From 4d090a2ac5684658fde5fc8c5b3eda66d775a691 Mon Sep 17 00:00:00 2001
From: Kipp Cannon <kipp.cannon@ligo.org>
Date: Tue, 12 Feb 2019 12:16:02 +0900
Subject: [PATCH] gstlal: add "ezligolw" library
- due to push back from collaboration over our desire to switch ilwd:char
IDs in LIGO LW tables to integers, we are forced to decouple ourselves
from the I/O code in LAL.
---
gstlal/doc/gobject/Makefile.am | 2 +
gstlal/lib/gstlal/Makefile.am | 8 +-
gstlal/lib/gstlal/ezligolw.c | 548 +++++++++++++++++
gstlal/lib/gstlal/ezligolw.h | 87 +++
gstlal/lib/gstlal/ezxml.c | 1015 ++++++++++++++++++++++++++++++++
gstlal/lib/gstlal/ezxml.h | 167 ++++++
6 files changed, 1825 insertions(+), 2 deletions(-)
create mode 100644 gstlal/lib/gstlal/ezligolw.c
create mode 100644 gstlal/lib/gstlal/ezligolw.h
create mode 100644 gstlal/lib/gstlal/ezxml.c
create mode 100644 gstlal/lib/gstlal/ezxml.h
diff --git a/gstlal/doc/gobject/Makefile.am b/gstlal/doc/gobject/Makefile.am
index 5739a68510..c1a694dd72 100644
--- a/gstlal/doc/gobject/Makefile.am
+++ b/gstlal/doc/gobject/Makefile.am
@@ -72,6 +72,8 @@ EXTRA_HFILES=
# Header files or dirs to ignore when scanning. Use base file/dir names
# e.g. IGNORE_HFILES=gtkdebug.h gtkintl.h private_code
IGNORE_HFILES= \
+ ezxml.c ezxml.h \
+ ezligolw.c ezligolw.h \
gstlal_peakfinder_top.h \
gstlal_peakfinder_types.h
diff --git a/gstlal/lib/gstlal/Makefile.am b/gstlal/lib/gstlal/Makefile.am
index c13e316900..c156d54f98 100644
--- a/gstlal/lib/gstlal/Makefile.am
+++ b/gstlal/lib/gstlal/Makefile.am
@@ -1,9 +1,13 @@
EXTRA_DIST =
CLEANFILES =
-pkginclude_HEADERS = gstlal.h gstlal_debug.h gstlal_marshal.h gstlal_tags.h gstlal_cdf_weighted_chisq_P.h gstlal_segments.h gstaudioadapter.h gstlalcollectpads.h gstlal_peakfinder.h gstlal_autocorrelation_chi2.h gstlal_gps_clock.h gstlal_frhistory.h gstlal_audio_info.h
+pkginclude_HEADERS = ezligolw.h ezxml.h gstlal.h gstlal_debug.h gstlal_marshal.h gstlal_tags.h gstlal_cdf_weighted_chisq_P.h gstlal_segments.h gstaudioadapter.h gstlalcollectpads.h gstlal_peakfinder.h gstlal_autocorrelation_chi2.h gstlal_gps_clock.h gstlal_frhistory.h gstlal_audio_info.h
pkgconfig_DATA = gstlal.pc
-lib_LTLIBRARIES = libgstlal.la libgstlaltags.la libgstlaltypes.la
+lib_LTLIBRARIES = libezligolw.la libgstlal.la libgstlaltags.la libgstlaltypes.la
+
+libezligolw_la_SOURCES = ezligolw.c ezligolw.h ezxml.c ezxml.h
+libezligolw_la_CFLAGS = $(AM_CFLAGS) -I..
+libezligolw_la_LDFLAGS = -version-info $(LIBVERSION) $(AM_LDFLAGS)
libgstlal_la_SOURCES = gstlal.c gstlal.h gstlal_debug.h gstlal_fftw.c gstlal_marshal.c gstlal_marshal.h gstlal_cdf_weighted_chisq_P.c gstlal_cdf_weighted_chisq_P.h gstlal_segments.h gstlal_segments.c gstlal_peakfinder.h gstlal_peakfinder.c gstlal_autocorrelation_chi2.h gstlal_autocorrelation_chi2.c gstlal_audio_info.c
libgstlal_la_CFLAGS = $(AM_CFLAGS) $(FFTW_CFLAGS) $(GSL_CFLAGS) $(LAL_CFLAGS) $(gstreamer_CFLAGS)
diff --git a/gstlal/lib/gstlal/ezligolw.c b/gstlal/lib/gstlal/ezligolw.c
new file mode 100644
index 0000000000..413f71abb6
--- /dev/null
+++ b/gstlal/lib/gstlal/ezligolw.c
@@ -0,0 +1,548 @@
+/*
+ * $Id: ezligolw.c,v 1.4 2008/07/31 08:28:42 kipp Exp $
+ *
+ * Copyright (C) 2007 Kipp Cannon
+ *
+ * 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 2 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, write to the Free Software Foundation, Inc.,
+ * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
+ */
+
+#include <ctype.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <ezligolw.h>
+#include <ezxml.h>
+
+
+/*
+ * Extract the meaningful portion of a table name. Returns a pointer to
+ * the last colon-delimited substring before an optional ":table" suffix.
+ */
+
+
+static const char *ligolw_strip_table_name(const char *Name)
+{
+ char buff[strlen(Name) + 1];
+ char *pos = buff;
+ char *start;
+
+ strcpy(buff, Name);
+
+ do
+ start = strsep(&pos, ":");
+ while(pos && strncmp(pos, "table", 5));
+
+ return Name + (start - buff);
+}
+
+
+/*
+ * Extract the meaningful portion of a column name. Returns a pointer to
+ * the last colon-delimited substring.
+ */
+
+
+static const char *ligolw_strip_column_name(const char *Name)
+{
+ char buff[strlen(Name) + 1];
+ char *pos = buff;
+ char *start;
+
+ strcpy(buff, Name);
+
+ do
+ start = strsep(&pos, ":");
+ while(pos);
+
+ return Name + (start - buff);
+}
+
+
+/*
+ * Convert a LIGO Light Weight type name string to/from a numeric type
+ * index.
+ */
+
+
+static const struct name_to_enum {
+ const char *name;
+ enum ligolw_table_cell_type type;
+} name_to_enum[] = {
+ {"char_s", ligolw_cell_type_char_s},
+ {"char_v", ligolw_cell_type_char_v},
+ {"ilwd:char", ligolw_cell_type_ilwdchar},
+ {"ilwd:char_u", ligolw_cell_type_ilwdchar_u},
+ {"lstring", ligolw_cell_type_lstring},
+ {"string", ligolw_cell_type_lstring},
+ {"int_2s", ligolw_cell_type_int_2s},
+ {"int_2u", ligolw_cell_type_int_2u},
+ {"int_4s", ligolw_cell_type_int_4s},
+ {"int", ligolw_cell_type_int_4s},
+ {"int_4u", ligolw_cell_type_int_4u},
+ {"int_8s", ligolw_cell_type_int_8s},
+ {"int_8u", ligolw_cell_type_int_8u},
+ {"real_4", ligolw_cell_type_real_4},
+ {"float", ligolw_cell_type_real_4},
+ {"real_8", ligolw_cell_type_real_8},
+ {"double", ligolw_cell_type_real_8},
+ {NULL, -1}
+};
+
+
+enum ligolw_table_cell_type ligolw_table_type_name_to_enum(const char *name)
+{
+ const struct name_to_enum *n_to_e;
+
+ for(n_to_e = name_to_enum; n_to_e->name; n_to_e++)
+ if(!strcmp(n_to_e->name, name))
+ /* found it */
+ return n_to_e->type;
+
+ /* unrecognized type */
+ return -1;
+}
+
+
+const char *ligolw_table_type_enum_to_name(enum ligolw_table_cell_type t)
+{
+ const struct name_to_enum *n_to_e;
+
+ for(n_to_e = name_to_enum; n_to_e->name; n_to_e++)
+ if(n_to_e->type == t)
+ /* found it */
+ return n_to_e->name;
+
+ /* unrecognized type */
+ return NULL;
+}
+
+
+/*
+ * Default row builder call-back.
+ */
+
+
+int ligolw_table_default_row_callback(struct ligolw_table *table, struct ligolw_table_row row, void *ignored)
+{
+ table->rows = realloc(table->rows, (table->n_rows + 1) * sizeof(*table->rows));
+ table->rows[table->n_rows] = row;
+ table->n_rows++;
+ return 0;
+}
+
+
+/*
+ * Parse an ezxml_t Table element into a struct ligolw_table structure. If
+ * row_callback() is NULL, then the default row builder is used, which
+ * inserts the rows into the ligolw_table structure. Calling code can
+ * provide it's own function, which will be called after each row is
+ * constructed. This allows the rows to be "intercepted", so that some
+ * other thing can be done with them other than being inserted into the
+ * ligolw_table. The call-back function will be passed the pointer to the
+ * current ligolw_table structure as its first argument, the pointer to the
+ * new row as its second, and the callback_data pointer as its third
+ * argument. The row_callback() function must free the row's cells element
+ * if it will not be saving it, or memory will be leaked. The call-back
+ * returns 0 to indicate success, non-zero to indicate failure.
+ *
+ * ligolw_table_parse() return the pointer to the new struct ligolw_table
+ * structure on success, NULL on failure.
