* #3680: Read Aloud tidies in one linear pass and cuts space-free sentences
Finding 56. tidy() repeated whole-string replacements until nothing
changed; each pass removed one level of nested empty brackets, so 40,000
pairs took ~50 s in a debug build. It is now one pass that drops a closing
bracket's empty pair as it goes, checked against the old algorithm on
20,000 generated strings. Sentences with no space to split at are cut at
MAX_UTTERANCE (1,200) characters, and each sentence is sliced from a char
vector instead of skipping from the paragraph's start.
Fixes#3680
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
* #3692: Paste Special reads only a regular picture file, bounded
Finding 24. A copied path naming a device such as /dev/zero reports a
length of 0, so the size check passed and std::fs::read read forever.
The path and the opened handle must both be regular files, and at most
the picture limit + 1 bytes are read. A named pipe swapped in between the
check and the open can still block the open (no portable non-blocking
open without platform flags); noted in the code.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
* #3692: macOS signing import decodes the certificate into a private temp folder
Finding 25. The .p12 bundle was written to a predictable name in the
shared temp folder, where a planted link could redirect it. It now goes
into a mktemp -d folder under the same temp root, removed on exit. The
keychain is unchanged.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
* #3666: engine file reads open once and check the handle; a FIFO is refused without blocking
read_file_capped checked the path's metadata and then read the path
again, so a file swapped or grown after the check was read whole;
read_file_head opened the path with no type check, so saving to a FIFO
.xml path (the Word 2003 sniff) blocked until a writer appeared. Both
now go through open_regular_file (path checked before opening, handle
checked after, opened once) and a bounded read; read_regular_file uses
the same pieces. Tests: both helpers refuse a FIFO without blocking
and a folder; read_file_capped refuses content over the cap.
Fixes#3666
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
* #3692: Paste Special reads picture files through the engine's bounded read
Ultrareview finding 9: paste_picture kept a third copy of the regular-file
bounded read; it now calls wordcraft_engine::io::read_regular_file (public,
with a wasm stub).
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
* #3666: open a checked file non-blocking so a swap to a named pipe can't hang
CodeRabbit: a path replaced by a FIFO between the regular-file check and
File::open blocked in open, before the handle check could refuse it. On Unix the
open now uses O_NONBLOCK (no effect on regular files; libc is already in the
lockfile) and the handle check refuses the pipe. The FIFO tests' cleanup writers
open non-blocking too, so a regression fails the test instead of hanging it.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
* #3666: review — Paste Special refuses a picture file whose handle reports more than the cap before reading it
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01CX1hMKX28764Lrx5moxs1o
* #3692: review — macOS signing keychain lives in the private mktemp folder too; removed if the import fails
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01CX1hMKX28764Lrx5moxs1o
* #3666: review — libc comes from the workspace dependencies
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01CX1hMKX28764Lrx5moxs1o
* #3680: review — Read Aloud logs a warning when its sentence cap leaves text unread
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01CX1hMKX28764Lrx5moxs1o
* #3692: review — engine speech timing test gets a margin a busy CI machine can't break
tidy_is_linear_on_nested_brackets: 2 s -> 10 s. The one-pass tidy takes ~13 ms on a loaded
machine; tidy_by_replacing (the old tidy, kept in the tests) took 293 s on the same input.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01CX1hMKX28764Lrx5moxs1o
---------
Co-authored-by: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
* AutoFormat As You Type: optional period with double-space, off by default (#2848)
* Test the double-space option through the AutoCorrect dialog (#2848)
* File > Options > Proofing: an AutoCorrect Options button opens the dialog (#476)
* #2848: report doubleSpacePeriod in tools.autocorrect's params doc and result
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
---------
Co-authored-by: Brandon Thomas <echelon@gmail.com>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
* With Regular expressions on, Replace with reads $1…$n, ${n}, ${name}, $name,
$0/${0}, $$, \1…\9, \n (paragraph mark, splits the paragraph through the ^p
path), \t and \\, alongside the ^ codes. Named groups resolve through
meta::Regex::group_info(); missing, unknown and out-of-range groups expand to
nothing. Plain and wildcard Replace are unchanged.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
* #4177: review — special:false keeps regex group references
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
* #4177: review — \$ and \^ in regex Replace keep the character as typed
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Preview Results and Finish & Merge took any IF whose code held the word
MERGEFIELD anywhere, and with the type now matched in any case, a quoted
value counted too: IF { PAGE } = "mergefield" "yes" "no" was evaluated
as a rule on an empty field and showed "no". Both guards now ask what
condition() already reads: the first nested field, through
mergefield_name. An IF on other fields keeps its result.
SUBJECT Draft sets the subject to Draft, as the field-argument rules
allow a value without quotes when it has no white space (#4217). A
switch alone, as in TITLE \* Upper, still sets nothing.
`mergefield_name` matched with `strip_prefix("MERGEFIELD")`. A field typed as `mergefield City` was therefore no merge field at all: Check for Errors skipped it, and Preview Results and Finish & Merge left it alone. A code such as `MERGEFIELDX a` went the other way and was read as a merge field named `X`.
