espeak: escape XML-special characters in single-character mode

When speakup echoes a single character, espeakup wraps it in SSML:

    <say-as interpret-as="characters">%c</say-as>

If the character is one of XML's five special characters (< > & ' "),
the raw byte is injected straight into the markup, producing ill-formed
SSML.  espeak-ng then misparses the <say-as> element, which surfaces as
a spurious high-pitched "ringing" whenever one of these characters is
spoken.

Escape each special character to its corresponding XML entity
(&lt; &gt; &amp; &apos; &quot;) so the markup stays well-formed.
Non-printable characters (< 0x20 or > 0x7e) are no longer wrapped in
SSML either; they fall through to the existing raw-synthesis fallback
so they cannot corrupt the surrounding element either.

The space and ordinary-printable-character paths are unchanged.

.clang-format

@@ -0,0 +1,41 @@
+---
+Language:        Cpp
+BasedOnStyle: LLVM
+
+AlignAfterOpenBracket: Align
+AlignEscapedNewlines: Left
+AlignOperands:   AlignAfterOperator
+AlignTrailingComments: true
+AllowAllArgumentsOnNextLine: true
+AllowAllParametersOfDeclarationOnNextLine: true
+AllowShortEnumsOnASingleLine: false
+AllowShortBlocksOnASingleLine: Never
+AllowShortCaseLabelsOnASingleLine: false
+AllowShortFunctionsOnASingleLine: None
+AllowShortIfStatementsOnASingleLine: Never
+AllowShortLoopsOnASingleLine: false
+AlwaysBreakAfterDefinitionReturnType: None
+AlwaysBreakAfterReturnType: None
+AlwaysBreakBeforeMultilineStrings: false
+BinPackArguments: true
+BinPackParameters: true
+BreakBeforeBraces: Linux
+ColumnLimit:     80
+IndentPPDirectives: BeforeHash
+IndentWidth:     4
+InsertTrailingCommas: None
+KeepEmptyLinesAtTheStartOfBlocks: false
+PointerAlignment: Right
+ReflowComments:  true
+SortIncludes:    true
+SpaceAfterCStyleCast: true
+SpaceAfterLogicalNot: false
+SpaceBeforeAssignmentOperators: true
+SpaceBeforeParens: ControlStatements
+SpaceAroundPointerQualifiers: Before
+SpaceInEmptyBlock: false
+SpaceInEmptyParentheses: false
+SpacesBeforeTrailingComments: 5
+TabWidth:        4
+UseTab:          AlignWithSpaces
+...

.gitignore

@@ -1,4 +1 @@
-espeakup
+build*
-*.d
-*.o
-*.swp

Makefile

@@ -1,45 +0,0 @@
-PREFIX = /usr/local
-BINDIR = ${PREFIX}/bin
-MANDIR = ${PREFIX}/share/man
-
-DEPFLAGS = -MMD
-WARNFLAGS = -Wall
-CFLAGS += ${DEPFLAGS} ${WARNFLAGS}
-
-LDLIBS = -lespeak -lpthread
-
-INSTALL = install
-BINMODE = 0755
-MANMODE = 0644
-CHANGELOG_LIMIT?= --after="1 year ago"
-
-SRCS = cli.c \
-	espeak.c \
-	espeakup.c  \
-	queue.c \
-	signal.c \
-	softsynth.c \
-	stringhandling.c
-
-OBJS = ${SRCS:.c=.o}
-
-all: espeakup
-
-changelog:
-	git log ${CHANGELOG_LIMIT} --format=full > ChangeLog
-
-install: espeakup
-	${INSTALL} -d ${DESTDIR}${BINDIR}
-	${INSTALL} -m ${BINMODE} $< ${DESTDIR}${BINDIR}
-	${INSTALL} -d ${DESTDIR}${MANDIR}/man8
-	${INSTALL} -m ${MANMODE} espeakup.8 ${DESTDIR}${MANDIR}/man8
-
-espeakup: ${OBJS}
-
-clean:
-	${RM} *.d *.o
-
-distclean: clean
-	${RM} espeakup
-
--include ${SRCS:.c=.d}

README

@@ -1,70 +0,0 @@
-espeakup connector
-=======================
-
-espeakup is a program which makes it possible for speakup to use
-the espeak software synthesizer.  It does this by reading speakup's
-softsynth device and passing the text to espeak which actually speaks.
-
-Requirements
-============
-
-This program works with the speakup screen reader, which can be obtained
-from http://linux-speakup.org, and the espeak software speech
-synthesizer which can be obtained from http://espeak.sourceforge.net.
-You must have both of these installed and operational.  Setting them up
-is beyond the scope of this document.
-
-Installation
-============
-
-The preferred way to install espeakup is using your distribution's
-packaging system, but if your distribution does not have a package for
-espeakup yet, espeakup just uses a Makefile, so you should be able to
-change to the source directory, then type make, then as root, make
-install.
-
-Starting Up
-===========
-
-This program should be run after speakup is set up to communicate with a
-software synthesizer and after /dev/softsynth exists.  The way this is
-done is distribution specific, so it is beyond the scope of this
-documentation.
-
-Command Line Options
-====================
-
-Espeakup currently accepts the following command line options:
-
-  --default-voice=voice, -V voice	Set default voice.
-  --debug, -d				Debug mode (stay in the foreground).
-  --help, -h				Show this help.
-  --version, -v				Display the software version.
-
-Getting the Latest Version
-==========================
-
-It is possible to download a tarball from github of any released version
-as follows:
-
-wget http://www.github.com/williamh/espeakup/tarball/vx.y
-
-If you need a tarball for packaging purposes, one is available from
-ftp://ftp.linux-speakup.org/pub/linux/goodies/espeakup-x.y.tar.bz2.
-
-The url for the git repository is git://github.com/williamh/espeakup.git.
-
-Acknowledgements
-================
-
-I would like to thank Marc Mulcahy, the author of the speakup to
-TTSynth connector, on which this work is based.  Also, I would like
-to thank Kirk Reiser and Jonathan Duddington, the
-authors of Speakup and Espeak, respectively, for their work.
-
-Questions
-=========
-
-You can contact me with questions, bugs, patches, etc, at
-w.d.hubbs@gmail.com or on the speakup mailing list.  I hope you find
-this software to be useful.

README.md

@@ -0,0 +1,48 @@
+# espeakup connector
+
+espeakup is a program which makes it possible for speakup to use
+the espeak-ng software synthesizer.  It does this by reading speakup's
+softsynth device and passing the text to espeak-ng which actually speaks.
+
+## Requirements
+
+This program works with the speakup screen reader, which can be obtained
+from http://linux-speakup.org, and the
+[espeak-ng](https://github.com/espeak-ng/espeak-ng) software speech
+synthesizer.
+
+You must have both of these installed and operational.  Setting them up
+is beyond the scope of this document.
+
+## Installation
+
+The preferred way to install espeakup is using your distribution's
+packaging system, but if your distribution does not have a package for
+espeakup yet, espeakup just uses meson, so you should be able to
+change to the source directory, then type:
+
+```bash
+meson setup . ./build
+cd ./build
+ninja
+sudo ninja install
+```
+
+## Starting Up
+
+This program should be run after speakup is set up to communicate with a
+software synthesizer and after /dev/softsynth exists.  The way this is
+done is distribution specific, so it is beyond the scope of this
+documentation.
+
+## Acknowledgements
+
+I would like to thank Marc Mulcahy, the author of the speakup to
+TTSynth connector, on which this work is based.  Also, I would like
+to thank Kirk Reiser and Jonathan Duddington, the
+authors of Speakup and Espeak, respectively, for their work.
+
+## Filing Bugs
+
+Bugs should be filed on our bug tracker at
+https://github.com/linux-speakup/issues.

TODO

@@ -1,11 +0,0 @@
-This file contains a list of improvements that need to be made to
-espeakup.
-
-Voice and language support should be added at some point.
-This will consist of the following:
-- add a function which will allow switching voices.  (completed on 9/27)
-- add a function/command which will allow speakup to find out which
-  voices espeak supports.
-- define and add the command to speakup which will allow it to switch
-  voices.
-

doc/espeakup.8.ronn

@@ -0,0 +1,77 @@
+<!-- markdownlint-disable MD036 -->
+
+# espeakup(8) --- connect Speakup to the espeak-ng TTS engine
+
+## SYNOPSIS
+
+`espeakup` [`--pid-path=`<path>] [`--alsa-volume`]
+[`--default-voice=`[<voicename>]] [`--debug`] [`--help`] [`--version`]
+
+## OPTIONS
+
+  * `-P` <path>, `--pid-path=`<path>:
+    Set the full path for the pid file espeakup uses when in daemon mode.
+
+  * `--alsa-volume`:
+    Drive the ALSA volume. useful for live environments where volume
+    adjustments maybe impossible.
+
+  * `-V` <voicename>, `--default-voice=`<voicename>:
+    Set the espeak-ng voice to be used by default.
+
+  * `-d`, `--debug`:
+    run in the foreground, rather than becoming a daemon process.
+
+  * `-h`, `--help`:
+    display a brief help message and exit.
+
+  * `-v`, `--version`:
+    output version information and exit.
+
+## DESCRIPTION
+
+espeakup bridges the gap between two tools: the Speakup screen review system and
+the espeak-ng text-to-speech engine. Each of these tools performs a
+well-defined task.
+
+Speakup is a kernel-based screen reader for the Linux console. It extracts and
+processes the text that is displayed on the foreground virtual console. It
+supports several hardware based speech synthesizers directly. However, since it
+is in kernel space, it cannot support a software speech synthesizer directly
+since these are in user space.
+
+espeak-ng is a popular software speech synthesizer. It is small, light weight,
+very responsive, and supports multiple languages. espeakup is a connector which
+will read text sent to it by speakup and forward it to espeak-ng. This allows
+Speakup to use espeak-ng as its speech synthesizer.
+
+espeakup is a daemon. Typically, it is started at boot time, and it terminates
+when the system is halted or rebooted. It should be started by the system's init
+scripts. This process varies among Linux distributions, but the details are
+usually managed by the person who packaged espeakup for your distribution. From
+the perspective of an average user, espeakup's operation is invisible.
+
+## BUGS
+
+If you find a bug, please create a
+[github issue](https://github.com/linux-speakup/espeakup/issues)
+You might also consider mentioning it on the mailing list for the Speakup
+screenreader. Visit
+[list page](https://linux-speakup.org/cgi-bin/mailman/listinfo/speakup)
+to learn more about the mailing list.
+
+## SEE ALSO
+
+For more information about Speakup, visit its
+[homepage](https://linux-speakup.org).
+
+espeak-ng can be found at [github](https://github.com/espeak-ng/espeak-ng)
+
+## AUTHOR
+
+William Hubbs <w.d.hubbs@gmail.com> is the author of espeakup.
+
+This manual page was written by Chris Brannon <chris@the-brannons.com>.
+
+current authors and maintainers can be found at
+[github](https://github.com/linux-speakup/espeakup/graphs/contributors)