+ */
+
+
+static void next_token(char **start, char **end, char **end_end, char delimiter)
+{
+ char *s, *e, *ee;
+ int quoted = 0;
+
+ /* find the token's start */
+ for(s = *start; *s && isspace(*s) && *s != delimiter && !(quoted = *s == '"'); s++);
+ if(!*s || *s == delimiter) {
+ /* token has zero length */
+ *start = *end = *end_end = s;
+ return;
+ }
+ /* find the end end of the token */
+ for(ee = s + 1; *ee && (quoted || *ee != delimiter); ee++)
+ if(quoted && *ee == '"')
+ quoted = 0;
+ /* move backwards from end end to find end */
+ for(e = ee - 1; e > s && isspace(*e); e--);
+ /* adjust */
+ if(*s == '"')
+ s++;
+ if(*e != '"')
+ e++;
+ /* answer */
+ *start = s;
+ *end = e;
+ *end_end = ee;
+}
+
+
+struct ligolw_table *ligolw_table_parse(ezxml_t elem, int (row_callback)(struct ligolw_table *, struct ligolw_table_row, void *), void *callback_data)
+{
+ struct ligolw_table *table;
+ char *txt;
+ ezxml_t column;
+ ezxml_t stream;
+
+ table = malloc(sizeof(*table));
+ if(!table)
+ return NULL;
+
+ table->name = ligolw_strip_table_name(ezxml_attr(elem, "Name"));
+
+ table->n_columns = 0;
+ table->columns = NULL;
+ table->n_rows = 0;
+ table->rows = NULL;
+
+ for(column = ezxml_child(elem, "Column"); column; column = column->next) {
+ table->columns = realloc(table->columns, (table->n_columns + 1) * sizeof(*table->columns));
+
+ table->columns[table->n_columns].name = ligolw_strip_column_name(ezxml_attr(column, "Name"));
+ table->columns[table->n_columns].table = table;
+ table->columns[table->n_columns].type = ligolw_table_type_name_to_enum(ezxml_attr(column, "Type"));
+
+ table->n_columns++;
+ }
+
+ stream = ezxml_child(elem, "Stream");
+ if(!stream) {
+ /* DTD allows Table to have 0 Stream children */
+ table->delimiter = '\0';
+ return table;
+ }
+
+ table->delimiter = *ezxml_attr(stream, "Delimiter");
+
+ if(!row_callback)
+ row_callback = ligolw_table_default_row_callback;
+
+ for(txt = stream->txt; txt && *txt; ) {
+ struct ligolw_table_row row;
+ int c;
+
+ row.table = table;
+ row.cells = malloc(table->n_columns * sizeof(*row.cells));
+
+ for(c = 0; c < table->n_columns; c++) {
+ char *end, *end_end;
+
+ next_token(&txt, &end, &end_end, table->delimiter);
+
+ switch(table->columns[c].type) {
+ case ligolw_cell_type_char_s:
+ case ligolw_cell_type_char_v:
+ case ligolw_cell_type_ilwdchar:
+ case ligolw_cell_type_ilwdchar_u:
+ case ligolw_cell_type_blob:
+ case ligolw_cell_type_lstring:
+ /* FIXME: move into a separate buffer so
+ * that the original document is not
+ * modified (see null terminator below) */
+ /* FIXME: binary types need to be sent
+ * through a decoder following this */
+ row.cells[c].as_string = txt;
+ break;
+
+ case ligolw_cell_type_int_2s:
+ case ligolw_cell_type_int_4s:
+ case ligolw_cell_type_int_8s:
+ row.cells[c].as_int = strtoll(txt, NULL, 0);
+ break;
+
+ case ligolw_cell_type_int_2u:
+ case ligolw_cell_type_int_4u:
+ case ligolw_cell_type_int_8u:
+ row.cells[c].as_uint = strtoull(txt, NULL, 0);
+ break;
+
+ case ligolw_cell_type_real_4:
+ row.cells[c].as_float = strtod(txt, NULL);
+ break;
+
+ case ligolw_cell_type_real_8:
+ row.cells[c].as_double = strtod(txt, NULL);
+ break;
+ }
+
+ /* don't set null at end until after testing
+ * end_end incase they point to the same place */
+ if(!*end_end) {
+ *end = '\0';
+ txt = end_end;
+ break;
+ }
+ *end = '\0';
+ txt = end_end + 1;
+ }
+
+ if(row_callback(table, row, callback_data)) {
+ ligolw_table_free(table);
+ return NULL;
+ }
+ }
+
+ return table;
+}
+
+
+/*
+ * Free a struct ligolw_table.
+ */
+
+
+void ligolw_table_free(struct ligolw_table *table)
+{
+
+ if(table) {
+ int i;
+ for(i = 0; i < table->n_rows; i++)
+ free(table->rows[i].cells);
+ free(table->rows);
+ free(table->columns);
+ }
+ free(table);
+}
+
+
+/*
+ * Get a column index by name from within a table. Returns the index of
+ * the column within table's columns array (and thus of the corresponding
+ * cell within each row's cell array) or -1 on failure. If type is not
+ * NULL, the place it points to is set to the columns's table_cell_type.
+ */
+
+
+int ligolw_table_get_column(struct ligolw_table *table, const char *name, enum ligolw_table_cell_type *type)
+{
+ int i;
+
+ for(i = 0; i < table->n_columns; i++)
+ if(!strcmp(table->columns[i].name, name)) {
+ /* found it */
+ if(type)
+ *type = table->columns[i].type;
+ return i;
+ }
+
+ /* couldn't find that column name */
+ if(type)
+ *type = -1;
+ return -1;
+}
+
+
+/*
+ * Retrieve the value stored in a cell within a table row. No error
+ * checking is done, you should ensure the requested column is present
+ * before calling this function.
+ */
+
+
+union ligolw_table_cell ligolw_row_get_cell(struct ligolw_table_row row, const char *name)
+{
+ return row.cells[ligolw_table_get_column(row.table, name, NULL)];
+}
+
+
+/*
+ * Find an ezxml_t Table element in a document.
+ */
+
+
+ezxml_t ligolw_table_get(ezxml_t xmldoc, const char *table_name)
+{
+ int n = strlen(table_name);
+ ezxml_t table;
+
+ for(table = ezxml_child(xmldoc, "Table"); table; table = table->next)
+ if(!strncmp(ligolw_strip_table_name(ezxml_attr(table, "Name")), table_name, n))
+ break;
+
+ return table;
+}
+
+
+/*
+ * Generic unpacking row builder.
+ */
+
+
+int ligolw_unpacking_row_builder(struct ligolw_table *table, struct ligolw_table_row row, void *data)
+{
+ struct ligolw_unpacking_spec *spec;
+
+ for(spec = data; spec->name; spec++) {
+ int c;
+ enum ligolw_table_cell_type type;
+ if((c = ligolw_table_get_column(table, spec->name, &type)) < 0) {
+ /* no column by that name */
+ if(!(spec->flags & LIGOLW_UNPACKING_REQUIRED))
+ /* not required */
+ continue;
+ free(row.cells);
+ return spec - (struct ligolw_unpacking_spec *) data + 1;
+ }
+ if(spec->type != type) {
+ /* type mismatch */
+ free(row.cells);
+ return -(spec - (struct ligolw_unpacking_spec *) data + 1);
+ }
+ if(!spec->dest)
+ /* column has a valid name and the correct type,
+ * but is ignored */
+ continue;
+
+ switch(spec->type) {
+ case ligolw_cell_type_char_s:
+ case ligolw_cell_type_char_v:
+ case ligolw_cell_type_ilwdchar:
+ case ligolw_cell_type_ilwdchar_u:
+ case ligolw_cell_type_lstring:
+ *(const char **) spec->dest = row.cells[c].as_string;
+ break;
+
+ case ligolw_cell_type_blob:
+ *(const unsigned char **) spec->dest = row.cells[c].as_blob;
+ break;
+
+ case ligolw_cell_type_int_2s:
+ *(int16_t *) spec->dest = row.cells[c].as_int;
+ break;
+
+ case ligolw_cell_type_int_2u:
+ *(int16_t *) spec->dest = row.cells[c].as_uint;
+ break;
+
+ case ligolw_cell_type_int_4s:
+ *(int32_t *) spec->dest = row.cells[c].as_int;
+ break;
+
+ case ligolw_cell_type_int_4u:
+ *(uint32_t *) spec->dest = row.cells[c].as_uint;
+ break;
+
+ case ligolw_cell_type_int_8s:
+ *(int64_t *) spec->dest = row.cells[c].as_int;
+ break;
+
+ case ligolw_cell_type_int_8u:
+ *(uint64_t *) spec->dest = row.cells[c].as_uint;
+ break;
+
+ case ligolw_cell_type_real_4:
+ *(float *) spec->dest = row.cells[c].as_float;
+ break;
+
+ case ligolw_cell_type_real_8:
+ *(double *) spec->dest = row.cells[c].as_double;
+ break;
+ }
+ }
+
+ free(row.cells);
+
+ if(spec - (struct ligolw_unpacking_spec *) data != table->n_columns) {
+ /* table has more columns than allowed */
+ /* FIXME: if this is an error, return an error code */
+ /* FIXME: this test doesn't work if the same column gets
+ * unpacked into more than one location, which the design
+ * of the loop above would allow */
+ }
+
+ return 0;
+}
+
+
+/*
+ * Print a struct ligolw_table structure
+ */
+
+
+int ligolw_table_print(FILE *f, struct ligolw_table *table)
+{
+ char short_name[strlen(table->name) + 1];
+ int r, c;
+
+ /* create a version of the table name with the optional :table
+ * suffix removed */
+ strcpy(short_name, table->name);
+ {
+ char *x = strchr(short_name, ':');
+ if(x)
+ *x = '\0';
+ }
+
+ /* print the table metadata */
