Fix symbol list matching regexp performance
Allow empty face lists, improve the face list matching regexp (see
discussion in Bug#69714) based on relint's comments, add tests:
* test/lisp/emacs-lisp/ert-font-lock-tests.el: Add tests.
* lisp/emacs-lisp/ert-font-lock.el: Fix regexps.
Fail on files with no assertions, parser now accepts multiple
carets per line and face lists:
* lisp/emacs-lisp/ert-font-lock.el: Assertion parser fix.
* test/lisp/emacs-lisp/ert-font-lock-resources/no-asserts.js:
* test/lisp/emacs-lisp/ert-font-lock-tests.el
(test-parse-comments--no-assertion-error)
(test-syntax-highlight-inline--caret-negated-wrong-face)
(test-macro-test--file-no-asserts): New test cases.
* doc/misc/ert.texi (Syntax Highlighting Tests): More syntax examples.
The `pcase-mutually-exclusive-predicates` table was not very
efficient since it grew like O(N²) with the number of
predicates. Replace it with an O(N) table that's auto-generated
from the `built-in-class` objects.
* lisp/emacs-lisp/pcase.el (pcase-mutually-exclusive-predicates):
Delete variable.
(pcase--subtype-bitsets): New function and constant.
(pcase--mutually-exclusive-p): Use them.
* lisp/emacs-lisp/cl-preloaded.el (built-in-class): Don't inline.
The `pp-fill` code sometimes end up generating things like:
(foo .
bar)
instead of
(foo
. bar)
so make sure we cut before rather than after the dot (and open
parens while we're at it).
* lisp/emacs-lisp/pp.el (pp-fill): Cut before parens and dots.
* test/lisp/emacs-lisp/pp-tests.el (pp-tests--dimensions): New function.
(pp-tests--cut-before): New test.
Not sure why earlier tests did not catch it, but there are more
places where we bump into problems because `eieio--class-precedence-list`
now returns also non-EIEIO classes.
* lisp/obsolete/eieio-compat.el
(eieio--generic-static-object-generalizer):
* lisp/emacs-lisp/eieio-core.el (eieio--generic-generalizer)
(eieio--generic-subclass-specializers): Handle non-EIEIO parents.
* test/lisp/emacs-lisp/eieio-tests/eieio-test-methodinvoke.el
(eieio-test-method-order-list-7): Adjust test.
A function or macro body consisting of a single string literal now only
uses it as a return value. Previously, it had the dual uses as return
value and doc string, which was never what the programmer wanted and
had some inconvenient consequences (bug#69387).
This change applies to `lambda`, `defun`, `defsubst` and `defmacro`
forms; most other defining forms already worked in the sensible way.
* lisp/emacs-lisp/bytecomp.el (byte-compile-lambda):
Don't use a lone string literal as doc string.
* test/lisp/emacs-lisp/bytecomp-resources/warn-wide-docstring-defun.el
(foo): Update docstring warning test.
* doc/lispref/functions.texi (Function Documentation): Update.
* etc/NEWS: Announce.
This reverts commit 4a0d430bdc.
AFAICT that commit was made to accomodate regressions introduced
in the new `cl-preloaded.el` code and these have been fixed.
* doc/lispref/modes.texi (Tabulated List Mode):
Add defvar tabulated-list-groups.
* lisp/buff-menu.el (Buffer-menu-group-by): New defcustom.
(Buffer-menu-unmark-all-buffers): Use tabulated-list-get-entry
to check whether the current line contains an entry.
(list-buffers-noselect): Enable outline-minor-mode
for tabulated-list-groups.
(list-buffers--refresh): When Buffer-menu-group-by is non-nil,
set tabulated-list-groups.
(Buffer-menu-group-by-mode, Buffer-menu-group-by-root):
New functions.
* lisp/emacs-lisp/tabulated-list.el (tabulated-list-groups):
New buffer-local variable.
(tabulated-list-print-fake-header): Add distinct overlay
property 'fake-header'.
(tabulated-list-header-overlay-p): Filter out overlays that
don't have the property 'fake-header'.
(tabulated-list-print): Use the variable 'tabulated-list-groups'
to sort entries in groups separately.
(tabulated-list-print-entries): New function factored out from
'tabulated-list-print'.
* test/lisp/emacs-lisp/tabulated-list-tests.el (tabulated-list-groups):
New test.
* test/lisp/emacs-lisp/edebug-tests.el (edebug-tests-trivial-backquote):
Don't use obsolete `edebug-eval-defun`.
(edebug-tests-trivial-comma): Use `inhibit-read-only`; don't use
obsolete `edebug-eval-defun`; and fix bug#69406 by binding
`eval-expression-debug-on-error`.
Since that fix, we made other changes (put arg names in allcaps)
which also happen to fix bug#54628, so we can remove the original fix
which was suboptimal when the type includes quotes.
* lisp/emacs-lisp/cl-generic.el (cl--generic-describe):
Don't rebind `print-quoted` to nil.