doc/meson.build

@@ -0,0 +1,8 @@
+ronn = find_program('ronn', required: get_option('man'))
+if ronn.found()
+  custom_target('man',
+    input : files('espeakup.8.ronn'),
+    output : 'espeakup.8',
+    command : [ronn, '--output-dir', '@OUTDIR@', '--roff', '@INPUT@'],
+    install : true, install_dir: join_paths(get_option('mandir'),'man8'))
+endif

espeak.c

@@ -1,328 +0,0 @@
-/*
- *  espeakup - interface which allows speakup to use espeak
- *
- *  Copyright (C) 2008 William Hubbs
- *
- *  This program is free software: you can redistribute it and/or modify
- *  it under the terms of the GNU General Public License as published by
- *  the Free Software Foundation, either version 3 of the License, or
- *  (at your option) any later version.
- *
- *   This program is distributed in the hope that it will be useful,
- *   but WITHOUT ANY WARRANTY; without even the implied warranty of
- *   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
- *   GNU General Public License for more details.
- *
- *  You should have received a copy of the GNU General Public License
- *   along with this program.  If not, see <http://www.gnu.org/licenses/>.
- */
-
-#define _GNU_SOURCE
-#include <assert.h>
-#include <stdio.h>
-#include <stdlib.h>
-#include <string.h>
-
-#include "espeakup.h"
-
-/* default voice settings */
-const int defaultFrequency = 5;
-const int defaultPitch = 5;
-const int defaultRate = 2;
-const int defaultVolume = 5;
-char *defaultVoice = NULL;
-
-/* multipliers and offsets */
-const int frequencyMultiplier = 11;
-const int pitchMultiplier = 11;
-const int rateMultiplier = 41;
-const int rateOffset = 80;
-const int volumeMultiplier = 22;
-
-volatile int stop_requested = 0;
-
-static int acsint_callback(short *wav, int numsamples, espeak_EVENT * events)
-{
-	int i;
-	for (i = 0; events[i].type !=  espeakEVENT_LIST_TERMINATED; i++) {
-		if (events[i].type == espeakEVENT_MARK) {
-			int mark = atoi(events[i].id.name);
-			if ((mark < 0) || (mark > 255))
-				continue;
-			putchar(mark);
-			fflush(stdout);
-		}
-	}
-	return 0;
-}
-
-static espeak_ERROR set_frequency(struct synth_t *s, int freq,
-								  enum adjust_t adj)
-{
-	espeak_ERROR rc;
-
-	if (adj == ADJ_DEC)
-		freq = -freq;
-	if (adj != ADJ_SET)
-		freq += s->frequency;
-	rc = espeak_SetParameter(espeakRANGE, freq * frequencyMultiplier, 0);
-	if (rc == EE_OK)
-		s->frequency = freq;
-	return rc;
-}
-
-static espeak_ERROR set_pitch(struct synth_t *s, int pitch,
-							  enum adjust_t adj)
-{
-	espeak_ERROR rc;
-
-	if (adj == ADJ_DEC)
-		pitch = -pitch;
-	if (adj != ADJ_SET)
-		pitch += s->pitch;
-	rc = espeak_SetParameter(espeakPITCH, pitch * pitchMultiplier, 0);
-	if (rc == EE_OK)
-		s->pitch = pitch;
-	return rc;
-}
-
-static espeak_ERROR set_punctuation(struct synth_t *s, int punct,
-									enum adjust_t adj)
-{
-	espeak_ERROR rc;
-
-	if (adj == ADJ_DEC)
-		punct = -punct;
-	if (adj != ADJ_SET)
-		punct += s->punct;
-	rc = espeak_SetParameter(espeakPUNCTUATION, punct, 0);
-	if (rc == EE_OK)
-		s->punct = punct;
-	return rc;
-}
-
-static espeak_ERROR set_rate(struct synth_t *s, int rate,
-							 enum adjust_t adj)
-{
-	espeak_ERROR rc;
-
-	if (adj == ADJ_DEC)
-		rate = -rate;
-	if (adj != ADJ_SET)
-		rate += s->rate;
-	rc = espeak_SetParameter(espeakRATE,
-							 rate * rateMultiplier + rateOffset, 0);
-	if (rc == EE_OK)
-		s->rate = rate;
-	return rc;
-}
-
-static espeak_ERROR set_voice(struct synth_t *s, char *voice)
-{
-	espeak_ERROR rc;
-	espeak_VOICE voice_select;
-
-	rc = espeak_SetVoiceByName(voice);
-	if (rc != EE_OK)
-	{
-		memset(&voice_select, 0, sizeof(voice_select));
-		voice_select.languages = voice;
-		rc = espeak_SetVoiceByProperties(&voice_select);
-	}
-	if (rc == EE_OK)
-		strcpy(s->voice, voice);
-	return rc;
-}
-
-static espeak_ERROR set_volume(struct synth_t *s, int vol,
-							   enum adjust_t adj)
-{
-	espeak_ERROR rc;
-
-	if (adj == ADJ_DEC)
-		vol = -vol;
-	if (adj != ADJ_SET)
-		vol += s->volume;
-	rc = espeak_SetParameter(espeakVOLUME, (vol + 1) * volumeMultiplier,
-							 0);
-	if (rc == EE_OK)
-		s->volume = vol;
-
-	return rc;
-}
-
-static espeak_ERROR stop_speech(void)
-{
-	espeak_ERROR rc;
-
-	rc = espeak_Cancel();
-	return rc;
-}
-
-static espeak_ERROR speak_text(struct synth_t *s)
-{
-	espeak_ERROR rc;
-	int synth_mode = 0;
-
-	if (espeakup_mode == ESPEAKUP_MODE_ACSINT)
-		synth_mode |= espeakSSML;
-
-	if (espeakup_mode == ESPEAKUP_MODE_SPEAKUP && (s->len == 1)) {
-		char *buf;
-		int n;
-		n = asprintf(&buf,
-			     "<say-as interpret-as=\"characters\">%c</say-as>",
-			     s->buf[0]);
-		if (n == -1) {
-			/* D'oh.  Not much to do on allocation failure.
-			 * Perhaps espeak will happen to say the character */
-			rc = espeak_Synth(s->buf, s->len + 1, 0, POS_CHARACTER,
-					  0, synth_mode, NULL, NULL);
-		} else {
-			rc = espeak_Synth(buf, n + 1, 0, POS_CHARACTER, 0,
-					  espeakSSML, NULL, NULL);
-			free(buf);
-		}
-	} else
-		rc = espeak_Synth(s->buf, s->len + 1, 0, POS_CHARACTER, 0,
-				  synth_mode, NULL, NULL);
-	return rc;
-}
-
-static void free_espeak_entry(struct espeak_entry_t *entry)
-{
-	assert(entry);
-	if (entry->cmd == CMD_SPEAK_TEXT)
-		free(entry->buf);
-	free(entry);
-}
-
-static void synth_queue_clear()
-{
-	struct espeak_entry_t *current;
-
-	while (queue_peek(synth_queue)) {
-		current = (struct espeak_entry_t *) queue_remove(synth_queue);
-		free_espeak_entry(current);
-	}
-}
-
-static void queue_process_entry(struct synth_t *s)
-{
-	espeak_ERROR error;
-	static struct espeak_entry_t *current = NULL;
-
-	if (current != queue_peek(synth_queue)) {
-		if (current)
-			free_espeak_entry(current);
-		current = (struct espeak_entry_t *) queue_remove(synth_queue);
-	}
-	pthread_mutex_unlock(&queue_guard);
-	switch (current->cmd) {
-	case CMD_SET_FREQUENCY:
-		error = set_frequency(s, current->value, current->adjust);
-		break;
-	case CMD_SET_PITCH:
-		error = set_pitch(s, current->value, current->adjust);
-		break;
-	case CMD_SET_PUNCTUATION:
-		error = set_punctuation(s, current->value, current->adjust);
-		break;
-	case CMD_SET_RATE:
-		error = set_rate(s, current->value, current->adjust);
-		break;
-	case CMD_SET_VOICE:
-		error = EE_OK;
-		break;
-	case CMD_SET_VOLUME:
-		error = set_volume(s, current->value, current->adjust);
-		break;
-	case CMD_SPEAK_TEXT:
-		s->buf = current->buf;
-		s->len = current->len;
-		error = speak_text(s);
-		break;
-	default:
-		break;
-	}
-
-	if (error == EE_OK) {
-		free_espeak_entry(current);
-		current = NULL;
-	}
-}
-
-int initialize_espeak(struct synth_t *s)
-{
-	int rate;
-
-	/* initialize espeak */
-	rate = espeak_Initialize(AUDIO_OUTPUT_PLAYBACK, 50, NULL, 0);
-	if (rate < 0) {
-		fprintf(stderr, "Unable to initialize espeak.\n");
-		return -1;
-	}
-
-	/* We need a callback in acsint mode, but not in speakup mode. */
-	if (espeakup_mode == ESPEAKUP_MODE_ACSINT)
-		espeak_SetSynthCallback(acsint_callback);
-
-	/* Setup initial voice parameters */
-	if (defaultVoice) {
-		set_voice(s, defaultVoice);
-		free(defaultVoice);
-		defaultVoice = NULL;
-	}
-	set_frequency(s, defaultFrequency, ADJ_SET);
-	set_pitch(s, defaultPitch, ADJ_SET);
-	set_rate(s, defaultRate, ADJ_SET);
-	set_volume(s, defaultVolume, ADJ_SET);
-	espeak_SetParameter(espeakCAPITALS, 0, 0);
-	return 0;
-}
-
-/* espeak_thread is the "main" function of our secondary (queue-processing)
- * thread.
- * First, lock queue_guard, because it needs to be locked when we call
- * pthread_cond_wait on the runner_awake condition variable.
- * Next, enter an infinite loop.
- * The wait call also unlocks queue_guard, so that the other thread can
- * manipulate the queue.
- * When runner_awake is signaled, the pthread_cond_wait call re-locks
- * queue_guard, and the "queue processor" thread has access to the queue.
- * While there is an entry in the queue, call queue_process_entry.
- * queue_process_entry unlocks queue_guard after removing an item from the
- * queue, so that the main thread doesn't have to wait for us to finish
- * processing the entry.  So re-lock queue_guard after each call to
- * queue_process_entry.
- *
- * The main thread can add items to the queue in exactly two situations:
- * 1. We are waiting on runner_awake, or
- * 2. We are processing an entry that has just been removed from the queue.
-*/
-void *espeak_thread(void *arg)
-{
-	struct synth_t *s = (struct synth_t *) arg;
-
-	pthread_mutex_lock(&queue_guard);
-	while (should_run) {
-
-		while (should_run && !queue_peek(synth_queue) && !stop_requested)
-			pthread_cond_wait(&runner_awake, &queue_guard);
-
-		if (stop_requested) {
-			stop_speech();
-			synth_queue_clear();
-			stop_requested = 0;
-			pthread_cond_signal(&stop_acknowledged);
-		}
-
-		while (should_run && queue_peek(synth_queue) && !stop_requested) {
-			queue_process_entry(s);
-			pthread_mutex_lock(&queue_guard);
-		}
-	}
-	pthread_cond_signal(&stop_acknowledged);
-	pthread_mutex_unlock(&queue_guard);
-	return NULL;
-}