+ fprintf(f, "<Table Name=\"%s\">\n", table->name);
+ for(c = 0; c < table->n_columns; c++)
+ fprintf(f, "\t<Column Name=\"%s:%s\" Type=\"%s\"/>\n", short_name, table->columns[c].name, ligolw_table_type_enum_to_name(table->columns[c].type));
+ fprintf(f, "\t<Stream Name=\"%s\" Type=\"Local\" Delimiter=\"%c\">\n", table->name, table->delimiter);
+
+ /* print the rows */
+ for(r = 0; r < table->n_rows; r++) {
+ if(r)
+ fprintf(f, ",\n\t\t");
+ else
+ fprintf(f, "\t\t");
+ for(c = 0; c < table->n_columns; c++) {
+ if(c)
+ fprintf(f, "%c", table->delimiter);
+
+ switch(table->columns[c].type) {
+ case ligolw_cell_type_char_s:
+ case ligolw_cell_type_char_v:
+ case ligolw_cell_type_ilwdchar:
+ case ligolw_cell_type_ilwdchar_u:
+ case ligolw_cell_type_blob:
+ case ligolw_cell_type_lstring:
+ /* FIXME: binary types need to pass through
+ * encoders first */
+ /* FIXME: string types need to pass through
+ * encoders first */
+ fprintf(f, "\"%s\"", table->rows[r].cells[c].as_string);
+ break;
+
+ case ligolw_cell_type_int_2s:
+ case ligolw_cell_type_int_4s:
+ case ligolw_cell_type_int_8s:
+ fprintf(f, "%lld", (long long) table->rows[r].cells[c].as_int);
+ break;
+
+ case ligolw_cell_type_int_2u:
+ case ligolw_cell_type_int_4u:
+ case ligolw_cell_type_int_8u:
+ fprintf(f, "%llu", (unsigned long long) table->rows[r].cells[c].as_uint);
+ break;
+
+ case ligolw_cell_type_real_4:
+ fprintf(f, "%.7g", (double) table->rows[r].cells[c].as_float);
+ break;
+
+ case ligolw_cell_type_real_8:
+ fprintf(f, "%.16g", table->rows[r].cells[c].as_double);
+ break;
+ }
+ }
+ }
+
+ /* finish 'er off */
+ fprintf(f, "\n\t</Stream>\n</Table>\n");
+
+ return 0;
+}
diff --git a/gstlal/lib/gstlal/ezligolw.h b/gstlal/lib/gstlal/ezligolw.h
new file mode 100644
index 0000000000..6a1d0c74f4
--- /dev/null
+++ b/gstlal/lib/gstlal/ezligolw.h
@@ -0,0 +1,87 @@
+/*
+ * $Id: ezligolw.h,v 1.4 2008/07/31 08:28:42 kipp Exp $
+ *
+ * Copyright (C) 2007 Kipp Cannon
+ *
+ * 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 2 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, write to the Free Software Foundation, Inc.,
+ * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
+ */
+
+
+#include <stdio.h>
+#include <stdint.h>
+#include <gstlal/ezxml.h>
+
+
+struct ligolw_table {
+ const char *name;
+ char delimiter;
+ int n_columns;
+ struct ligolw_table_column {
+ const char *name;
+ struct ligolw_table *table;
+ enum ligolw_table_cell_type {
+ ligolw_cell_type_char_s,
+ ligolw_cell_type_char_v,
+ ligolw_cell_type_ilwdchar,
+ ligolw_cell_type_ilwdchar_u,
+ ligolw_cell_type_blob,
+ ligolw_cell_type_lstring,
+ ligolw_cell_type_int_2s,
+ ligolw_cell_type_int_2u,
+ ligolw_cell_type_int_4s,
+ ligolw_cell_type_int_4u,
+ ligolw_cell_type_int_8s,
+ ligolw_cell_type_int_8u,
+ ligolw_cell_type_real_4,
+ ligolw_cell_type_real_8
+ } type;
+ } *columns;
+ int n_rows;
+ struct ligolw_table_row {
+ struct ligolw_table *table;
+ union ligolw_table_cell {
+ int64_t as_int;
+ uint64_t as_uint;
+ float as_float;
+ double as_double;
+ const char *as_string;
+ const unsigned char *as_blob;
+ } *cells;
+ } *rows;
+};
+
+
+ezxml_t ligolw_table_get(ezxml_t, const char *);
+enum ligolw_table_cell_type ligolw_table_type_name_to_enum(const char *);
+const char *ligolw_table_type_enum_to_name(enum ligolw_table_cell_type);
+int ligolw_table_default_row_callback(struct ligolw_table *, struct ligolw_table_row, void *);
+struct ligolw_table *ligolw_table_parse(ezxml_t, int (*)(struct ligolw_table *, struct ligolw_table_row, void *), void *);
+union ligolw_table_cell ligolw_row_get_cell(struct ligolw_table_row, const char *);
+void ligolw_table_free(struct ligolw_table *);
+int ligolw_table_get_column(struct ligolw_table *, const char *, enum ligolw_table_cell_type *);
+int ligolw_table_print(FILE *, struct ligolw_table *);
+
+
+#define LIGOLW_UNPACKING_REQUIRED 0x1
+
+
+struct ligolw_unpacking_spec {
+ void *dest;
+ const char *name;
+ enum ligolw_table_cell_type type;
+ int flags;
+};
+
+int ligolw_unpacking_row_builder(struct ligolw_table *, struct ligolw_table_row, void *);
diff --git a/gstlal/lib/gstlal/ezxml.c b/gstlal/lib/gstlal/ezxml.c
new file mode 100644
index 0000000000..e964b40db6
--- /dev/null
+++ b/gstlal/lib/gstlal/ezxml.c
@@ -0,0 +1,1015 @@
+/* ezxml.c
+ *
+ * Copyright 2004-2006 Aaron Voisine <aaron@voisine.org>
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining
+ * a copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sublicense, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included
+ * in all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
+ * IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY
+ * CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
+ * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
+ * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ */
+
+#include <stdlib.h>
+#include <stdio.h>
+#include <stdarg.h>
+#include <string.h>
+#include <ctype.h>
+#include <unistd.h>
+#include <sys/types.h>
+#ifndef EZXML_NOMMAP
+#include <sys/mman.h>
+#endif // EZXML_NOMMAP
+#include <sys/stat.h>
+#include "ezxml.h"
+
+#define EZXML_WS "\t\r\n " // whitespace
+#define EZXML_ERRL 128 // maximum error string length
+
+typedef struct ezxml_root *ezxml_root_t;
+struct ezxml_root { // additional data for the root tag
+ struct ezxml xml; // is a super-struct built on top of ezxml struct
+ ezxml_t cur; // current xml tree insertion point
+ char *m; // original xml string
+ size_t len; // length of allocated memory for mmap, -1 for malloc
+ char *u; // UTF-8 conversion of string if original was UTF-16
+ char *s; // start of work area
+ char *e; // end of work area
+ char **ent; // general entities (ampersand sequences)
+ char ***attr; // default attributes
+ char ***pi; // processing instructions
+ short standalone; // non-zero if <?xml standalone="yes"?>
+ char err[EZXML_ERRL]; // error string
+};
+
+char *EZXML_NIL[] = { NULL }; // empty, null terminated array of strings
+
+// returns the first child tag with the given name or NULL if not found
+ezxml_t ezxml_child(ezxml_t xml, const char *name)
+{
+ xml = (xml) ? xml->child : NULL;
+ while (xml && strcmp(name, xml->name)) xml = xml->sibling;
+ return xml;
+}
+
+// returns the Nth tag with the same name in the same subsection or NULL if not
+// found
+ezxml_t ezxml_idx(ezxml_t xml, int idx)
+{
+ for (; xml && idx; idx--) xml = xml->next;
+ return xml;
+}
+
+// returns the value of the requested tag attribute or NULL if not found
+const char *ezxml_attr(ezxml_t xml, const char *attr)
+{
+ int i = 0, j = 1;
+ ezxml_root_t root = (ezxml_root_t)xml;
+
+ if (! xml || ! xml->attr) return NULL;
+ while (xml->attr[i] && strcmp(attr, xml->attr[i])) i += 2;
+ if (xml->attr[i]) return xml->attr[i + 1]; // found attribute
+
+ while (root->xml.parent) root = (ezxml_root_t)root->xml.parent; // root tag
+ for (i = 0; root->attr[i] && strcmp(xml->name, root->attr[i][0]); i++);
+ if (! root->attr[i]) return NULL; // no matching default attributes
+ while (root->attr[i][j] && strcmp(attr, root->attr[i][j])) j += 3;
+ return (root->attr[i][j]) ? root->attr[i][j + 1] : NULL; // found default
+}
+
+// same as ezxml_get but takes an already initialized va_list
+ezxml_t ezxml_vget(ezxml_t xml, va_list ap)
+{
+ char *name = va_arg(ap, char *);
+ int idx = -1;
+
+ if (name && *name) {
+ idx = va_arg(ap, int);
+ xml = ezxml_child(xml, name);
+ }
+ return (idx < 0) ? xml : ezxml_vget(ezxml_idx(xml, idx), ap);
+}
+
+// Traverses the xml tree to retrieve a specific subtag. Takes a variable
+// length list of tag names and indexes. The argument list must be terminated
+// by either an index of -1 or an empty string tag name. Example:
+// title = ezxml_get(library, "shelf", 0, "book", 2, "title", -1);
+// This retrieves the title of the 3rd book on the 1st shelf of library.
+// Returns NULL if not found.
+ezxml_t ezxml_get(ezxml_t xml, ...)