* test/lisp/emacs-lisp/cl-generic-tests.el
(cl-generic-tests--print-quoted): New test.
* lisp/emacs-lisp/cconv.el (cconv-convert):
Warn about (F ...) where F is a non-symbol function value (bytecode
object etc), but let it pass for compatibility's sake (bug#68931).
* test/lisp/emacs-lisp/bytecomp-tests.el (bytecomp--fun-value-as-head):
New test.
(comp-cstr-test-ts): Move out of `cl-eval-when`.
(comp-cstr-typespec-test): Delete.
(comp-cstr-synthesize-tests): Make it take the tests as an argument.
(comp-cstr-typespec-tests-alist): Delete var, pass its value to the
macro instead.
* test/lisp/emacs-lisp/bytecomp-tests.el
(bytecomp-tests--error-frame, bytecomp--byte-op-error-backtrace):
Make test pass again and simplify, using handler-bind instead
of the previous debugger hack.
Now that `ert.el` uses `handler-bind` instead of `debugger`, some
details of the behavior have changed. More specifically,
three tests are now broken, but these basically tested the failure
of ERT's machinery to record errors when ERT was run within
a `condition-case`.
AFAICT, these tests do not check for a behavior that we want,
so rather than "fix" them, I deleted them (bug#67862).
* test/lisp/emacs-lisp/ert-tests.el (ert-test-error-debug)
(ert-test-fail-debug-with-condition-case): Delete.
(ert-test-should-failure-debugging): Don't use `ert-debug-on-error`.
(ert-test-with-demoted-errors): It now passes. Bug#11218 is fixed!
* lisp/emacs-lisp/checkdoc.el (checkdoc-this-string-valid-engine):
Avoid false positive when a variable contains a keybinding (for
example, "C-g"). (Bug#68002)
* test/lisp/emacs-lisp/checkdoc-tests.el
(checkdoc-docstring-avoid-false-positive-ok): New test.
* lisp/emacs-lisp/package.el (package--prepare-dependencies): Move
from here...
* lisp/emacs-lisp/lisp-mnt.el (lm--prepare-package-dependencies):
...to here.
(lm-package-requires): New function.
(package-buffer-info): Use above new function.
* test/lisp/emacs-lisp/lisp-mnt-tests.el
(lm--tests-lm-package-requires): New test.
Add ert-font-lock as well as unit tests and testing resources.
* lisp/emacs-lisp/ert-font-lock.el: New library.
* test/lisp/emacs-lisp/ert-font-lock-resources/broken.js:
* test/lisp/emacs-lisp/ert-font-lock-resources/correct.js:
* test/lisp/emacs-lisp/ert-font-lock-tests.el: Unit tests.
(Bug#67460)
Clones of instances of subclasses of 'eieio-instance-inheritor' didn't
delegate to their ':parent-instance' field when reading object fields
using ':accessor'.
* lisp/emacs-lisp/eieio.el (defclass): Remove 'slot-boundp' check for
:accessor's getter
* test/lisp/emacs-lisp/eieio-tests/eieio-tests.el
(eieio-test-use-accessor-function-with-cloned-object): New test.
Copyright-paperwork-exempt: yes
The compiler didn't cancel aliasing if the aliased variable was
modified in a variable binding in the same `let` that created
the alias. For example,
(let ((x A))
(let ((y x)
(z (setq x B)))
y))
would incorrectly substitute y->x in the body form despite x being
already modified at that point, which normally should have cancelled
the aliasing.
Bug reported by Alan Mackenzie.
* lisp/emacs-lisp/byte-opt.el (byte-optimize--aliased-vars):
Now an alist that also contains the aliases; update the doc string.
* lisp/emacs-lisp/byte-opt.el (byte-optimize-form-code-walker):
* lisp/emacs-lisp/byte-opt.el (byte-optimize-let-form):
Detect aliasing early for `let`-bound variables as well.
* test/lisp/emacs-lisp/bytecomp-tests.el (bytecomp-tests--test-cases):
Add test cases.
Adjust the type operations to account for the fact that types can have
several parents.
* lisp/emacs-lisp/comp-cstr.el (comp--cl-class-hierarchy):
Use `cl--class-allparents`. Add FIXME.
(comp--direct-supertype): Declare obsolete.
(comp--direct-supertypes): New function.
(comp--normalize-typeset0): Rewrite to use `comp--direct-supertypes`;
adjust to account for the DAG structure; use `cl-set-difference`.
(comp--direct-subtypes): Rewrite.
(comp--intersection): New function.
(comp-supertypes): Rewrite and change return type.
(comp-subtype-p): Simplify.
(comp-union-typesets): Use `comp-supertypes` instead of iterating over
`comp-cstr-ctxt-typeof-types`.
* lisp/emacs-lisp/comp.el (comp--native-compile): Don't catch
errors if we're debugging.
* test/lisp/emacs-lisp/comp-cstr-tests.el: Adjust tests.