espeakup.8

@@ -1,78 +0,0 @@
-.\" Hey, Emacs!  This is an -*- nroff -*- source file.
-.\" Espeakup is Copyright 2008 by William Hubbs.
-.\" This is free software; see the GNU General Public Licence version 3
-.\" or later for copying conditions.  There is NO warranty.
-.TH ESPEAKUP 8 "5 Nov 2008" "0.60"
-.nh
-.SH NAME
-espeakup \(em connect Speakup to the ESpeak TTS engine
-.SH SYNOPSIS
-.B espeakup
-[
-.B \-\^\-pid-path=path
-]
-[
-.B \-\^\-default-voice=voicename
-]
-[
-.B \-\^\-debug
-]
-[
-.B \-\^\-help
-]
-[
-.B \-\^\-version
-]
-.SH OPTIONS
-.TP
-.B \-P path, \-\^\-pid-path=path
-Set the full path for the pid file espeakup uses when in daemon mode.
-.TP
-.B \-V voicename, \-\^\-default-voice=voicename
-Set the espeak voice to be used by default.
-.TP
-.B \-d, \-\^\-debug
-run in the foreground, rather than becoming a daemon process.
-.TP
-.B \-h, \-\^\-help
-display a brief help message and exit.
-.TP
-.B \-v, \-\^\-version
-output version information and exit.
-.SH DESCRIPTION
-Espeakup bridges the gap between two tools: the Speakup screen review
-system and the ESppeak text-to-speech engine.  Each of these tools
-performs a well-defined task.  Speakup is a kernel-based screen reader
-for the Linux console.  It extracts and processes the text that is
-displayed on the foreground virtual console.  It supports several
-hardware based speech synthesizers directly.  However, since it is in
-kernel space, it cannot support a software speech synthesizer directly
-since these are in user space.
-ESpeak is a popular software speech synthesizer.  It is small, light
-weight, very responsive, and supports multiple languages.
-Espeakup is a connector which will read text sent to it by speakup and
-forward it to ESpeak.  This allows Speakup to use ESpeak as its speech
-synthesizer.
-.PP
-Espeakup is a daemon.  Typically, it is started at boot time, and it terminates
-when the system is halted or rebooted.  It should be started by the
-system's init scripts.  This process varies among Linux distributions,
-but the details are usually managed by the person who packaged Espeakup for
-your distribution.
-From the perspective of an average user, Espeakup's operation is invisible.
-.SH BUGS
-.PP
-Espeakup is still classified as alpha software.  Bugs are periodically found
-and fixed.  If you find a bug, please do report it to the author.  You
-might also consider mentioning it on the mailing list for the Speakup
-screenreader.  Visit http://speech.braille.uwo.ca/mailman/listinfo/speakup
-to learn more about the mailing list.
-.SH SEE ALSO
-.PP
-For more information about Speakup, visit its homepage: http://linux-speakup.org.
-ESpeak's home page is http://espeak.sourceforge.net.
-.SH AUTHOR
-.PP
-William Hubbs is the author and maintainer of Espeakup.  He may be reached
-via the email address <w.d.hubbs@gmail.com>.  This manual page was written
-by Chris Brannon, and his email address is <cmbrannon79@gmail.com>.

meson.build

@@ -0,0 +1,25 @@
+project('espeakup', 'c',
+  default_options : [
+    'buildtype=debugoptimized',
+    'c_std=gnu11',
+    'warning_level=3'
+    ],
+  license : 'GPL-3.0-or-later',
+  version : '0.90',
+  meson_version : '>=0.51.0')
+
+cc = meson.get_compiler('c')
+thread_dep = dependency('threads')
+espeak_dep = dependency('espeak-ng')
+alsa_dep = dependency('alsa')
+math_dep = cc.find_library('m', required : false)
+
+subdir('doc')
+subdir('services')
+subdir('src')
+
+executable('espeakup',
+  espeakup_version,
+  espeakup_sources,
+  dependencies : [thread_dep, espeak_dep, alsa_dep, math_dep],
+  install : true)

meson_options.txt

@@ -0,0 +1,4 @@
+option('man', type : 'feature', value : 'auto',
+       description : 'build manpage with ronn')
+option('systemd', type : 'feature', value : 'auto',
+       description :'enable systemd support')

services/meson.build

@@ -0,0 +1,4 @@
+systemd = dependency('systemd')
+if (systemd.found() and get_option('systemd').allowed()) or get_option('systemd').enabled()
+    subdir('systemd')
+endif

services/systemd/espeakup.service.in

@@ -0,0 +1,18 @@
+[Unit]
+Description=Software speech output for Speakup
+Documentation=man:espeakup(8)
+Wants=modprobe@speakup_soft.service
+After=modprobe@speakup_soft.service sound.target
+
+[Service]
+Type=forking
+PIDFile=/run/espeakup.pid
+Environment="default_voice="
+ExecStart=@bindir@/espeakup --default-voice=${default_voice}
+ExecReload=kill -HUP $MAINPID
+Restart=always
+Nice=-10
+OOMScoreAdjust=-900
+
+[Install]
+WantedBy=sound.target

services/systemd/meson.build

@@ -0,0 +1,20 @@
+if systemd.found()
+  unitdir = systemd.get_variable(pkgconfig: 'systemdsystemunitdir')
+else
+  unitdir = join_paths(prefixdir, get_option('libdir'), 'systemd', 'system')
+endif
+prefixdir = get_option('prefix')
+bindir = join_paths(prefixdir, get_option('bindir'))
+
+unit_conf = configuration_data()
+unit_conf.set('bindir', bindir)
+
+service_file = configure_file(
+  input : 'espeakup.service.in',
+  output : 'espeakup.service',
+  configuration : unit_conf
+)
+
+install_data(service_file,
+  install_dir : unitdir
+)

src/cli.c

@@ -1,5 +1,5 @@
 /*
- *  espeakup - interface which allows speakup to use espeak
+ *  espeakup - interface which allows speakup to use espeak-ng
  *
  *  Copyright (C) 2008 William Hubbs
  *
@@ -23,6 +23,8 @@
 #include <string.h>
 
 #include "espeakup.h"
+#include "stringhandling.h"
+#include "version.h"
 
 /* pid path */
 extern char *pidPath;
@@ -30,17 +32,20 @@ extern char *pidPath;
 /* default voice */
 extern char *defaultVoice;
 