+{
+ va_list ap;
+ ezxml_t r;
+
+ va_start(ap, xml);
+ r = ezxml_vget(xml, ap);
+ va_end(ap);
+ return r;
+}
+
+// returns a null terminated array of processing instructions for the given
+// target
+const char **ezxml_pi(ezxml_t xml, const char *target)
+{
+ ezxml_root_t root = (ezxml_root_t)xml;
+ int i = 0;
+
+ if (! root) return (const char **)EZXML_NIL;
+ while (root->xml.parent) root = (ezxml_root_t)root->xml.parent; // root tag
+ while (root->pi[i] && strcmp(target, root->pi[i][0])) i++; // find target
+ return (const char **)((root->pi[i]) ? root->pi[i] + 1 : EZXML_NIL);
+}
+
+// set an error string and return root
+ezxml_t ezxml_err(ezxml_root_t root, char *s, const char *err, ...)
+{
+ va_list ap;
+ int line = 1;
+ char *t, fmt[EZXML_ERRL];
+
+ for (t = root->s; t < s; t++) if (*t == '\n') line++;
+ snprintf(fmt, EZXML_ERRL, "[error near line %d]: %s", line, err);
+
+ va_start(ap, err);
+ vsnprintf(root->err, EZXML_ERRL, fmt, ap);
+ va_end(ap);
+
+ return &root->xml;
+}
+
+// Recursively decodes entity and character references and normalizes new lines
+// ent is a null terminated array of alternating entity names and values. set t
+// to '&' for general entity decoding, '%' for parameter entity decoding, 'c'
+// for cdata sections, ' ' for attribute normalization, or '*' for non-cdata
+// attribute normalization. Returns s, or if the decoded string is longer than
+// s, returns a malloced string that must be freed.
+char *ezxml_decode(char *s, char **ent, char t)
+{
+ char *e, *r = s, *m = s;
+ long b, c, d, l;
+
+ for (; *s; s++) { // normalize line endings
+ while (*s == '\r') {
+ *(s++) = '\n';
+ if (*s == '\n') memmove(s, (s + 1), strlen(s));
+ }
+ }
+
+ for (s = r; ; ) {
+ while (*s && *s != '&' && (*s != '%' || t != '%') && !isspace(*s)) s++;
+
+ if (! *s) break;
+ else if (t != 'c' && ! strncmp(s, "&#", 2)) { // character reference
+ if (s[2] == 'x') c = strtol(s + 3, &e, 16); // base 16
+ else c = strtol(s + 2, &e, 10); // base 10
+ if (! c || *e != ';') { s++; continue; } // not a character ref
+
+ if (c < 0x80) *(s++) = c; // US-ASCII subset
+ else { // multi-byte UTF-8 sequence
+ for (b = 0, d = c; d; d /= 2) b++; // number of bits in c
+ b = (b - 2) / 5; // number of bytes in payload
+ *(s++) = (0xFF << (7 - b)) | (c >> (6 * b)); // head
+ while (b) *(s++) = 0x80 | ((c >> (6 * --b)) & 0x3F); // payload
+ }
+
+ memmove(s, strchr(s, ';') + 1, strlen(strchr(s, ';')));
+ }
+ else if ((*s == '&' && (t == '&' || t == ' ' || t == '*')) ||
+ (*s == '%' && t == '%')) { // entity reference
+ for (b = 0; ent[b] && strncmp(s + 1, ent[b], strlen(ent[b]));
+ b += 2); // find entity in entity list
+
+ if (ent[b++]) { // found a match
+ if ((c = strlen(ent[b])) - 1 > (e = strchr(s, ';')) - s) {
+ l = (d = (s - r)) + c + strlen(e); // new length
+ r = (r == m) ? strcpy(malloc(l), r) : realloc(r, l);
+ e = strchr((s = r + d), ';'); // fix up pointers
+ }
+
+ memmove(s + c, e + 1, strlen(e)); // shift rest of string
+ strncpy(s, ent[b], c); // copy in replacement text
+ }
+ else s++; // not a known entity
+ }
+ else if ((t == ' ' || t == '*') && isspace(*s)) *(s++) = ' ';
+ else s++; // no decoding needed
+ }
+
+ if (t == '*') { // normalize spaces for non-cdata attributes
+ for (s = r; *s; s++) {
+ if ((l = strspn(s, " "))) memmove(s, s + l, strlen(s + l) + 1);
+ while (*s && *s != ' ') s++;
+ }
+ if (--s >= r && *s == ' ') *s = '\0'; // trim any trailing space
+ }
+ return r;
+}
+
+// called when parser finds start of new tag
+void ezxml_open_tag(ezxml_root_t root, char *name, char **attr)
+{
+ ezxml_t xml = root->cur;
+
+ if (xml->name) xml = ezxml_add_child(xml, name, strlen(xml->txt));
+ else xml->name = name; // first open tag
+
+ xml->attr = attr;
+ root->cur = xml; // update tag insertion point
+}
+
+// called when parser finds character content between open and closing tag
+void ezxml_char_content(ezxml_root_t root, char *s, size_t len, char t)
+{
+ ezxml_t xml = root->cur;
+ char *m = s;
+ size_t l;
+
+ if (! xml || ! xml->name || ! len) return; // sanity check
+
+ s[len] = '\0'; // null terminate text (calling functions anticipate this)
+ len = strlen(s = ezxml_decode(s, root->ent, t)) + 1;
+
+ if (! *(xml->txt)) xml->txt = s; // initial character content
+ else { // allocate our own memory and make a copy
+ xml->txt = (xml->flags & EZXML_TXTM) // allocate some space
+ ? realloc(xml->txt, (l = strlen(xml->txt)) + len)
+ : strcpy(malloc((l = strlen(xml->txt)) + len), xml->txt);
+ strcpy(xml->txt + l, s); // add new char content
+ if (s != m) free(s); // free s if it was malloced by ezxml_decode()
+ }
+
+ if (xml->txt != m) ezxml_set_flag(xml, EZXML_TXTM);
+}
+
+// called when parser finds closing tag
+ezxml_t ezxml_close_tag(ezxml_root_t root, char *name, char *s)
+{
+ if (! root->cur || ! root->cur->name || strcmp(name, root->cur->name))
+ return ezxml_err(root, s, "unexpected closing tag </%s>", name);
+
+ root->cur = root->cur->parent;
+ return NULL;
+}
+
+// checks for circular entity references, returns non-zero if no circular
+// references are found, zero otherwise
+int ezxml_ent_ok(char *name, char *s, char **ent)
+{
+ int i;
+
+ for (; ; s++) {
+ while (*s && *s != '&') s++; // find next entity reference
+ if (! *s) return 1;
+ if (! strncmp(s + 1, name, strlen(name))) return 0; // circular ref.
+ for (i = 0; ent[i] && strncmp(ent[i], s + 1, strlen(ent[i])); i += 2);
+ if (ent[i] && ! ezxml_ent_ok(name, ent[i + 1], ent)) return 0;
+ }
+}
+
+// called when the parser finds a processing instruction
+void ezxml_proc_inst(ezxml_root_t root, char *s, size_t len)
+{
+ int i = 0, j = 1;
+ char *target = s;
+
+ s[len] = '\0'; // null terminate instruction
+ if (*(s += strcspn(s, EZXML_WS))) {
+ *s = '\0'; // null terminate target
+ s += strspn(s + 1, EZXML_WS) + 1; // skip whitespace after target
+ }
+
+ if (! strcmp(target, "xml")) { // <?xml ... ?>
+ if ((s = strstr(s, "standalone")) && ! strncmp(s + strspn(s + 10,
+ EZXML_WS "='\"") + 10, "yes", 3)) root->standalone = 1;
+ return;
+ }
+
+ if (! root->pi[0]) *(root->pi = malloc(sizeof(char **))) = NULL; //first pi
+
+ while (root->pi[i] && strcmp(target, root->pi[i][0])) i++; // find target
+ if (! root->pi[i]) { // new target
+ root->pi = realloc(root->pi, sizeof(char **) * (i + 2));
+ root->pi[i] = malloc(sizeof(char *) * 3);
+ root->pi[i][0] = target;
+ root->pi[i][1] = (char *)(root->pi[i + 1] = NULL); // terminate pi list
+ root->pi[i][2] = strdup(""); // empty document position list
+ }
+
+ while (root->pi[i][j]) j++; // find end of instruction list for this target
+ root->pi[i] = realloc(root->pi[i], sizeof(char *) * (j + 3));
+ root->pi[i][j + 2] = realloc(root->pi[i][j + 1], j + 1);
+ strcpy(root->pi[i][j + 2] + j - 1, (root->xml.name) ? ">" : "<");
+ root->pi[i][j + 1] = NULL; // null terminate pi list for this target
+ root->pi[i][j] = s; // set instruction
+}
+
+// called when the parser finds an internal doctype subset
+short ezxml_internal_dtd(ezxml_root_t root, char *s, size_t len)
+{
+ char q, *c, *t, *n = NULL, *v, **ent, **pe;
+ int i, j;
+
+ pe = memcpy(malloc(sizeof(EZXML_NIL)), EZXML_NIL, sizeof(EZXML_NIL));
+
+ for (s[len] = '\0'; s; ) {
+ while (*s && *s != '<' && *s != '%') s++; // find next declaration
+
+ if (! *s) break;
+ else if (! strncmp(s, "<!ENTITY", 8)) { // parse entity definitions
+ c = s += strspn(s + 8, EZXML_WS) + 8; // skip white space separator
+ n = s + strspn(s, EZXML_WS "%"); // find name
+ *(s = n + strcspn(n, EZXML_WS)) = ';'; // append ; to name
+
+ v = s + strspn(s + 1, EZXML_WS) + 1; // find value
+ if ((q = *(v++)) != '"' && q != '\'') { // skip externals
+ s = strchr(s, '>');
+ continue;
+ }
+
+ for (i = 0, ent = (*c == '%') ? pe : root->ent; ent[i]; i++);
+ ent = realloc(ent, (i + 3) * sizeof(char *)); // space for next ent
+ if (*c == '%') pe = ent;
+ else root->ent = ent;
+
+ *(++s) = '\0'; // null terminate name
+ if ((s = strchr(v, q))) *(s++) = '\0'; // null terminate value
+ ent[i + 1] = ezxml_decode(v, pe, '%'); // set value
+ ent[i + 2] = NULL; // null terminate entity list
+ if (! ezxml_ent_ok(n, ent[i + 1], ent)) { // circular reference
+ if (ent[i + 1] != v) free(ent[i + 1]);
+ ezxml_err(root, v, "circular entity declaration &%s", n);
+ break;
+ }
+ else ent[i] = n; // set entity name
+ }
+ else if (! strncmp(s, "<!ATTLIST", 9)) { // parse default attributes
+ t = s + strspn(s + 9, EZXML_WS) + 9; // skip whitespace separator
+ if (! *t) { ezxml_err(root, t, "unclosed <!ATTLIST"); break; }
+ if (*(s = t + strcspn(t, EZXML_WS ">")) == '>') continue;
+ else *s = '\0'; // null terminate tag name
+ for (i = 0; root->attr[i] && strcmp(n, root->attr[i][0]); i++);
+
+ while (*(n = ++s + strspn(s, EZXML_WS)) && *n != '>') {
+ if (*(s = n + strcspn(n, EZXML_WS))) *s = '\0'; // attr name
+ else { ezxml_err(root, t, "malformed <!ATTLIST"); break; }
+
+ s += strspn(s + 1, EZXML_WS) + 1; // find next token
+ c = (strncmp(s, "CDATA", 5)) ? "*" : " "; // is it cdata?
