el (tramp-fingerprint-prompt-regexp):
Use American spelling “centered”, to match current libfprintf.
* lisp/org/org-fold-core.el (org-fold-core--optimize-for-huge-buffers):
Rename from org-fold-core--optimise-for-huge-buffers.
(org-fold-core-update-optimization):
Rename from org-fold-core-update-optimisation,
leaving an alias behind.
(org-fold-core-remove-optimization):
Rename from org-fold-core-remove-optimisation,
leaving an alias behind.
* lisp/org/org.el (org-advertized-archive-subtree):
This alias is now obsolete.
* lisp/play/zone.el (zone-ignored-buffers):
Fix misspelling of ‘zone--buffer-encrypted-p’.
* lisp/progmodes/csharp-mode.el (csharp-ts-mode-faces):
Fix misspelling of ‘csharp’ group.
* lisp/vc/vc.el (vc-clonable-backends-custom-type):
Rename from vc-cloneable-backends-custom-type,
leaving an alias behind.
* test/lisp/emacs-lisp/bytecomp-tests.el:
(bytecomp-tests--warn-arity-non-compiled-callee):
Rename from bytecomp-tests--warn-arity-noncompiled-callee.
(bytecomp-test-defface-spec): Reword a deliberate misspelling
of “default” that is so common I don’t want it to pollute the
spelling dictionary.
* test/lisp/emacs-lisp/package-vc-tests.el:
(package-vc-tests-preserve-artifacts):
Rename from package-vc-tests-preserve-artifacts.
* test/lisp/eshell/em-prompt-tests.el:
(em-prompt-test/forward-backward-paragraph-1):
Reword a deliberate misspelling of “goodbye” that is so common I
don’t want it to pollute the spelling dictionary.
* test/lisp/progmodes/project-tests.el
(project-vc-supports-project-in-different-dir):
Account for difference in behavior when failing over to 'find'
(comments in bug#79809).
* lisp/progmodes/project.el (project-files):
Use 'file-in-directory-p' to dispatch to 'project--vc-list-files'
when listing subdirectory files too (bug#79809).
* lisp/progmodes/ruby-mode.el (ruby-smie-rules): Check for ':' and
'=>' as previous tokens, and handle symbols ending with ':' to
properly indent keyword argument arrays and hashes when
ruby-bracketed-args-indent is nil.
* lisp/progmodes/ruby-ts-mode.el (ruby-ts--parent-call-or-bol):
Handle arrays/hashes that are children of 'pair' nodes (keyword
arguments) to ensure consistent indentation.
* test/lisp/progmodes/ruby-mode-resources/ruby-bracketed-args-indent.rb:
Add test cases for keyword argument arrays and hashes with both
symbol-colon and hash-rocket syntax.
When ruby-bracketed-args-indent is nil, arrays and hashes used as
keyword argument values now indent by ruby-indent-level from the
line start, matching the documented behavior and fixing inconsistent
indentation (bug#74517).
(https://lists.gnu.org/archive/html/emacs-devel/2025-11/msg00939.html)
The tests are pretty basic, I don't expect them to catch
anything else than giant bugs.
* test/lisp/progmodes/eglot-tests.el (eglot-test-semtok-basic)
(eglot-test-semtok-refontify): New tests.
(eglot--semtok-faces, eglot--semtok-wait): New helpers.
Newer/newest clangd support more encodings and actually
negotiate with us, so test shouldn't assume UTF-16..
* test/lisp/progmodes/eglot-tests.el
(eglot-tests--lsp-abiding-column-1): Tweak.
Replace the use of 'elisp-scope-face(-1)' for analyzing face
specifications with 'elisp-scope-1' calls with an
appropriate OUTSPEC argument. This allows us to analyze
face specifications even when they are not passed directly
to relevant functions, but rather appear in a tail position
of a form that evaluates to a face specification.
* lisp/emacs-lisp/elisp-scope.el
(elisp-scope--match-spec-to-arg): Add new 'face' spec.
Use it instead of...
(elisp-scope-face, elisp-scope-face-1): ...these functions.
Remove them, no longer used.
* test/lisp/progmodes/elisp-mode-resources/semantic-highlighting.el:
Add test.
So far, 'scope-elisp-analyze-form' would disable
'read-symbol-shorthands' while reading a source form in
order to obtain the "original" length of symbols and pass
that length to the callback called for each analyzed symbol.
However, 'scope-elisp-analyze-form' could report an
incorrect length when a symbol was written with redundant
escaping (e.g. 'f\oo'). Moreover, disabling
'read-symbol-shorthands' breaks macro-expansion during
analysis, because macros may expect "expanded" symbols,
without shorthands.
In this commit we address these issues by leaving
'scope-elisp-analyze-form' enabled (so we get expended
symbols for macro-expansion) and recovering the original
length lazily in the callback, if needed, by going to the
beginning of the symbol and searching forward for its end.
* lisp/emacs-lisp/elisp-scope.el (elisp-scope--report):
Replace LEN argument with SYM, the analyzed symbol itself.
Adapt all callers.