+/* Whether to drive ALSA volume */
+extern int alsaVolume;
+
 /* command line options */
 const char *shortOptions = "P:V:adhv";
 const struct option longOptions[] = {
 	{"pid-path", required_argument, NULL, 'P'},
 	{"default-voice", required_argument, NULL, 'V'},
+	{"alsa-volume", no_argument, &alsaVolume, 1},
 	{"acsint", no_argument, NULL, 'a'},
 	{"debug", no_argument, NULL, 'd'},
 	{"help", no_argument, NULL, 'h'},
 	{"version", no_argument, NULL, 'v'},
-	{0, 0, 0, 0}
+	{0, 0, 0, 0}};
-};
 
 static void show_help()
 {
@@ -48,6 +53,7 @@ static void show_help()
 	printf("Options are as follows:\n");
 	printf("  --pid-path=path, -P path\t\tSet path for pid file.\n");
 	printf("  --default-voice=voice, -V voice\tSet default voice.\n");
+	printf("  --alsa-volume\t\t\t\tDrive the ALSA volume.\n");
 	printf("  --debug, -d\t\t\t\tDebug mode (stay in the foreground).\n");
 	printf("  --help, -h\t\t\t\tShow this help.\n");
 	printf("  --version, -v\t\t\t\tDisplay the software version.\n");
@@ -67,18 +73,15 @@ static void show_version(void)
 void process_cli(int argc, char **argv)
 {
 	int opt;
-	char *cp;
 
 	do {
 		opt = getopt_long(argc, argv, shortOptions, longOptions, NULL);
 		switch (opt) {
-		case 'p':
+		case 'P':
-			cp = strdup(optarg);
+			pidPath = dupeString(optarg);
-			if (cp != NULL)
-				pidPath = cp;
 			break;
 		case 'V':
-			defaultVoice = strdup(optarg);
+			defaultVoice = dupeString(optarg);
 			break;
 		case 'a':
 			espeakup_mode = ESPEAKUP_MODE_ACSINT;
@@ -93,6 +96,7 @@ void process_cli(int argc, char **argv)
 			show_version();
 			break;
 		case -1:
+		case 0:
 			break;
 		default:
 			show_help();