+ if (! strncmp(s, "NOTATION", 8))
+ s += strspn(s + 8, EZXML_WS) + 8;
+ s = (*s == '(') ? strchr(s, ')') : s + strcspn(s, EZXML_WS);
+ if (! s) { ezxml_err(root, t, "malformed <!ATTLIST"); break; }
+
+ s += strspn(s, EZXML_WS ")"); // skip white space separator
+ if (! strncmp(s, "#FIXED", 6))
+ s += strspn(s + 6, EZXML_WS) + 6;
+ if (*s == '#') { // no default value
+ s += strcspn(s, EZXML_WS ">") - 1;
+ if (*c == ' ') continue; // cdata is default, nothing to do
+ v = NULL;
+ }
+ else if ((*s == '"' || *s == '\'') && // default value
+ (s = strchr(v = s + 1, *s))) *s = '\0';
+ else { ezxml_err(root, t, "malformed <!ATTLIST"); break; }
+
+ if (! root->attr[i]) { // new tag name
+ root->attr = (! i) ? malloc(2 * sizeof(char **))
+ : realloc(root->attr,
+ (i + 2) * sizeof(char **));
+ root->attr[i] = malloc(2 * sizeof(char *));
+ root->attr[i][0] = t; // set tag name
+ root->attr[i][1] = (char *)(root->attr[i + 1] = NULL);
+ }
+
+ for (j = 1; root->attr[i][j]; j += 3); // find end of list
+ root->attr[i] = realloc(root->attr[i],
+ (j + 4) * sizeof(char *));
+
+ root->attr[i][j + 3] = NULL; // null terminate list
+ root->attr[i][j + 2] = c; // is it cdata?
+ root->attr[i][j + 1] = (v) ? ezxml_decode(v, root->ent, *c)
+ : NULL;
+ root->attr[i][j] = n; // attribute name
+ }
+ }
+ else if (! strncmp(s, "<!--", 4)) s = strstr(s + 4, "-->"); // comments
+ else if (! strncmp(s, "<?", 2)) { // processing instructions
+ if ((s = strstr(c = s + 2, "?>")))
+ ezxml_proc_inst(root, c, s++ - c);
+ }
+ else if (*s == '<') s = strchr(s, '>'); // skip other declarations
+ else if (*(s++) == '%' && ! root->standalone) break;
+ }
+
+ free(pe);
+ return ! *root->err;
+}
+
+// Converts a UTF-16 string to UTF-8. Returns a new string that must be freed
+// or NULL if no conversion was needed.
+char *ezxml_str2utf8(char **s, size_t *len)
+{
+ char *u;
+ size_t l = 0, sl, max = *len;
+ long c, d;
+ int b, be = (**s == '\xFE') ? 1 : (**s == '\xFF') ? 0 : -1;
+
+ if (be == -1) return NULL; // not UTF-16
+
+ u = malloc(max);
+ for (sl = 2; sl < *len - 1; sl += 2) {
+ c = (be) ? (((*s)[sl] & 0xFF) << 8) | ((*s)[sl + 1] & 0xFF) //UTF-16BE
+ : (((*s)[sl + 1] & 0xFF) << 8) | ((*s)[sl] & 0xFF); //UTF-16LE
+ if (c >= 0xD800 && c <= 0xDFFF && (sl += 2) < *len - 1) { // high-half
+ d = (be) ? (((*s)[sl] & 0xFF) << 8) | ((*s)[sl + 1] & 0xFF)
+ : (((*s)[sl + 1] & 0xFF) << 8) | ((*s)[sl] & 0xFF);
+ c = (((c & 0x3FF) << 10) | (d & 0x3FF)) + 0x10000;
+ }
+
+ while (l + 6 > max) u = realloc(u, max += EZXML_BUFSIZE);
+ if (c < 0x80) u[l++] = c; // US-ASCII subset
+ else { // multi-byte UTF-8 sequence
+ for (b = 0, d = c; d; d /= 2) b++; // bits in c
+ b = (b - 2) / 5; // bytes in payload
+ u[l++] = (0xFF << (7 - b)) | (c >> (6 * b)); // head
+ while (b) u[l++] = 0x80 | ((c >> (6 * --b)) & 0x3F); // payload
+ }
+ }
+ return *s = realloc(u, *len = l);
+}
+
+// frees a tag attribute list
+void ezxml_free_attr(char **attr) {
+ int i = 0;
+ char *m;
+
+ if (! attr || attr == EZXML_NIL) return; // nothing to free
+ while (attr[i]) i += 2; // find end of attribute list
+ m = attr[i + 1]; // list of which names and values are malloced
+ for (i = 0; m[i]; i++) {
+ if (m[i] & EZXML_NAMEM) free(attr[i * 2]);
+ if (m[i] & EZXML_TXTM) free(attr[(i * 2) + 1]);
+ }
+ free(m);
+ free(attr);
+}
+
+// parse the given xml string and return an ezxml structure
+ezxml_t ezxml_parse_str(char *s, size_t len)
+{
+ ezxml_root_t root = (ezxml_root_t)ezxml_new(NULL);
+ char q, e, *d, **attr, **a = NULL; // initialize a to avoid compile warning
+ int l, i, j;
+
+ root->m = s;
+ if (! len) return ezxml_err(root, NULL, "root tag missing");
+ root->u = ezxml_str2utf8(&s, &len); // convert utf-16 to utf-8
+ root->e = (root->s = s) + len; // record start and end of work area
+
+ e = s[len - 1]; // save end char
+ s[len - 1] = '\0'; // turn end char into null terminator
+
+ while (*s && *s != '<') s++; // find first tag
+ if (! *s) return ezxml_err(root, s, "root tag missing");
+
+ for (; ; ) {
+ attr = (char **)EZXML_NIL;
+ d = ++s;
+
+ if (isalpha(*s) || *s == '_' || *s == ':' || *s < '\0') { // new tag
+ if (! root->cur)
+ return ezxml_err(root, d, "markup outside of root element");
+
+ s += strcspn(s, EZXML_WS "/>");
+ while (isspace(*s)) *(s++) = '\0'; // null terminate tag name
+
+ if (*s && *s != '/' && *s != '>') // find tag in default attr list
+ for (i = 0; (a = root->attr[i]) && strcmp(a[0], d); i++);
+
+ for (l = 0; *s && *s != '/' && *s != '>'; l += 2) { // new attrib
+ attr = (l) ? realloc(attr, (l + 4) * sizeof(char *))
+ : malloc(4 * sizeof(char *)); // allocate space
+ attr[l + 3] = (l) ? realloc(attr[l + 1], (l / 2) + 2)
+ : malloc(2); // mem for list of maloced vals
+ strcpy(attr[l + 3] + (l / 2), " "); // value is not malloced
+ attr[l + 2] = NULL; // null terminate list
+ attr[l + 1] = ""; // temporary attribute value
+ attr[l] = s; // set attribute name
+
+ s += strcspn(s, EZXML_WS "=/>");
+ if (*s == '=' || isspace(*s)) {
+ *(s++) = '\0'; // null terminate tag attribute name
+ q = *(s += strspn(s, EZXML_WS "="));
+ if (q == '"' || q == '\'') { // attribute value
+ attr[l + 1] = ++s;
+ while (*s && *s != q) s++;
+ if (*s) *(s++) = '\0'; // null terminate attribute val
+ else {
+ ezxml_free_attr(attr);
+ return ezxml_err(root, d, "missing %c", q);
+ }
+
+ for (j = 1; a && a[j] && strcmp(a[j], attr[l]); j +=3);
+ attr[l + 1] = ezxml_decode(attr[l + 1], root->ent, (a
+ && a[j]) ? *a[j + 2] : ' ');
+ if (attr[l + 1] < d || attr[l + 1] > s)
+ attr[l + 3][l / 2] = EZXML_TXTM; // value malloced
+ }
+ }
+ while (isspace(*s)) s++;
+ }
+
+ if (*s == '/') { // self closing tag
+ *(s++) = '\0';
+ if ((*s && *s != '>') || (! *s && e != '>')) {
+ if (l) ezxml_free_attr(attr);
+ return ezxml_err(root, d, "missing >");
+ }
+ ezxml_open_tag(root, d, attr);
+ ezxml_close_tag(root, d, s);
+ }
+ else if ((q = *s) == '>' || (! *s && e == '>')) { // open tag
+ *s = '\0'; // temporarily null terminate tag name
+ ezxml_open_tag(root, d, attr);
+ *s = q;
+ }
+ else {
+ if (l) ezxml_free_attr(attr);
+ return ezxml_err(root, d, "missing >");
+ }
+ }
+ else if (*s == '/') { // close tag
+ s += strcspn(d = s + 1, EZXML_WS ">") + 1;
+ if (! (q = *s) && e != '>') return ezxml_err(root, d, "missing >");
+ *s = '\0'; // temporarily null terminate tag name
+ if (ezxml_close_tag(root, d, s)) return &root->xml;
+ if (isspace(*s = q)) s += strspn(s, EZXML_WS);
+ }
+ else if (! strncmp(s, "!--", 3)) { // xml comment