* lisp/emacs-lisp/cl-macs.el (cl-defstruct): Fix mishap when we
evaluate (cl-defstruct cl-structure-object ..) during the compilation
of `cl-preloaded.el`.
* lisp/emacs-lisp/cl-preloaded.el: Add corresponding assertion.
This warning is much more appropriate for the compiler, since lexical
binding affects what it can reason and warn about, than for checkdoc
as the warning has no bearing to documentation at all.
The move also improves the reach of the warning.
* etc/NEWS: Update.
* lisp/emacs-lisp/checkdoc.el (checkdoc-lexical-binding-flag)
(checkdoc-file-comments-engine): Move warning from here....
* lisp/emacs-lisp/bytecomp.el (byte-compile-file): ...to here.
* test/lisp/emacs-lisp/bytecomp-resources/no-byte-compile.el:
* test/lisp/emacs-lisp/bytecomp-tests.el
(bytecomp-tests--unescaped-char-literals)
(bytecomp-tests-function-put, bytecomp-tests--not-writable-directory)
(bytecomp-tests--target-file-no-directory):
Update tests.
(bytecomp-tests--log-from-compilation)
(bytecomp-tests--lexical-binding-cookie): New test.
Assuming 'number-or-marker' is a type (as present multiple times in
cl--typeof-types) adding some missing entries for coherency.
* lisp/emacs-lisp/cl-preloaded.el (cl--typeof-types): Add
'number-or-marker' as supertype of 'number' in the 'float' branch.
* lisp/emacs-lisp/cl-macs.el (cl-deftype-satisfies): Add
'number-or-marker'.
* test/lisp/emacs-lisp/comp-cstr-tests.el (comp-cstr-typespec-tests-alist):
Update test.
* test/src/comp-tests.el (comp-tests-type-spec-tests): Update two testes.
* test/lisp/emacs-lisp/comp-cstr-tests.el
(comp-cstr-typespec-tests-alist): Add four tests.
* lisp/emacs-lisp/comp-cstr.el (comp--sym-lessp)
(comp--direct-supertype, comp--normalize-typeset0): New functions.
(comp-normalize-typeset): Rework to make use of
'comp--normalize-typeset0'.
(comp--direct-subtypes): New function.
I.e. commit the needed change to the test suite.
* test/lisp/emacs-lisp/cl-print-tests.el
(cl-print-tests-ellipsis-string): Bind cl-print-string-length.
Only bind print-length and print-level where they are
specifically needed.
It is bad user experience when two menu items have identical labels.
* lisp/emacs-lisp/bytecomp.el (bytecomp--check-cus-type): Add check.
* test/lisp/emacs-lisp/bytecomp-tests.el
(bytecomp-test-defcustom-type): Add test case.
This fixes bug#65680. The former use of print-length as a
bound on the length of the printed string was erroneous,
causing error messages preceding backtracees to get
unnecessarily truncated to 50 characters.
* lisp/emacs-lisp/cl-print.el (cl-print-object <string>):
Remove the substitution of ellipses for long strings.
* test/lisp/emacs-lisp/cl-print-tests.el
(cl-print-tests-ellipsis-string): Remove this test.
Warn about more kinds of mistakes in :type arguments of `defcustom`
and `define-widget`. These include:
- misplaced keyword args, as in (const red :tag "A reddish hue")
- missing subordinate types, as in (repeat :tag "List of names")
or (choice list string)
- duplicated values, as in (choice (const yes) (const yes))
- misplaced `other` member, as in
(choice (const red) (other nil) (const blue))
- various type name mistakes, as in (vector bool functionp)
* lisp/emacs-lisp/bytecomp.el (byte-compile--defcustom-type-quoted)
(byte-compile-nogroup-warn): Remove.
(byte-compile-normal-call): Remove call to the above.
(bytecomp--cus-warn, bytecomp--check-cus-type)
(bytecomp--custom-declare): New.
* lisp/emacs-lisp/cl-macs.el (cl-defstruct): Split the first line of
generated docstrings if either the struct name or a field name is very
long. This reduces the likelihood of "docstring wider than 80
characters" errors. (Bug#65790)
* lisp/emacs-lisp/cl-macs.el (cl-defsubst): Split the first line of
the docstring into 2 lines if the function name is very long.
Additionally, remove the word "inlining" in the generated docstring as
it is not very useful and increases the likelihood of "docstring wider
than 80 characters" errors. (Bug#65790)
Warn about clauses after the default clause, as in
(cond ((= x 0) (say "none"))
(t (say "some"))
(say "goodbye"))
because they are very much an indicator of a mistake (such as
misplaced brackets), and since they are deleted by the optimiser, any
other warnings there are lost and the user wouldn't know that
something is wrong otherwise.
* lisp/emacs-lisp/macroexp.el (macroexp--expand-all): Add warning.
* etc/NEWS: Announce.
* test/lisp/emacs-lisp/bytecomp-tests.el
(bytecomp-test--with-suppressed-warnings): Add test case.