src/espeak.c

@@ -0,0 +1,638 @@
+/*
+ *  espeakup - interface which allows speakup to use espeak-ng
+ *
+ *  Copyright (C) 2008 William Hubbs
+ *
+ *  This program is free software: you can redistribute it and/or modify
+ *  it under the terms of the GNU General Public License as published by
+ *  the Free Software Foundation, either version 3 of the License, or
+ *  (at your option) any later version.
+ *
+ *   This program is distributed in the hope that it will be useful,
+ *   but WITHOUT ANY WARRANTY; without even the implied warranty of
+ *   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ *   GNU General Public License for more details.
+ *
+ *  You should have received a copy of the GNU General Public License
+ *   along with this program.  If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#define _GNU_SOURCE
+#include <alsa/asoundlib.h>
+#include <assert.h>
+#include <math.h>
+#include <stdatomic.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <time.h>
+#include <unistd.h>
+
+#include "espeakup.h"
+
+/* default voice settings */
+const int defaultFrequency = 5;
+const int defaultPitch = 5;
+const int defaultRange = 5;
+const int defaultRate = 2;
+const int defaultVolume = 5;
+char *defaultVoice = NULL;
+int alsaVolume = 0;
+
+/* multipliers and offsets */
+const int frequencyMultiplier = 11;
+const int pitchMultiplier = 11;
+const int rangeMultiplier = 11;
+const int rateMultiplier = 41;
+const int rateOffset = 80;
+const int volumeMultiplier = 22;
+
+volatile int stop_requested = 0;
+int paused_espeak = 1;
+
+/* Wedged-engine detection.  Espeak may legitimately refuse entries for a
+ * while (EE_BUFFER_FULL while a long backlog is being played back), so a
+ * failing entry is normally just retried.  But when entries keep failing
+ * while the synth callback reports no progress at all, the audio output
+ * is most likely wedged (e.g. an ALSA device stuck returning EBUSY).
+ * After ESPEAK_STALL_RETRIES consecutive such retries (about one second
+ * each), restart the engine; after ESPEAK_MAX_RESTARTS restarts without
+ * the engine having been healthy for ESPEAK_HEALTHY_SECS in between,
+ * give up and exit, so that the init system can respawn us. */
+#define ESPEAK_STALL_RETRIES 10
+#define ESPEAK_MAX_RESTARTS 3
+#define ESPEAK_HEALTHY_SECS 60
+
+/* Set by the synth callback whenever espeak makes synthesis progress;
+ * used to tell a merely backlogged engine from a wedged one.  The
+ * callback runs in espeak's own thread, so the flag is atomic. */
+static atomic_int synth_progressed = 0;
+static int stalled_retries = 0;
+static int restart_attempts = 0;
+static struct timespec last_restart;
+
+static int callback(short *wav, int numsamples, espeak_EVENT *events)
+{
+	int i;
+	atomic_store(&synth_progressed, 1);
+	for (i = 0; events[i].type != espeakEVENT_LIST_TERMINATED; i++) {
+		if (events[i].type == espeakEVENT_MARK) {
+			int mark = atoi(events[i].id.name);
+			if ((mark < 0) || (mark > 255))
+				continue;
+			softsynth_reportindex(mark);
+		}
+	}
+	return 0;
+}
+
+static espeak_ERROR set_frequency(struct synth_t *s, int freq,
+                                  enum adjust_t adj)
+{
+	espeak_ERROR rc;
+
+	if (adj == ADJ_DEC)
+		freq = -freq;
+	if (adj != ADJ_SET)
+		freq += s->frequency;
+	rc = espeak_SetParameter(espeakRANGE, freq * frequencyMultiplier, 0);
+	if (rc == EE_OK)
+		s->frequency = freq;
+	return rc;
+}
+
+static espeak_ERROR set_pitch(struct synth_t *s, int pitch, enum adjust_t adj)
+{
+	espeak_ERROR rc;
+
+	if (adj == ADJ_DEC)
+		pitch = -pitch;
+	if (adj != ADJ_SET)
+		pitch += s->pitch;
+	rc = espeak_SetParameter(espeakPITCH, pitch * pitchMultiplier, 0);
+	if (rc == EE_OK)
+		s->pitch = pitch;
+	return rc;
+}
+
+static espeak_ERROR set_range(struct synth_t *s, int range, enum adjust_t adj)
+{
+	espeak_ERROR rc;
+
+	if (adj == ADJ_DEC)
+		range = -range;
+	if (adj != ADJ_SET)
+		range += s->range;
+	rc = espeak_SetParameter(espeakRANGE, range * rangeMultiplier, 0);
+	if (rc == EE_OK)
+		s->range = range;
+	return rc;
+}
+
+static espeak_ERROR set_punctuation(struct synth_t *s, int punct,
+                                    enum adjust_t adj)
+{
+	espeak_ERROR rc;
+	espeak_PUNCT_TYPE espeak_punct;
+
+	if (adj == ADJ_DEC)
+		punct = -punct;
+	if (adj != ADJ_SET)
+		punct += s->punct;
+
+	switch (punct) {
+		case 0:
+			espeak_punct = espeakPUNCT_NONE;
+			break;
+		case 1:
+			espeak_punct = espeakPUNCT_SOME;
+			break;
+		case 2:
+			/* XXX: approximation */
+			espeak_punct = espeakPUNCT_SOME;
+			break;
+		case 3:
+		default:
+			espeak_punct = espeakPUNCT_ALL;
+			break;
+	}
+
+	rc = espeak_SetParameter(espeakPUNCTUATION, espeak_punct, 0);
+	if (rc == EE_OK)
+		s->punct = punct;
+	return rc;
+}
+
+static espeak_ERROR set_rate(struct synth_t *s, int rate, enum adjust_t adj)
+{
+	espeak_ERROR rc;
+
+	if (adj == ADJ_DEC)
+		rate = -rate;
+	if (adj != ADJ_SET)
+		rate += s->rate;
+	rc = espeak_SetParameter(espeakRATE, rate * rateMultiplier + rateOffset, 0);
+	if (rc == EE_OK)
+		s->rate = rate;
+	return rc;
+}
+
+static espeak_ERROR set_voice(struct synth_t *s, char *voice)
+{
+	espeak_ERROR rc;
+	espeak_VOICE voice_select;
+
+	rc = espeak_SetVoiceByName(voice);
+	if (rc != EE_OK) {
+		memset(&voice_select, 0, sizeof(voice_select));
+		voice_select.languages = voice;
+		rc = espeak_SetVoiceByProperties(&voice_select);
+	}
+	if (rc == EE_OK)
+		strcpy(s->voice, voice);
+	return rc;
+}
+
+static void set_alsa_volume(int vol)
+{
+	snd_mixer_t *m;
+	snd_mixer_elem_t *e;
+	int err;
+
+	err = snd_mixer_open(&m, 0);
+	if (err < 0) {
+		fprintf(stderr, "ALSA mixer open error: %s\n", snd_strerror(err));
+		return;
+	}
+
+	err = snd_mixer_attach(m, "default");
+	if (err < 0) {
+		fprintf(stderr, "ALSA mixer attach error: %s\n", snd_strerror(err));
+		return;
+	}
+	err = snd_mixer_selem_register(m, NULL, NULL);
+	if (err < 0) {
+		fprintf(stderr, "ALSA mixer load error: %s\n", snd_strerror(err));
+		return;
+	}
+	err = snd_mixer_load(m);
+	if (err < 0) {
+		fprintf(stderr, "ALSA mixer load error: %s\n", snd_strerror(err));
+		return;
+	}
+
+	/* Turn vol value to volume %.
+	 * We do not want to soften that much with ALSA, espeak is already
+	 * doing it. We want the default value (5) to be the usual default
+	 * volume (80%), and make higher values increase ALSA volume, up to
+	 * 100%. */
+
+	int volume = (vol + 1) * 50 / 10 + 50;
+
+	for (e = snd_mixer_first_elem(m); e; e = snd_mixer_elem_next(e)) {
+		if (snd_mixer_elem_get_type(e) != SND_MIXER_ELEM_SIMPLE)
+			continue;
+		if (snd_mixer_selem_is_enumerated(e))
+			continue;
+
+		if (snd_mixer_selem_has_playback_switch(e)) {
+			snd_mixer_selem_set_playback_switch_all(e, 1);
+		}
+
+		if (snd_mixer_selem_has_playback_volume(e)) {
+			long min, max, set;
+
+			err = snd_mixer_selem_get_playback_dB_range(e, &min, &max);
+			if (err == 0 && min < max) {
+				if (max - min < 2400) {
+					/* 24dB amplitude is too small for using a logscale */
+					set = min + volume * (max - min) / 100;
+				} else {
+					/* Use a logscale */
+					double volf = volume / 100.;
+					if (min != SND_CTL_TLV_DB_GAIN_MUTE) {
+						double minf = pow(10, (min - max) / 6000.);
+						volf = volf * (1 - minf) + minf;
+					}
+					set = 6000. * log10(volf) + max;
+				}
+				snd_mixer_selem_set_playback_dB_all(e, set, 0);
+			} else {
+				/* No dB setting, try a linear scale */
+				snd_mixer_selem_get_playback_volume_range(e, &min, &max);
+				set = min + volume * (max - min) / 100;
+				snd_mixer_selem_set_playback_volume_all(e, set);
+			}
+		}
+	}
+}
+
+static espeak_ERROR set_volume(struct synth_t *s, int vol, enum adjust_t adj)
+{
+	espeak_ERROR rc;
+
+	if (adj == ADJ_DEC)
+		vol = -vol;
+	if (adj != ADJ_SET)
+		vol += s->volume;
+	rc = espeak_SetParameter(espeakVOLUME, (vol + 1) * volumeMultiplier, 0);
+	if (rc == EE_OK) {
+		s->volume = vol;
+		if (alsaVolume)
+			set_alsa_volume(vol);
+	}
+
+	return rc;
+}
+
+static espeak_ERROR stop_speech(void)
+{
+	espeak_ERROR rc;
+
+	rc = espeak_Cancel();
+	return rc;
+}
+
+static espeak_ERROR speak_text(struct synth_t *s)
+{
+	espeak_ERROR rc;
+	int synth_mode = 0;
+
+	if (espeakup_mode == ESPEAKUP_MODE_ACSINT)
+		synth_mode |= espeakSSML;
+
+	if (espeakup_mode == ESPEAKUP_MODE_SPEAKUP && (s->len == 1)) {
+		char *buf = NULL;
+		int n;
+		unsigned char c = s->buf[0];
+		if (c == ' ')
+			n = asprintf(&buf,
+			             "<say-as interpret-as=\"tts:char\">&#32;</say-as>");
+		else if (c < 0x20 || c > 0x7e) {
+			/* Not a printable character; do not embed it in SSML, as
+			 * that would produce invalid markup.  Fall through to the
+			 * raw-synthesis path below. */
+			n = -1;
+		} else {
+			/* Escape characters that are special in XML/SSML so the
+			 * resulting markup stays well-formed; otherwise espeak-ng
+			 * misparses the <say-as> element, which is the source of
+			 * the spurious high-pitched "ringing" on these characters. */
+			const char *entity = NULL;
+			switch (c) {
+			case '<': entity = "&lt;"; break;
+			case '>': entity = "&gt;"; break;
+			case '&': entity = "&amp;"; break;
+			case '\'': entity = "&apos;"; break;
+			case '"': entity = "&quot;"; break;
+			}
+			if (entity)
+				n = asprintf(&buf,
+				             "<say-as interpret-as=\"characters\">%s</say-as>",
+				             entity);
+			else
+				n = asprintf(&buf,
+				             "<say-as interpret-as=\"characters\">%c</say-as>",
+				             c);
+		}
+		if (n == -1) {
+			/* D'oh.  Not much to do on allocation failure.
+			 * Perhaps espeak will happen to say the character */
+			rc = espeak_Synth(s->buf, s->len + 1, 0, POS_CHARACTER, 0,
+			                  synth_mode, NULL, NULL);
+		} else {
+			rc = espeak_Synth(buf, n + 1, 0, POS_CHARACTER, 0, espeakSSML, NULL,
+			                  NULL);
+			free(buf);
+		}
+	} else
+		rc = espeak_Synth(s->buf, s->len + 1, 0, POS_CHARACTER, 0, synth_mode,
+		                  NULL, NULL);
+	return rc;
+}
+
+static void free_espeak_entry(struct espeak_entry_t *entry)
+{
+	assert(entry);
+	if (entry->cmd == CMD_SPEAK_TEXT)
+		free(entry->buf);
+	free(entry);
+}
+
+static void synth_queue_clear()
+{
+	struct espeak_entry_t *current;
+
+	while (queue_peek(synth_queue)) {
+		current = (struct espeak_entry_t *) queue_remove(synth_queue);
+		free_espeak_entry(current);
+	}
+}
+
+static int reinitialize_espeak(struct synth_t *s)
+{
+	int rate;
+
+	/* Re-initialize espeak */
+	rate = espeak_Initialize(AUDIO_OUTPUT_PLAYBACK, 0, NULL, 0);
+	if (rate < 0) {
+		fprintf(stderr, "Unable to initialize espeak.\n");
+		return -1;
+	}
+
+	espeak_SetSynthCallback(callback);
+
+	/* Set parameters again */
+	espeak_SetVoiceByName(s->voice);
+	espeak_SetParameter(espeakRANGE, s->frequency * frequencyMultiplier, 0);
+	espeak_SetParameter(espeakPITCH, s->pitch * pitchMultiplier, 0);
+	espeak_SetParameter(espeakRATE, s->rate * rateMultiplier + rateOffset, 0);
+	espeak_SetParameter(espeakVOLUME, (s->volume + 1) * volumeMultiplier, 0);
+	espeak_SetParameter(espeakCAPITALS, 0, 0);
+	paused_espeak = 0;
+	return 0;
+}
+
+/* Wait for up to a second before retrying an entry which could not be
+ * processed, so that we do not busy-loop on a persistent error.  Called
+ * and returns with queue_guard held.  Wakes up immediately if a stop is
+ * requested. */
+static void espeak_wait_retry(void)
+{
+	struct timespec timeout;
+
+	clock_gettime(CLOCK_MONOTONIC, &timeout);
+	timeout.tv_sec++;
+	pthread_cond_timedwait(&wake_stop, &queue_guard, &timeout);
+}
+
+/* Handle an entry which could not be processed.  Called and returns
+ * with queue_guard held.
+ * Normally just back off before the retry, but watch out for a wedged
+ * engine: if entries keep failing while the synth callback shows no
+ * progress at all, restart the engine, and if restarting does not help
+ * either, exit so that the init system respawns us in a clean state. */
+static void espeak_handle_failure(struct synth_t *s)
+{
+	if (atomic_exchange(&synth_progressed, 0)) {
+		/* Espeak is making progress, it is merely backlogged. */
+		stalled_retries = 0;
+	} else if (++stalled_retries >= ESPEAK_STALL_RETRIES) {
+		stalled_retries = 0;
+		if (++restart_attempts > ESPEAK_MAX_RESTARTS) {
+			fprintf(stderr, "espeakup: espeak keeps failing without "
+			        "making progress and restarting it did not help, "
+			        "aborting\n");
+			/* Use _exit because exit could hang in library destructors
+			 * while the audio device is wedged. */
+			_exit(3);
+		}
+		fprintf(stderr, "espeakup: espeak has been failing without "
+		        "making progress for %d seconds, restarting it\n",
+		        ESPEAK_STALL_RETRIES);
+		/* Call into espeak with queue_guard released: these calls can
+		 * take time, or block on a wedged audio device.  If they do
+		 * block forever, the stop-acknowledgement timeout in the
+		 * softsynth thread is our last resort. */
+		pthread_mutex_unlock(&queue_guard);
+		if (!paused_espeak) {
+			espeak_Cancel();
+			espeak_Terminate();
+			paused_espeak = 1;
+		}
+		reinitialize_espeak(s);
+		clock_gettime(CLOCK_MONOTONIC, &last_restart);
+		pthread_mutex_lock(&queue_guard);
+		return;
+	}
+
+	espeak_wait_retry();
+}
+
+static struct espeak_entry_t *current = NULL;
+static void queue_process_entry(struct synth_t *s)
+{
+	espeak_ERROR error = EE_OK;
+	char markbuff[50];
+
+	if (current != queue_peek(synth_queue)) {
+		if (current)
+			free_espeak_entry(current);
+		current = queue_peek(synth_queue);
+	}
+	pthread_mutex_unlock(&queue_guard);
+
+	if (current->cmd != CMD_PAUSE && paused_espeak) {
+		if (reinitialize_espeak(s) < 0) {
+			/* Espeak is unavailable, so the entry cannot be processed.
+			 * Calling espeak functions on a terminated engine would
+			 * just fail (or worse).  Leave the entry queued and retry
+			 * after a small pause. */
+			pthread_mutex_lock(&queue_guard);
+			espeak_handle_failure(s);
+			return;
+		}
+	}
+
+	switch (current->cmd) {
+	case CMD_SET_FREQUENCY:
+		error = set_frequency(s, current->value, current->adjust);
+		break;
+	case CMD_SET_MARK:
+		snprintf(markbuff, sizeof(markbuff), "<mark name=\"%d\"/>",
+		         current->value);
+		error = espeak_Synth(markbuff, strlen(markbuff) + 1, 0, POS_CHARACTER,
+		                     0, espeakSSML, NULL, NULL);
+		break;
+	case CMD_SET_PITCH:
+		error = set_pitch(s, current->value, current->adjust);
+		break;
+	case CMD_SET_RANGE:
+		error = set_range(s, current->value, current->adjust);
+		break;
+	case CMD_SET_PUNCTUATION:
+		error = set_punctuation(s, current->value, current->adjust);
+		break;
+	case CMD_SET_RATE:
+		error = set_rate(s, current->value, current->adjust);
+		break;
+	case CMD_SET_VOICE:
+		error = EE_OK;
+		break;
+	case CMD_SET_VOLUME:
+		error = set_volume(s, current->value, current->adjust);
+		break;
+	case CMD_SPEAK_TEXT:
+		s->buf = current->buf;
+		s->len = current->len;
+		error = speak_text(s);
+		break;
+	case CMD_PAUSE:
+		if (!paused_espeak) {
+			error = espeak_Cancel();
+			if (error == EE_OK)
+				error = espeak_Terminate();
+			if (error == EE_OK)
+				paused_espeak = 1;
+		} else {
+			error = EE_OK;
+		}
+		break;
+	default:
+		/* Uh? */
+		error = EE_OK;
+		break;
+	}
+
+	pthread_mutex_lock(&queue_guard);
+	if (error == EE_OK) {
+		/* Processed, drop it */
+		struct espeak_entry_t *unqueued = queue_remove(synth_queue);
+		assert(unqueued == current);
+		free_espeak_entry(current);
+		current = NULL;
+		stalled_retries = 0;
+		if (restart_attempts) {
+			/* Forget about past restarts once the engine has been
+			 * healthy for a while.  Entries can spuriously succeed
+			 * right after a restart while the audio output is still
+			 * wedged (espeak's internal queue is empty again), so a
+			 * quick success must not reset the counter. */
+			struct timespec now;
+			clock_gettime(CLOCK_MONOTONIC, &now);
+			if (now.tv_sec - last_restart.tv_sec >= ESPEAK_HEALTHY_SECS)
+				restart_attempts = 0;
+		}
+	} else {
+		if (error != EE_BUFFER_FULL)
+			fprintf(stderr, "espeak error: %d\n", error);
+		/* The entry stays queued and will be retried.  Give espeak a
+		 * little break before that, whatever the error: previously only
+		 * EE_BUFFER_FULL throttled, and any other persistent error made
+		 * us retry the same entry in a tight loop, burning a whole CPU. */
+		espeak_handle_failure(s);
+	}
+}
+
+int initialize_espeak(struct synth_t *s)
+{
+	int rate;
+
+	/* initialize espeak */
+	rate = espeak_Initialize(AUDIO_OUTPUT_PLAYBACK, 0, NULL, 0);
+	if (rate < 0) {
+		fprintf(stderr, "Unable to initialize espeak.\n");
+		return -1;
+	}
+
+	espeak_SetSynthCallback(callback);
+
+	/* Setup initial voice parameters */
+	if (defaultVoice && defaultVoice[0]) {
+		set_voice(s, defaultVoice);
+		free(defaultVoice);
+		defaultVoice = NULL;
+	}
+	set_frequency(s, defaultFrequency, ADJ_SET);
+	set_pitch(s, defaultPitch, ADJ_SET);
+	set_range(s, defaultRange, ADJ_SET);
+	set_rate(s, defaultRate, ADJ_SET);
+	set_volume(s, defaultVolume, ADJ_SET);
+	espeak_SetParameter(espeakCAPITALS, 0, 0);
+	paused_espeak = 0;
+	return 0;
+}
+
+/* espeak_thread is the "main" function of our secondary (queue-processing)
+ * thread.
+ * First, lock queue_guard, because it needs to be locked when we call
+ * pthread_cond_wait on the runner_awake condition variable.
+ * Next, enter an infinite loop.
+ * The wait call also unlocks queue_guard, so that the other thread can
+ * manipulate the queue.
+ * When runner_awake is signaled, the pthread_cond_wait call re-locks
+ * queue_guard, and the "queue processor" thread has access to the queue.
+ * While there is an entry in the queue, call queue_process_entry.
+ * queue_process_entry unlocks queue_guard after removing an item from the
+ * queue, so that the main thread doesn't have to wait for us to finish
+ * processing the entry.  So re-lock queue_guard after each call to
+ * queue_process_entry.
+ *
+ * The main thread can add items to the queue in exactly two situations:
+ * 1. We are waiting on runner_awake, or
+ * 2. We are processing an entry that has just been removed from the queue.
+ */
+void *espeak_thread(void *arg)
+{
+	struct synth_t *s = (struct synth_t *) arg;
+
+	pthread_mutex_lock(&queue_guard);
+	while (should_run) {
+		while (should_run && !queue_peek(synth_queue) && !stop_requested)
+			pthread_cond_wait(&runner_awake, &queue_guard);
+
+		if (stop_requested) {
+			current = NULL;
+			/* Call into espeak with queue_guard released: espeak_Cancel
+			 * can take time, or even block indefinitely when the audio
+			 * output is wedged, and holding the lock here would prevent
+			 * the other threads from ever making progress again.  The
+			 * queue cannot change concurrently: the only producer (the
+			 * softsynth thread) is blocked waiting for stop_acknowledged
+			 * as long as stop_requested is set. */
+			pthread_mutex_unlock(&queue_guard);
+			stop_speech();
+			pthread_mutex_lock(&queue_guard);
+			synth_queue_clear();
+			stop_requested = 0;
+			pthread_cond_signal(&stop_acknowledged);
+		}
+
+		while (should_run && queue_peek(synth_queue) && !stop_requested) {
+			queue_process_entry(s);
+		}
+	}
+	pthread_cond_signal(&stop_acknowledged);
+	pthread_mutex_unlock(&queue_guard);
+	return NULL;
+}