+ if (! (s = strstr(s + 3, "--")) || (*(s += 2) != '>' && *s) ||
+ (! *s && e != '>')) return ezxml_err(root, d, "unclosed <!--");
+ }
+ else if (! strncmp(s, "![CDATA[", 8)) { // cdata
+ if ((s = strstr(s, "]]>")))
+ ezxml_char_content(root, d + 8, (s += 2) - d - 10, 'c');
+ else return ezxml_err(root, d, "unclosed <![CDATA[");
+ }
+ else if (! strncmp(s, "!DOCTYPE", 8)) { // dtd
+ for (l = 0; *s && ((! l && *s != '>') || (l && (*s != ']' ||
+ *(s + strspn(s + 1, EZXML_WS) + 1) != '>')));
+ l = (*s == '[') ? 1 : l) s += strcspn(s + 1, "[]>") + 1;
+ if (! *s && e != '>')
+ return ezxml_err(root, d, "unclosed <!DOCTYPE");
+ d = (l) ? strchr(d, '[') + 1 : d;
+ if (l && ! ezxml_internal_dtd(root, d, s++ - d)) return &root->xml;
+ }
+ else if (*s == '?') { // <?...?> processing instructions
+ do { s = strchr(s, '?'); } while (s && *(++s) && *s != '>');
+ if (! s || (! *s && e != '>'))
+ return ezxml_err(root, d, "unclosed <?");
+ else ezxml_proc_inst(root, d + 1, s - d - 2);
+ }
+ else return ezxml_err(root, d, "unexpected <");
+
+ if (! s || ! *s) break;
+ *s = '\0';
+ d = ++s;
+ if (*s && *s != '<') { // tag character content
+ while (*s && *s != '<') s++;
+ if (*s) ezxml_char_content(root, d, s - d, '&');
+ else break;
+ }
+ else if (! *s) break;
+ }
+
+ if (! root->cur) return &root->xml;
+ else if (! root->cur->name) return ezxml_err(root, d, "root tag missing");
+ else return ezxml_err(root, d, "unclosed tag <%s>", root->cur->name);
+}
+
+// Wrapper for ezxml_parse_str() that accepts a file stream. Reads the entire
+// stream into memory and then parses it. For xml files, use ezxml_parse_file()
+// or ezxml_parse_fd()
+ezxml_t ezxml_parse_fp(FILE *fp)
+{
+ ezxml_root_t root;
+ size_t l, len = 0;
+ char *s;
+
+ if (! (s = malloc(EZXML_BUFSIZE))) return NULL;
+ do {
+ len += (l = fread((s + len), 1, EZXML_BUFSIZE, fp));
+ if (l == EZXML_BUFSIZE) s = realloc(s, len + EZXML_BUFSIZE);
+ } while (s && l == EZXML_BUFSIZE);
+
+ if (! s) return NULL;
+ root = (ezxml_root_t)ezxml_parse_str(s, len);
+ root->len = -1; // so we know to free s in ezxml_free()
+ return &root->xml;
+}
+
+// A wrapper for ezxml_parse_str() that accepts a file descriptor. First
+// attempts to mem map the file. Failing that, reads the file into memory.
+// Returns NULL on failure.
+ezxml_t ezxml_parse_fd(int fd)
+{
+ ezxml_root_t root;
+ struct stat st;
+ size_t l;
+ void *m;
+
+ if (fd < 0) return NULL;
+ fstat(fd, &st);
+
+#ifndef EZXML_NOMMAP
+ l = (st.st_size + sysconf(_SC_PAGESIZE) - 1) & ~(sysconf(_SC_PAGESIZE) -1);
+ if ((m = mmap(NULL, l, PROT_READ | PROT_WRITE, MAP_PRIVATE, fd, 0)) !=
+ MAP_FAILED) {
+ madvise(m, l, MADV_SEQUENTIAL); // optimize for sequential access
+ root = (ezxml_root_t)ezxml_parse_str(m, st.st_size);
+ madvise(m, root->len = l, MADV_NORMAL); // put it back to normal
+ }
+ else { // mmap failed, read file into memory
+#endif // EZXML_NOMMAP
+ l = read(fd, m = malloc(st.st_size), st.st_size);
+ root = (ezxml_root_t)ezxml_parse_str(m, l);
+ root->len = -1; // so we know to free s in ezxml_free()
+#ifndef EZXML_NOMMAP
+ }
+#endif // EZXML_NOMMAP
+ return &root->xml;
+}
+
+// a wrapper for ezxml_parse_fd that accepts a file name
+ezxml_t ezxml_parse_file(const char *file)
+{
+ int fd = open(file, O_RDONLY, 0);
+ ezxml_t xml = ezxml_parse_fd(fd);
+
+ if (fd >= 0) close(fd);
+ return xml;
+}
+
+// Encodes ampersand sequences appending the results to *dst, reallocating *dst
+// if length excedes max. a is non-zero for attribute encoding. Returns *dst
+char *ezxml_ampencode(const char *s, size_t len, char **dst, size_t *dlen,
+ size_t *max, short a)
+{
+ const char *e;
+
+ for (e = s + len; s != e; s++) {
+ while (*dlen + 10 > *max) *dst = realloc(*dst, *max += EZXML_BUFSIZE);
+
+ switch (*s) {
+ case '\0': return *dst;
+ case '&': *dlen += sprintf(*dst + *dlen, "&"); break;
+ case '<': *dlen += sprintf(*dst + *dlen, "<"); break;
+ case '>': *dlen += sprintf(*dst + *dlen, ">"); break;
+ case '"': *dlen += sprintf(*dst + *dlen, (a) ? """ : "\""); break;
+ case '\n': *dlen += sprintf(*dst + *dlen, (a) ? "
" : "\n"); break;
+ case '\t': *dlen += sprintf(*dst + *dlen, (a) ? "	" : "\t"); break;
+ case '\r': *dlen += sprintf(*dst + *dlen, "
"); break;
+ default: (*dst)[(*dlen)++] = *s;
+ }
+ }
+ return *dst;
+}
+
+// Recursively converts each tag to xml appending it to *s. Reallocates *s if
+// its length excedes max. start is the location of the previous tag in the
+// parent tag's character content. Returns *s.
+char *ezxml_toxml_r(ezxml_t xml, char **s, size_t *len, size_t *max,
+ size_t start, char ***attr)
+{
+ int i, j;
+ char *txt = (xml->parent) ? xml->parent->txt : "";
+ size_t off = 0;
+
+ // parent character content up to this tag
+ *s = ezxml_ampencode(txt + start, xml->off - start, s, len, max, 0);
+
+ while (*len + strlen(xml->name) + 4 > *max) // reallocate s
+ *s = realloc(*s, *max += EZXML_BUFSIZE);
+
+ *len += sprintf(*s + *len, "<%s", xml->name); // open tag
+ for (i = 0; xml->attr[i]; i += 2) { // tag attributes
+ if (ezxml_attr(xml, xml->attr[i]) != xml->attr[i + 1]) continue;
+ while (*len + strlen(xml->attr[i]) + 7 > *max) // reallocate s
+ *s = realloc(*s, *max += EZXML_BUFSIZE);
+
+ *len += sprintf(*s + *len, " %s=\"", xml->attr[i]);
+ ezxml_ampencode(xml->attr[i + 1], -1, s, len, max, 1);
+ *len += sprintf(*s + *len, "\"");
+ }
+
+ for (i = 0; attr[i] && strcmp(attr[i][0], xml->name); i++);
+ for (j = 1; attr[i] && attr[i][j]; j += 3) { // default attributes
+ if (! attr[i][j + 1] || ezxml_attr(xml, attr[i][j]) != attr[i][j + 1])
+ continue; // skip duplicates and non-values
+ while (*len + strlen(attr[i][j]) + 7 > *max) // reallocate s
+ *s = realloc(*s, *max += EZXML_BUFSIZE);
+
+ *len += sprintf(*s + *len, " %s=\"", attr[i][j]);
+ ezxml_ampencode(attr[i][j + 1], -1, s, len, max, 1);
+ *len += sprintf(*s + *len, "\"");
+ }
+ *len += sprintf(*s + *len, ">");
+
+ *s = (xml->child) ? ezxml_toxml_r(xml->child, s, len, max, 0, attr) //child
+ : ezxml_ampencode(xml->txt, -1, s, len, max, 0); //data
+
+ while (*len + strlen(xml->name) + 4 > *max) // reallocate s
+ *s = realloc(*s, *max += EZXML_BUFSIZE);
+
+ *len += sprintf(*s + *len, "</%s>", xml->name); // close tag
+
+ while (txt[off] && off < xml->off) off++; // make sure off is within bounds
+ return (xml->ordered) ? ezxml_toxml_r(xml->ordered, s, len, max, off, attr)
+ : ezxml_ampencode(txt + off, -1, s, len, max, 0);
+}
+
+// Converts an ezxml structure back to xml. Returns a string of xml data that
+// must be freed.