Field-type names ignore case (ECMA-376 Part 1, 17.16.1: "The field-type names DATE, Date, dAtE, and date are equivalent"). The other keywords in this file already go through the upper-cased `keyword()`. Now the type ends at the first space or quote and is compared ignoring case, which is the rule the shared HYPERLINK parser in #4222 follows. The two IF checks for a nested MERGEFIELD ignore case as well, so `IF { mergefield X } = ...` merges like its upper-case form.
Old and new compared on 5,940 generated codes: 2,760 give the same name, 960 are case variants that are now found, and 2,220 start with a longer word such as `MERGEFIELDX` that no longer counts as MERGEFIELD.
Test: merge_field_types_ignore_case.
NUMWORDS, NUMCHARS and REVNUM wrote a count past what the field's \*
number format covers as the format's last value, so a revision number
of 1,000,000 under \* ROMAN showed 50 M's. Such a count now keeps its
digits, as a table formula result does.
A field with \* Lower \* Caps applied only \* Lower, so a title of
WEEKLY REPORT showed as weekly report. Every capitalization switch in a
field code now applies in turn, and the example shows Weekly Report.
A row content control that the cut crosses reopened below the
converted rows with its w:id, so the saved file had two controls with
the same id. Table::split_off (#4162) clears the id and ends the
vertical merges the same way, so Convert to Text and Split Table
follow one rule. Rows without cells leave an empty paragraph.
A field or table formula with \* DollarText or \* Hex showed its digits,
and PAGE \* ArabicDash showed the plain page number. \* DollarText now
writes the result in lowercase cardinal text with "and" and the first two
decimals (rounded) over 100, so { = 9.20 + 5.35 \* DollarText \* Upper }
shows FOURTEEN AND 55/100. \* Hex writes uppercase hexadecimal digits
(458 shows 1CA). PAGE \* ArabicDash shows the page number between
hyphens ("- 31 -"); other fields, PAGEREF included, skip it. The
page-number, SEQ, PAGEREF and count fields and table formulas read both
switches. Cardinal, ordinal and dollar text keep the digits of a number
past 999,999. Before, \* CardText showed such a number as "nine hundred
ninety-nine thousand nine hundred ninety-nine" and \* OrdText as "nine
hundred ninety-nine thousand nine hundred ninety-ninth".
SUBJECT Draft left the field showing Draft and set the document's
subject to it. A TITLE, SUBJECT, AUTHOR, KEYWORDS or COMMENTS field now
replaces its property only when quoted text follows the field's name,
and otherwise shows the property as it is.
In a document that is read only or open to comments only, an exception
that starts a paragraph puts its start marker at the line start. Home,
Left, Right from the line above and Down into an empty line leave the
caret before that marker, outside the exception, so Enter and typing
there were refused. An editable line right after a protected line could
not be broken or typed into from its start. The same happened at the end
of a line that an exception ends, where End leaves the caret after the
end marker.
Typing now moves a caret right before the start of an exception the
author may edit to just inside it, and one right after its end back
inside it, as Document::typing_pos already does for the markers of a
block-level content control.
Update Field and the page-number fields ignored the capitalization
switches, so TITLE \* Upper or FILENAME \* Lower showed the text as
stored. A \* CardText or \* OrdText result started with a capital
("Twelve"), where Word shows it in lowercase unless a capitalization
switch follows. A table formula with a \* switch, such as
=SUM(A1:B2) \* CardText or =SUM(ABOVE) \* MERGEFORMAT, was a syntax
error.
The results of the document information fields, DOCPROPERTY,
FILENAME, the dates, REF, PAGEREF, SEQ, STYLEREF and table formulas
now take the field's \* Upper, \* Lower, \* Caps or \* FirstCap switch
when Update Field computes them. PAGE, NUMPAGES, SECTION,
SECTIONPAGES, PAGEREF and STYLEREF take it when they are laid out.
Cardinal and ordinal text is lowercase ("twelve", "seventh") unless
one of the switches follows. WordCraft now reads a table formula's \*
switches separately from the formula. A whole-number result that a \*
number format such as CardText or roman covers is written in that
format; any other result, and any other \* switch, leaves the digits.
Table Layout › Data › Formula keeps the \* switches as typed.
Update Field left the SUBJECT, KEYWORDS, COMMENTS, LASTSAVEDBY, REVNUM
and NUMCHARS fields at their stored result, so one inserted in
WordCraft showed nothing, and the Field dialog did not list them.
These fields now update with the others and are listed under Document
Information.
A quoted text after TITLE, SUBJECT, AUTHOR, KEYWORDS or COMMENTS
replaces that property, at most 255 characters, as in Word, and later
fields, DOCPROPERTY included, show the new value. NUMWORDS, NUMCHARS
and REVNUM take a \* number format.
With *Bold* and _italic_ as real formatting on, the opening mark is not
part of the word AutoCorrect reads, so typing `*teh end* ` gives bold
`The end ` instead of `teh end `. A sentence capital is decided from the
text before the mark.
Dylan M. Taylor