src/espeakup.c

@@ -1,5 +1,5 @@
 /*
- *  espeakup - interface which allows speakup to use espeak
+ *  espeakup - interface which allows speakup to use espeak-ng
  *
  *  Copyright (C) 2008 William Hubbs
  *
@@ -18,17 +18,18 @@
  */
 
 #include <errno.h>
+#include <fcntl.h>
 #include <signal.h>
 #include <stdio.h>
 #include <stdlib.h>
 #include <string.h>
-#include <unistd.h>
-#include <fcntl.h>
 #include <sys/file.h>
+#include <time.h>
+#include <unistd.h>
 
 #include "espeakup.h"
 
-/* path to our pid file */
+// path to our pid file
 char *pidPath = "/var/run/espeakup.pid";
 
 int debug = 0;
@@ -40,7 +41,9 @@ volatile int should_run = 1;
 espeak_AUDIO_OUTPUT audio_mode;
 
 pthread_cond_t runner_awake = PTHREAD_COND_INITIALIZER;
-pthread_cond_t stop_acknowledged = PTHREAD_COND_INITIALIZER;
+/* Initialized in main: use the monotonic clock for timed waits. */
+pthread_cond_t wake_stop;
+pthread_cond_t stop_acknowledged;
 pthread_mutex_t queue_guard = PTHREAD_MUTEX_INITIALIZER;
 
 int espeakup_start_daemon(void)
@@ -60,7 +63,7 @@ int espeakup_start_daemon(void)
 		exit(1);
 	}
 	if (pid) {
-		/* Parent, just wait for daemon */
+		// Parent, just wait for daemon
 		if (read(fds[0], &c, 1) < 0) {
 			printf("Espeakup is already running!\n");
 			exit(1);
@@ -68,17 +71,17 @@ int espeakup_start_daemon(void)
 		exit(c);
 	}
 
-	/* Child, create new session */
+	// Child, create new session
 	setsid();
 	pid = fork();
 	if (pid)
-		/* Intermediate child, just exit */
+		// Intermediate child, just exit
 		exit(0);
 
-	/* Child */
+	// Child
 	if (chdir("/") < 0) {
 		c = 1;
-		(void)write(fds[1], &c, 1);
+		(void) write(fds[1], &c, 1);
 		exit(1);
 	}
 	return fds[1];
@@ -99,20 +102,18 @@ int espeakup_is_running(void)
 	}
 
 	if (flock(pidFile, LOCK_EX) < 0) {
-		printf("Can not lock the pid file %s: %s\n", pidPath,
+		printf("Can not lock the pid file %s: %s\n", pidPath, strerror(errno));
-		       strerror(errno));
 		goto error;
 	}
 	n = read(pidFile, s, sizeof(s) - 1);
 	if (n < 0) {
-		printf("Can not read the pid file %s: %s\n", pidPath,
+		printf("Can not read the pid file %s: %s\n", pidPath, strerror(errno));
-		       strerror(errno));
 		goto error;
 	}
 	s[n] = 0;
 	n = sscanf(s, "%d", &pid);
 	if (n == 1 && (!kill(pid, 0) || errno != ESRCH)) {
-		/* Already running */
+		// Already running
 		close(pidFile);
 		return 1;
 	}
@@ -148,6 +149,18 @@ int main(int argc, char **argv)
 	struct synth_t s = {
 		.voice = "",
 	};
+	pthread_condattr_t monotonic_attr;
+
+	/* Condition variables used with pthread_cond_timedwait must use the
+	 * monotonic clock, so that wall-clock adjustments (NTP, an
+	 * installer setting the system time) cannot make the timeouts fire
+	 * too early or far too late. */
+	pthread_condattr_init(&monotonic_attr);
+	pthread_condattr_setclock(&monotonic_attr, CLOCK_MONOTONIC);
+	pthread_cond_init(&wake_stop, &monotonic_attr);
+	pthread_cond_init(&stop_acknowledged, &monotonic_attr);
+	pthread_condattr_destroy(&monotonic_attr);
+
 	synth_queue = new_queue();
 
 	if (!synth_queue) {
@@ -155,13 +168,13 @@ int main(int argc, char **argv)
 		return 2;
 	}
 