+char *ezxml_toxml(ezxml_t xml)
+{
+ ezxml_t p = (xml) ? xml->parent : NULL, o = (xml) ? xml->ordered : NULL;
+ ezxml_root_t root = (ezxml_root_t)xml;
+ size_t len = 0, max = EZXML_BUFSIZE;
+ char *s = strcpy(malloc(max), ""), *t, *n;
+ int i, j, k;
+
+ if (! xml || ! xml->name) return realloc(s, len + 1);
+ while (root->xml.parent) root = (ezxml_root_t)root->xml.parent; // root tag
+
+ for (i = 0; ! p && root->pi[i]; i++) { // pre-root processing instructions
+ for (k = 2; root->pi[i][k - 1]; k++);
+ for (j = 1; (n = root->pi[i][j]); j++) {
+ if (root->pi[i][k][j - 1] == '>') continue; // not pre-root
+ while (len + strlen(t = root->pi[i][0]) + strlen(n) + 7 > max)
+ s = realloc(s, max += EZXML_BUFSIZE);
+ len += sprintf(s + len, "<?%s%s%s?>\n", t, *n ? " " : "", n);
+ }
+ }
+
+ xml->parent = xml->ordered = NULL;
+ s = ezxml_toxml_r(xml, &s, &len, &max, 0, root->attr);
+ xml->parent = p;
+ xml->ordered = o;
+
+ for (i = 0; ! p && root->pi[i]; i++) { // post-root processing instructions
+ for (k = 2; root->pi[i][k - 1]; k++);
+ for (j = 1; (n = root->pi[i][j]); j++) {
+ if (root->pi[i][k][j - 1] == '<') continue; // not post-root
+ while (len + strlen(t = root->pi[i][0]) + strlen(n) + 7 > max)
+ s = realloc(s, max += EZXML_BUFSIZE);
+ len += sprintf(s + len, "\n<?%s%s%s?>", t, *n ? " " : "", n);
+ }
+ }
+ return realloc(s, len + 1);
+}
+
+// free the memory allocated for the ezxml structure
+void ezxml_free(ezxml_t xml)
+{
+ ezxml_root_t root = (ezxml_root_t)xml;
+ int i, j;
+ char **a, *s;
+
+ if (! xml) return;
+ ezxml_free(xml->child);
+ ezxml_free(xml->ordered);
+
+ if (! xml->parent) { // free root tag allocations
+ for (i = 10; root->ent[i]; i += 2) // 0 - 9 are default entites (<>&"')
+ if ((s = root->ent[i + 1]) < root->s || s > root->e) free(s);
+ free(root->ent); // free list of general entities
+
+ for (i = 0; (a = root->attr[i]); i++) {
+ for (j = 1; a[j++]; j += 2) // free malloced attribute values
+ if (a[j] && (a[j] < root->s || a[j] > root->e)) free(a[j]);
+ free(a);
+ }
+ if (root->attr[0]) free(root->attr); // free default attribute list
+
+ for (i = 0; root->pi[i]; i++) {
+ for (j = 1; root->pi[i][j]; j++);
+ free(root->pi[i][j + 1]);
+ free(root->pi[i]);
+ }
+ if (root->pi[0]) free(root->pi); // free processing instructions
+
+ if (root->len == (size_t) -1) free(root->m); // malloced xml data
+#ifndef EZXML_NOMMAP
+ else if (root->len) munmap(root->m, root->len); // mem mapped xml data
+#endif // EZXML_NOMMAP
+ if (root->u) free(root->u); // utf8 conversion
+ }
+
+ ezxml_free_attr(xml->attr); // tag attributes
+ if ((xml->flags & EZXML_TXTM)) free(xml->txt); // character content
+ if ((xml->flags & EZXML_NAMEM)) free(xml->name); // tag name
+ free(xml);
+}
+
+// return parser error message or empty string if none
+const char *ezxml_error(ezxml_t xml)
+{
+ while (xml && xml->parent) xml = xml->parent; // find root tag
+ return (xml) ? ((ezxml_root_t)xml)->err : "";
+}
+
+// returns a new empty ezxml structure with the given root tag name
+ezxml_t ezxml_new(const char *name)
+{
+ static char *ent[] = { "lt;", "<", "gt;", ">", "quot;", """,
+ "apos;", "'", "amp;", "&", NULL };
+ ezxml_root_t root = (ezxml_root_t)memset(malloc(sizeof(struct ezxml_root)),
+ '\0', sizeof(struct ezxml_root));
+ root->xml.name = (char *)name;
+ root->cur = &root->xml;
+ strcpy(root->err, root->xml.txt = "");
+ root->ent = memcpy(malloc(sizeof(ent)), ent, sizeof(ent));
+ root->attr = root->pi = (char ***)(root->xml.attr = EZXML_NIL);
+ return &root->xml;
+}
+
+// inserts an existing tag into an ezxml structure
+ezxml_t ezxml_insert(ezxml_t xml, ezxml_t dest, size_t off)
+{
+ ezxml_t cur, prev, head;
+
+ xml->next = xml->sibling = xml->ordered = NULL;
+ xml->off = off;
+ xml->parent = dest;
+
+ if ((head = dest->child)) { // already have sub tags
+ if (head->off <= off) { // not first subtag
+ for (cur = head; cur->ordered && cur->ordered->off <= off;
+ cur = cur->ordered);
+ xml->ordered = cur->ordered;
+ cur->ordered = xml;
+ }
+ else { // first subtag
+ xml->ordered = head;
+ dest->child = xml;
+ }
+
+ for (cur = head, prev = NULL; cur && strcmp(cur->name, xml->name);
+ prev = cur, cur = cur->sibling); // find tag type
+ if (cur && cur->off <= off) { // not first of type
+ while (cur->next && cur->next->off <= off) cur = cur->next;
+ xml->next = cur->next;
+ cur->next = xml;
+ }
+ else { // first tag of this type
+ if (prev && cur) prev->sibling = cur->sibling; // remove old first
+ xml->next = cur; // old first tag is now next
+ for (cur = head, prev = NULL; cur && cur->off <= off;
+ prev = cur, cur = cur->sibling); // new sibling insert point
+ xml->sibling = cur;
+ if (prev) prev->sibling = xml;
+ }
+ }
+ else dest->child = xml; // only sub tag
+
+ return xml;
+}
+
+// Adds a child tag. off is the offset of the child tag relative to the start
+// of the parent tag's character content. Returns the child tag.
+ezxml_t ezxml_add_child(ezxml_t xml, const char *name, size_t off)
+{
+ ezxml_t child;
+
+ if (! xml) return NULL;
+ child = (ezxml_t)memset(malloc(sizeof(struct ezxml)), '\0',
+ sizeof(struct ezxml));
+ child->name = (char *)name;
+ child->attr = EZXML_NIL;
+ child->txt = "";
+
+ return ezxml_insert(child, xml, off);
+}
+
+// sets the character content for the given tag and returns the tag
+ezxml_t ezxml_set_txt(ezxml_t xml, const char *txt)
+{
+ if (! xml) return NULL;
+ if (xml->flags & EZXML_TXTM) free(xml->txt); // existing txt was malloced
+ xml->flags &= ~EZXML_TXTM;
+ xml->txt = (char *)txt;
+ return xml;
+}
+
+// Sets the given tag attribute or adds a new attribute if not found. A value
+// of NULL will remove the specified attribute. Returns the tag given.