(elisp-scope-analyze-form): Cease let-binding
'read-symbol-shorthands' to nil while reading. Wrap
analysis in 'save-excursion' when reading from current
buffer. Update docstring.
* lisp/progmodes/elisp-mode.el (elisp-local-references)
(elisp-fontify-symbol): Obtain symbol length from buffer.
* test/lisp/progmodes/elisp-mode-resources/semantic-highlighting.el:
Add test that incorporates 'read-symbol-shorthands'.
* test/lisp/progmodes/elisp-mode-tests.el (elisp-test-font-lock):
Set up 'read-symbol-shorthands' in test file.
Use argument specs to analyze complex widget types.
* lisp/emacs-lisp/elisp-scope.el (elisp-scope-widget-type)
(elisp-scope-widget-type-1)
(elisp-scope-widget-type-keyword-arguments)
(elisp-scope-widget-type-arguments)
(elisp-scope-widget-type-arguments-1): Delete, no longer used.
(custom-declare-variable, define-widget): Simplify analyzers.
(elisp-scope--match-spec-to-arg): Add new 'list', 'and', and
'plist-and-then' parametric specs, and add 'widget-type' as a
new recursive spec.
* test/lisp/progmodes/elisp-mode-resources/semantic-highlighting.el
Add test.
Rename a couple of faces to solidify the convention that the
face name 'elisp-foo' implies "references to foo", not "foo
definitions". For definitions we use 'elisp-deffoo' if foo is
only one word, or 'elisp-bar-baz-definition' otherwise.
* lisp/progmodes/elisp-mode.el (elisp-function-reference):
Rename to 'elisp-function'.
(elisp-macro-call): Rename to 'elisp-macro'.
(elisp-non-local-exit):
(elisp-unknown-call):
(elisp-special-form):
* lisp/emacs-lisp/elisp-scope.el:
* test/lisp/progmodes/elisp-mode-resources/semantic-highlighting.el:
Update references to renamed faces.
* lisp/emacs-lisp/elisp-scope.el (elisp-scope-if-let): Fix
handling of a plain symbol as one the bindings in an
'if-let*' form, as in (if-let* (foo) 'bar).
* test/lisp/progmodes/elisp-mode-resources/semantic-highlighting.el:
Test it.
* test/lisp/progmodes/elisp-mode-resources/semantic-highlighting.el:
Add test form with 'when-let*'.
* test/lisp/progmodes/elisp-mode-tests.el (elisp-test-font-lock):
Trust temporary test buffer.
* test/lisp/progmodes/elisp-mode-tests.el
(elisp-test-font-lock): New test.
* test/lisp/progmodes/elisp-mode-resources/semantic-highlighting.el:
New resource file.
* lisp/progmodes/ruby-mode.el (ruby-syntax-propertize):
Create an exception for assign-division operator not to be
recognized as regexp start (bug#79454).
* test/lisp/progmodes/ruby-mode-resources/ruby.rb: Add example.
Escape literal % characters in Eglot URIs
Otherwise, a literal % in a file-name will be interpreted (by the
language server) as if it were a part of a percent-encoded sequence.
See Bug#78984 for context on why `url-path-allowed-chars' cannot be
changed to escape literal % characters.
* lisp/progmodes/eglot.el (eglot--uri-path-allowed-chars): Escape %,
remove the redundant variable definition.
* test/lisp/progmodes/eglot-tests.el (eglot-test-path-to-uri-escape):
test it.
* lisp/progmodes/lua-ts-mode.el (lua-ts--comment-font-lock):
Apply 'font-lock-comment-delimiter-face' to the entire span of
initial dashes. In particular, this improves the appearance of
LuaCATS and EmmyLua style annotations which use "---".
* test/lisp/progmodes/lua-ts-mode-resources/font-lock.lua:
Add tests. (Bug#79258)
* lisp/progmodes/cperl-mode.el (cperl-indent-line): Make sure that
labels are indented in code only
* test/lisp/progmodes/cperl-mode-resources/cperl-indents.erts: Two
new testcases for non-indentable "labels" in a regex pattern and a
qw list
This fixes the second issue in Bug#79269.
* lisp/progmodes/cperl-mode.el (cperl-init-faces): Move handling
of signatures with initializers from the "anchor" to the
"anchored" matcher
* test/lisp/progmodes/cperl-mode-resources/proto-and-attrs.pl (sub_7):
Test case for Bug#79269, wrong face report
* test/lisp/progmodes/cperl-mode-tests.el
(cperl-test-fontify-attrs-and-signatures): Make sure that the test
catches sub_7 for Bug#79269
Don't signal an error when encountering invalid rust syntax. Without
this patch, invalid rust code would prevent a chunk of the buffer from
being highlighted (bug#79272).
* lisp/progmodes/rust-ts-mode.el (rust-ts-mode--fontify-scope):
(rust-ts-mode--fontify-pattern): Avoid calling `string-match-p' on nil
when a node is missing a parent.
* test/lisp/progmodes/rust-ts-mode-resources/font-lock-no-parent.rs:
Rust file that reproduces the issue.
* test/lisp/progmodes/rust-ts-mode-tests.el: Test case to reproduce the
issue.