-	/* set up the pipe used to wake the espeak thread */
+	// set up the pipe used to wake the espeak thread
 	if (pipe(self_pipe_fds) < 0) {
 		perror("Unable to create pipe");
 		return 5;
 	}
 
-	/* process command line options */
+	// process command line options
 	process_cli(argc, argv);
 
 	if (!debug && espeakup_mode == ESPEAKUP_MODE_SPEAKUP) {
@@ -181,7 +194,7 @@ int main(int argc, char **argv)
 			close(devnull);
 	}
 
-	/* create the signal processing thread here. */
+	// create the signal processing thread here.
 	err = pthread_create(&signal_thread_id, NULL, signal_thread, NULL);
 	if (err != 0) {
 		ret = 4;
@@ -197,26 +210,26 @@ int main(int argc, char **argv)
 	sigaddset(&sigset, SIGTERM);
 	sigprocmask(SIG_BLOCK, &sigset, NULL);
 
-/* Initialize espeak */
+	// Initialize espeak
 	if (initialize_espeak(&s) < 0) {
 		ret = 2;
 		goto out;
 	}
 
-/* open the softsynth */
+	// open the softsynth
 	if (open_softsynth() < 0) {
 		ret = 2;
 		goto out;
 	}
 
-	/* Spawn our softsynth thread. */
+	// Spawn our softsynth thread.
 	err = pthread_create(&softsynth_thread_id, NULL, softsynth_thread, &s);
 	if (err != 0) {
 		ret = 4;
 		goto out;
 	}
 
-	/* Spawn our espeak-interacting thread. */
+	// Spawn our espeak-interacting thread.
 	err = pthread_create(&espeak_thread_id, NULL, espeak_thread, &s);
 	if (err != 0) {
 		ret = 4;
@@ -224,14 +237,15 @@ int main(int argc, char **argv)
 	}
 
 	if (!debug && espeakup_mode == ESPEAKUP_MODE_SPEAKUP)
-		(void)write(fd, &ret, 1);
+		(void) write(fd, &ret, 1);
 
-	/* wait for the threads to shut down. */
+	// wait for the threads to shut down.
 	pthread_join(signal_thread_id, NULL);
 	pthread_join(softsynth_thread_id, NULL);
 	pthread_join(espeak_thread_id, NULL);
 
-	espeak_Terminate();
+	if (!paused_espeak)
+		espeak_Terminate();
 	close_softsynth();
 
 out:
@@ -239,9 +253,8 @@ out:
 		if (ret != 1)
 			unlink(pidPath);
 		if (ret != 0)
-			(void)write(fd, &ret, 1);
+			(void) write(fd, &ret, 1);
-		/* If ret was 0, the status byte was written before joining
+		// If ret was 0, the status byte was written before joining the threads.
-		 * the threads. */
 	}
 	return ret;
 }

src/espeakup.h

@@ -1,5 +1,5 @@
 /*
- *  espeakup - interface which allows speakup to use espeak
+ *  espeakup - interface which allows speakup to use espeak-ng
  *
  *  Copyright (C) 2008 William Hubbs
  *
@@ -20,35 +20,40 @@
 #ifndef __ESPEAKUP_H
 #define __ESPEAKUP_H
 
-/* This was added for gcc 4.3 */
+// This was added for gcc 4.3
-#include <stddef.h>
 #include <pthread.h>
+#include <stddef.h>
 
-#include <espeak/speak_lib.h>
+#include <espeak-ng/speak_lib.h>
 
 #include "queue.h"
 
-#define PACKAGE_VERSION "0.80"
+#define PACKAGE_BUGREPORT "https://github.com/linux-speakup/espeakup/issues"
-#define PACKAGE_BUGREPORT "http://github.com/williamh/espeakup/issues"
 
-enum espeakup_mode_t {
+enum espeakup_mode_t
+{
 	ESPEAKUP_MODE_SPEAKUP,
 	ESPEAKUP_MODE_ACSINT
 };
 
-enum command_t {
+enum command_t
+{
 	CMD_SET_FREQUENCY,
+	CMD_SET_MARK,
 	CMD_SET_PITCH,
+	CMD_SET_RANGE,
 	CMD_SET_PUNCTUATION,
 	CMD_SET_RATE,
 	CMD_SET_VOICE,
 	CMD_SET_VOLUME,
 	CMD_SPEAK_TEXT,
 	CMD_FLUSH,
+	CMD_PAUSE,
 	CMD_UNKNOWN,
 };
 
-enum adjust_t {
+enum adjust_t
+{
 	ADJ_DEC,
 	ADJ_SET,
 	ADJ_INC,
@@ -65,9 +70,10 @@ struct espeak_entry_t {
 struct synth_t {
 	int frequency;
 	int pitch;
+	int range;
 	int punct;
 	int rate;
-	char voice[10];
+	char voice[20];
 	int volume;
 	char *buf;
 	int len;
@@ -84,13 +90,16 @@ extern void *espeak_thread(void *arg);
 extern int open_softsynth(void);
 extern void close_softsynth(void);
 extern void *softsynth_thread(void *arg);
+extern void softsynth_reportindex(int index);
 extern volatile int should_run;
 extern volatile int stop_requested;
+extern int paused_espeak;
 extern int self_pipe_fds[2];
 #define PIPE_READ_FD (self_pipe_fds[0])
 #define PIPE_WRITE_FD (self_pipe_fds[1])
 
 extern pthread_cond_t runner_awake;
+extern pthread_cond_t wake_stop;
 extern pthread_cond_t stop_acknowledged;
 extern pthread_mutex_t queue_guard;
 

src/meson.build

@@ -0,0 +1,10 @@
+espeakup_sources = files([
+        'cli.c',
+        'espeak.c',
+        'espeakup.c',
+        'queue.c',
+        'signal.c',
+        'softsynth.c',
+        'stringhandling.c'
+])
+espeakup_version = vcs_tag(input : 'version.h.in', output : 'version.h')

src/queue.c

@@ -1,5 +1,5 @@
 /*
- *  espeakup - interface which allows speakup to use espeak
+ *  espeakup - interface which allows speakup to use espeak-ng
  *
  * Note that these functions are meant to be used in either a single or
  * multi-threaded environment, so they know nothing about mutexes, etc.

src/queue.h

@@ -1,5 +1,5 @@
 /*
- *  espeakup - interface which allows speakup to use espeak
+ *  espeakup - interface which allows speakup to use espeak-ng
  *
  *  Copyright (C) 2008 William Hubbs
  *
@@ -20,7 +20,7 @@
 #ifndef __QUEUE_H
 #define __QUEUE_H
 
-struct queue_t;					/* An opaque type. */
+struct queue_t;     // An opaque type.
 
 extern struct queue_t *new_queue(void);
 extern int queue_add(struct queue_t *q, void *entry);

src/signal.c

@@ -1,5 +1,5 @@
 /*
- *  espeakup - interface which allows speakup to use espeak
+ *  espeakup - interface which allows speakup to use espeak-ng
  *
  *  Copyright (C) 2008 William Hubbs
  *
@@ -21,10 +21,11 @@
 #include <stdio.h>
 #include <string.h>
 #include <unistd.h>
-#define STOP_MSG "s"
 
 #include "espeakup.h"
 
+#define STOP_MSG "s"
+
 /*
  * We install a dummy signal handler to let the o/s know that we
  * do not want the default action to be performed since we are
@@ -40,8 +41,8 @@ void *signal_thread(void *arg)
 	sigset_t sigset;
 	int sig;
 
-	memset(&temp, 0, sizeof (struct sigaction));
+	memset(&temp, 0, sizeof(struct sigaction));
-	/* install dummy handlers for the signals we want to process */
+	// install dummy handlers for the signals we want to process
 	temp.sa_handler = dummy_handler;
 	sigemptyset(&temp.sa_mask);
 	sigaction(SIGINT, &temp, NULL);
@@ -57,14 +58,24 @@ void *signal_thread(void *arg)
 		case SIGTERM:
 			pthread_mutex_lock(&queue_guard);
 			should_run = 0;
+			/* Wake up any thread waiting on a condition variable so
+			 * that it notices the shutdown request: the softsynth
+			 * thread may be waiting for a stop acknowledgement, and
+			 * the espeak thread may be waiting for work or throttling
+			 * before a retry. */
+			pthread_cond_broadcast(&runner_awake);
+			pthread_cond_broadcast(&wake_stop);
+			pthread_cond_broadcast(&stop_acknowledged);
 			pthread_mutex_unlock(&queue_guard);
 			break;
 		default:
 			printf("espeakup caught signal %d\n", sig);
 			break;
 		}
+		pthread_mutex_lock(&queue_guard);
 	}
-	/* Tell the reader to stop, if it is in a select() call. */
+	pthread_mutex_unlock(&queue_guard);
+	// Tell the reader to stop, if it is in a select() call.
 	write(PIPE_WRITE_FD, STOP_MSG, strlen(STOP_MSG));
 	return NULL;
 }

src/softsynth.c

@@ -1,5 +1,5 @@
 /*
- *  espeakup - interface which allows speakup to use espeak
+ *  espeakup - interface which allows speakup to use espeak-ng
  *
  *  Copyright (C) 2008 William Hubbs
  *
@@ -17,27 +17,28 @@
  *   along with this program.  If not, see <http://www.gnu.org/licenses/>.
  */
 
+#include <ctype.h>
 #include <errno.h>
 #include <fcntl.h>
 #include <stdio.h>
 #include <stdlib.h>
-#include <unistd.h>
-#include <sys/select.h>
 #include <string.h>
-#include <ctype.h>
+#include <sys/select.h>
+#include <time.h>
+#include <unistd.h>
 