+ezxml_t ezxml_set_attr(ezxml_t xml, const char *name, const char *value)
+{
+ int l = 0, c;
+
+ if (! xml) return NULL;
+ while (xml->attr[l] && strcmp(xml->attr[l], name)) l += 2;
+ if (! xml->attr[l]) { // not found, add as new attribute
+ if (! value) return xml; // nothing to do
+ if (xml->attr == EZXML_NIL) { // first attribute
+ xml->attr = malloc(4 * sizeof(char *));
+ xml->attr[1] = strdup(""); // empty list of malloced names/vals
+ }
+ else xml->attr = realloc(xml->attr, (l + 4) * sizeof(char *));
+
+ xml->attr[l] = (char *)name; // set attribute name
+ xml->attr[l + 2] = NULL; // null terminate attribute list
+ xml->attr[l + 3] = realloc(xml->attr[l + 1],
+ (c = strlen(xml->attr[l + 1])) + 2);
+ strcpy(xml->attr[l + 3] + c, " "); // set name/value as not malloced
+ if (xml->flags & EZXML_DUP) xml->attr[l + 3][c] = EZXML_NAMEM;
+ }
+ else if (xml->flags & EZXML_DUP) free((char *)name); // name was strduped
+
+ for (c = l; xml->attr[c]; c += 2); // find end of attribute list
+ if (xml->attr[c + 1][l / 2] & EZXML_TXTM) free(xml->attr[l + 1]); //old val
+ if (xml->flags & EZXML_DUP) xml->attr[c + 1][l / 2] |= EZXML_TXTM;
+ else xml->attr[c + 1][l / 2] &= ~EZXML_TXTM;
+
+ if (value) xml->attr[l + 1] = (char *)value; // set attribute value
+ else { // remove attribute
+ if (xml->attr[c + 1][l / 2] & EZXML_NAMEM) free(xml->attr[l]);
+ memmove(xml->attr + l, xml->attr + l + 2, (c - l + 2) * sizeof(char*));
+ xml->attr = realloc(xml->attr, (c + 2) * sizeof(char *));
+ memmove(xml->attr[c + 1] + (l / 2), xml->attr[c + 1] + (l / 2) + 1,
+ (c / 2) - (l / 2)); // fix list of which name/vals are malloced
+ }
+ xml->flags &= ~EZXML_DUP; // clear strdup() flag
+ return xml;
+}
+
+// sets a flag for the given tag and returns the tag
+ezxml_t ezxml_set_flag(ezxml_t xml, short flag)
+{
+ if (xml) xml->flags |= flag;
+ return xml;
+}
+
+// removes a tag along with its subtags without freeing its memory
+ezxml_t ezxml_cut(ezxml_t xml)
+{
+ ezxml_t cur;
+
+ if (! xml) return NULL; // nothing to do
+ if (xml->next) xml->next->sibling = xml->sibling; // patch sibling list
+
+ if (xml->parent) { // not root tag
+ cur = xml->parent->child; // find head of subtag list
+ if (cur == xml) xml->parent->child = xml->ordered; // first subtag
+ else { // not first subtag
+ while (cur->ordered != xml) cur = cur->ordered;
+ cur->ordered = cur->ordered->ordered; // patch ordered list
+
+ cur = xml->parent->child; // go back to head of subtag list
+ if (strcmp(cur->name, xml->name)) { // not in first sibling list
+ while (strcmp(cur->sibling->name, xml->name))
+ cur = cur->sibling;
+ if (cur->sibling == xml) { // first of a sibling list
+ cur->sibling = (xml->next) ? xml->next
+ : cur->sibling->sibling;
+ }
+ else cur = cur->sibling; // not first of a sibling list
+ }
+
+ while (cur->next && cur->next != xml) cur = cur->next;
+ if (cur->next) cur->next = cur->next->next; // patch next list
+ }
+ }
+ xml->ordered = xml->sibling = xml->next = NULL;
+ return xml;
+}
+
+#ifdef EZXML_TEST // test harness
+int main(int argc, char **argv)
+{
+ ezxml_t xml;
+ char *s;
+ int i;
+
+ if (argc != 2) return fprintf(stderr, "usage: %s xmlfile\n", argv[0]);
+
+ xml = ezxml_parse_file(argv[1]);
+ printf("%s\n", (s = ezxml_toxml(xml)));
+ free(s);
+ i = fprintf(stderr, "%s", ezxml_error(xml));
+ ezxml_free(xml);
+ return (i) ? 1 : 0;
+}
+#endif // EZXML_TEST
diff --git a/gstlal/lib/gstlal/ezxml.h b/gstlal/lib/gstlal/ezxml.h
new file mode 100644
index 0000000000..3e020788b4
--- /dev/null
+++ b/gstlal/lib/gstlal/ezxml.h
@@ -0,0 +1,167 @@
+/* ezxml.h
+ *
+ * Copyright 2004-2006 Aaron Voisine <aaron@voisine.org>
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining
+ * a copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sublicense, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included
+ * in all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
+ * IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY
+ * CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
+ * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
+ * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ */
+
+#ifndef _EZXML_H
+#define _EZXML_H
+
+#include <stdlib.h>
+#include <stdio.h>
+#include <stdarg.h>
+#include <fcntl.h>
+
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+#define EZXML_BUFSIZE 1024 // size of internal memory buffers
+#define EZXML_NAMEM 0x80 // name is malloced
+#define EZXML_TXTM 0x40 // txt is malloced
+#define EZXML_DUP 0x20 // attribute name and value are strduped
+
+typedef struct ezxml *ezxml_t;
+struct ezxml {
+ char *name; // tag name
+ char **attr; // tag attributes { name, value, name, value, ... NULL }
+ char *txt; // tag character content, empty string if none
+ size_t off; // tag offset from start of parent tag character content
+ ezxml_t next; // next tag with same name in this section at this depth
+ ezxml_t sibling; // next tag with different name in same section and depth
+ ezxml_t ordered; // next tag, same section and depth, in original order
+ ezxml_t child; // head of sub tag list, NULL if none
+ ezxml_t parent; // parent tag, NULL if current tag is root tag
+ short flags; // additional information
+};
+
+// Given a string of xml data and its length, parses it and creates an ezxml
+// structure. For efficiency, modifies the data by adding null terminators
+// and decoding ampersand sequences. If you don't want this, copy the data and
+// pass in the copy. Returns NULL on failure.
+ezxml_t ezxml_parse_str(char *s, size_t len);
+
+// A wrapper for ezxml_parse_str() that accepts a file descriptor. First
+// attempts to mem map the file. Failing that, reads the file into memory.
+// Returns NULL on failure.
+ezxml_t ezxml_parse_fd(int fd);
+
+// a wrapper for ezxml_parse_fd() that accepts a file name
+ezxml_t ezxml_parse_file(const char *file);
+
+// Wrapper for ezxml_parse_str() that accepts a file stream. Reads the entire
+// stream into memory and then parses it. For xml files, use ezxml_parse_file()
+// or ezxml_parse_fd()
+ezxml_t ezxml_parse_fp(FILE *fp);
+
+// returns the first child tag (one level deeper) with the given name or NULL
+// if not found
+ezxml_t ezxml_child(ezxml_t xml, const char *name);
+
+// returns the next tag of the same name in the same section and depth or NULL
+// if not found
+#define ezxml_next(xml) ((xml) ? xml->next : NULL)
+
+// Returns the Nth tag with the same name in the same section at the same depth
+// or NULL if not found. An index of 0 returns the tag given.
+ezxml_t ezxml_idx(ezxml_t xml, int idx);
+
+// returns the name of the given tag
+#define ezxml_name(xml) ((xml) ? xml->name : NULL)
+
+// returns the given tag's character content or empty string if none
+#define ezxml_txt(xml) ((xml) ? xml->txt : "")
+
+// returns the value of the requested tag attribute, or NULL if not found
+const char *ezxml_attr(ezxml_t xml, const char *attr);
+
+// Traverses the ezxml sturcture to retrieve a specific subtag. Takes a
+// variable length list of tag names and indexes. The argument list must be
+// terminated by either an index of -1 or an empty string tag name. Example:
+// title = ezxml_get(library, "shelf", 0, "book", 2, "title", -1);
+// This retrieves the title of the 3rd book on the 1st shelf of library.
+// Returns NULL if not found.
+ezxml_t ezxml_get(ezxml_t xml, ...);
+
+// Converts an ezxml structure back to xml. Returns a string of xml data that
+// must be freed.
+char *ezxml_toxml(ezxml_t xml);
+
+// returns a NULL terminated array of processing instructions for the given
+// target
+const char **ezxml_pi(ezxml_t xml, const char *target);
+
+// frees the memory allocated for an ezxml structure
+void ezxml_free(ezxml_t xml);
+
+// returns parser error message or empty string if none
+const char *ezxml_error(ezxml_t xml);
+
+// returns a new empty ezxml structure with the given root tag name
+ezxml_t ezxml_new(const char *name);
+
+// wrapper for ezxml_new() that strdup()s name
+#define ezxml_new_d(name) ezxml_set_flag(ezxml_new(strdup(name)), EZXML_NAMEM)
+
+// Adds a child tag. off is the offset of the child tag relative to the start
+// of the parent tag's character content. Returns the child tag.
+ezxml_t ezxml_add_child(ezxml_t xml, const char *name, size_t off);
+
+// wrapper for ezxml_add_child() that strdup()s name
+#define ezxml_add_child_d(xml, name, off) \
+ ezxml_set_flag(ezxml_add_child(xml, strdup(name), off), EZXML_NAMEM)
+
+// sets the character content for the given tag and returns the tag
+ezxml_t ezxml_set_txt(ezxml_t xml, const char *txt);
+
+// wrapper for ezxml_set_txt() that strdup()s txt
+#define ezxml_set_txt_d(xml, txt) \
+ ezxml_set_flag(ezxml_set_txt(xml, strdup(txt)), EZXML_TXTM)
+
+// Sets the given tag attribute or adds a new attribute if not found. A value
+// of NULL will remove the specified attribute. Returns the tag given.
+ezxml_t ezxml_set_attr(ezxml_t xml, const char *name, const char *value);
+
+// Wrapper for ezxml_set_attr() that strdup()s name/value. Value cannot be NULL
+#define ezxml_set_attr_d(xml, name, value) \
+ ezxml_set_attr(ezxml_set_flag(xml, EZXML_DUP), strdup(name), strdup(value))
+
+// sets a flag for the given tag and returns the tag
+ezxml_t ezxml_set_flag(ezxml_t xml, short flag);
+
+// removes a tag along with its subtags without freeing its memory
+ezxml_t ezxml_cut(ezxml_t xml);
+
+// inserts an existing tag into an ezxml structure
+ezxml_t ezxml_insert(ezxml_t xml, ezxml_t dest, size_t off);
+
+// Moves an existing tag to become a subtag of dest at the given offset from
+// the start of dest's character content. Returns the moved tag.
+#define ezxml_move(xml, dest, off) ezxml_insert(ezxml_cut(xml), dest, off)
+
+// removes a tag along with all its subtags
+#define ezxml_remove(xml) ezxml_free(ezxml_cut(xml))
+
+#ifdef __cplusplus
+}
+#endif
+
+#endif // _EZXML_H
--
GitLab