 #include "espeakup.h"
 #include "stringhandling.h"
 
-/* max buffer size */
+// max buffer size
 static const size_t maxBufferSize = 16 * 1024 + 1;
 
-/* synth flush character */
+// synth flush character
 static const int synthFlushChar = 0x18;
 
 static int softFD = 0;
 
-/* Text accumulator: */
+// Text accumulator:
 char *textAccumulator;
 int textAccumulator_l;
 
@@ -124,15 +125,24 @@ static int process_command(struct synth_t *s, char *buf, int start)
 		case 'f':
 			cmd = CMD_SET_FREQUENCY;
 			break;
+		case 'i':
+			cmd = CMD_SET_MARK;
+			break;
 		case 'p':
 			cmd = CMD_SET_PITCH;
 			break;
+		case 'r':
+			cmd = CMD_SET_RANGE;
+			break;
 		case 's':
 			cmd = CMD_SET_RATE;
 			break;
 		case 'v':
 			cmd = CMD_SET_VOLUME;
 			break;
+		case 'P':
+			cmd = CMD_PAUSE;
+			break;
 		default:
 			cmd = CMD_UNKNOWN;
 			break;
@@ -146,8 +156,7 @@ static int process_command(struct synth_t *s, char *buf, int start)
 	}
 
 	if (cmd != CMD_FLUSH && cmd != CMD_UNKNOWN) {
-		if (espeakup_mode == ESPEAKUP_MODE_ACSINT
+		if (espeakup_mode == ESPEAKUP_MODE_ACSINT && textAccumulator_l != 0) {
-			&& textAccumulator_l != 0) {
 			queue_add_text(textAccumulator, textAccumulator_l);
 			free(textAccumulator);
 			textAccumulator = initString(&textAccumulator_l);
@@ -168,7 +177,8 @@ static void process_buffer(struct synth_t *s, char *buf, ssize_t length)
 	start = 0;
 	end = 0;
 	while (start < length) {
-		while ((buf[end] < 0 || buf[end] >= ' ') && end < length)
+		while ((buf[end] < 0 || buf[end] >= ' ' || buf[end] == '\n') &&
+		       end < length)
 			end++;
 		if (end != start) {
 			txtLen = end - start;
@@ -183,8 +193,7 @@ static void process_buffer(struct synth_t *s, char *buf, ssize_t length)
 	}
 }
 
-static void process_buffer_acsint(struct synth_t *s, char *buf,
+static void process_buffer_acsint(struct synth_t *s, char *buf, ssize_t length)
-								  ssize_t length)
 {
 	int start = 0;
 	int i;
@@ -198,8 +207,8 @@ static void process_buffer_acsint(struct synth_t *s, char *buf,
 				break;
 		}
 		if (i > start)
-			stringAndBytes(&textAccumulator, &textAccumulator_l,
+			stringAndBytes(&textAccumulator, &textAccumulator_l, buf + start,
-						   buf + start, i - start);
+			               i - start);
 		if (flushIt) {
 			if (textAccumulator != EMPTYSTRING) {
 				queue_add_text(textAccumulator, textAccumulator_l);
@@ -215,29 +224,56 @@ static void process_buffer_acsint(struct synth_t *s, char *buf,
 	}
 }
 
+/* How long to wait for the espeak thread to acknowledge a stop request
+ * before concluding that espeak is wedged beyond in-process recovery. */
+static const int stopAckTimeout = 10;
+
 static void request_espeak_stop(void)
 {
+	struct timespec timeout;
+	int err = 0;
+
 	pthread_mutex_lock(&queue_guard);
 	stop_requested = 1;
-	pthread_cond_signal(&runner_awake);	/* Wake runner, if necessary. */
+	pthread_cond_signal(&runner_awake);     // Wake runner, if necessary.
-	while (should_run && stop_requested)
+	pthread_cond_signal(&wake_stop);        // Wake runner, if necessary.
-		pthread_cond_wait(&stop_acknowledged, &queue_guard);	/* wait for acknowledgement. */
+	clock_gettime(CLOCK_MONOTONIC, &timeout);
+	timeout.tv_sec += stopAckTimeout;
+	while (should_run && stop_requested && err != ETIMEDOUT)
+		// wait for acknowledgement.
+		err = pthread_cond_timedwait(&stop_acknowledged, &queue_guard,
+		                             &timeout);
+	if (should_run && stop_requested) {
+		/* The espeak thread is stuck in a call into espeak, most likely
+		 * on a wedged audio device.  There is no way to recover from
+		 * within the process: exit so that the init system can respawn
+		 * us in a clean state, rather than staying silent, ignoring
+		 * SIGTERM, and stalling the whole console by not draining
+		 * /dev/softsynth anymore.  Use _exit because exit could hang in
+		 * library destructors while the audio device is wedged. */
+		fprintf(stderr, "espeakup: espeak did not acknowledge a stop "
+		        "request within %d seconds, aborting\n", stopAckTimeout);
+		_exit(3);
+	}
 	pthread_mutex_unlock(&queue_guard);
 }
 
 int open_softsynth(void)
 {
 	int rc = 0;
-	/* If we're in acsint mode, we read from stdin.  No need to open. */
+	// If we're in acsint mode, we read from stdin.  No need to open.
 	if (espeakup_mode == ESPEAKUP_MODE_ACSINT) {
 		softFD = STDIN_FILENO;
 		return 0;
 	}
 
-	/* open the softsynth. */
+	// open the softsynth.
-	softFD = open("/dev/softsynth", O_RDWR | O_NONBLOCK);
+	softFD = open("/dev/softsynthu", O_RDWR | O_NONBLOCK);
+	if (softFD < 0 && errno == ENOENT)
+		// Kernel without unicode support? Try without unicode.
+		softFD = open("/dev/softsynth", O_RDWR | O_NONBLOCK);
 	if (softFD < 0) {
-		perror("Unable to open the softsynth device");
+		perror("Unable to open the /dev/softsynth device");
 		rc = -1;
 	}
 	return rc;
@@ -319,3 +355,16 @@ void *softsynth_thread(void *arg)
 	pthread_mutex_unlock(&queue_guard);
 	return NULL;
 }
+
+void softsynth_reportindex(int index)
+{
+	if (espeakup_mode == ESPEAKUP_MODE_ACSINT) {
+		putchar(index);
+		fflush(stdout);
+	} else {
+		char buf[16];
+		snprintf(buf, sizeof(buf), "%d", index);
+		if (write(softFD, buf, strlen(buf)) < 0)
+			perror("Writing index failed");
+	}
+}

src/stringhandling.c

@@ -1,5 +1,5 @@
 /*
- *  espeakup - interface which allows speakup to use espeak
+ *  espeakup - interface which allows speakup to use espeak-ng
  *
  *  Copyright (C) 2011 William Hubbs
  *
@@ -62,6 +62,17 @@ void *reallocMem(void *p, size_t n)
 	return s;
 }
 
+char *dupeString(char *s)
+{
+	char *c;
+
+	if (!(c = strdup(s))) {
+		fprintf(stderr, "Out of memory!\n");
+		exit(1);
+	}
+	return c;
+}
+
 char *initString(int *l)
 {
 	*l = 0;
@@ -75,9 +86,9 @@ void stringAndString(char **s, int *l, const char *t)
 	oldlen = *l;
 	newlen = oldlen + strlen(t);
 	*l = newlen;
-	++newlen;					/* room for the 0 */
+	++newlen;     // room for the 0
 	x = oldlen ^ newlen;
-	if (x > oldlen) {			/* must realloc */
+	if (x > oldlen) {     // must realloc
 		newlen |= (newlen >> 1);
 		newlen |= (newlen >> 2);
 		newlen |= (newlen >> 4);
@@ -98,7 +109,7 @@ void stringAndBytes(char **s, int *l, const char *t, int cnt)
 	*l = newlen;
 	++newlen;
 	x = oldlen ^ newlen;
-	if (x > oldlen) {			/* must realloc */
+	if (x > oldlen) {     // must realloc
 		newlen |= (newlen >> 1);
 		newlen |= (newlen >> 2);
 		newlen |= (newlen >> 4);

src/stringhandling.h

@@ -1,5 +1,5 @@
 /*
- *  espeakup - interface which allows speakup to use espeak
+ *  espeakup - interface which allows speakup to use espeak-ng
  *
  *  Copyright (C) 2011 William Hubbs
  *
@@ -26,6 +26,7 @@ extern char *EMPTYSTRING;
 
 void *allocMem(size_t n);
 void *reallocMem(void *p, size_t n);
+char *dupeString(char *s);
 char *initString(int *l);
 void stringAndString(char **s, int *l, const char *t);
 void stringAndBytes(char **s, int *l, const char *t, int cnt);

src/version.h.in

@@ -0,0 +1,25 @@
+/*
+ *  espeakup - interface which allows speakup to use espeak-ng
+ *
+ *  Copyright (C) 2008 William Hubbs
+ *
+ *  This program is free software: you can redistribute it and/or modify
+ *  it under the terms of the GNU General Public License as published by
+ *  the Free Software Foundation, either version 3 of the License, or
+ *  (at your option) any later version.
+ *
+ *   This program is distributed in the hope that it will be useful,
+ *   but WITHOUT ANY WARRANTY; without even the implied warranty of
+ *   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ *   GNU General Public License for more details.
+ *
+ *  You should have received a copy of the GNU General Public License
+ *   along with this program.  If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#ifndef __VERSION_H
+#define __VERSION_H
+
+#define PACKAGE_VERSION "@VCS_TAG@"
+
+#endif