static void
show_lines(source_file *file, uint16_t line_min, uint16_t line_max)
{
char buf[LINE_BUF_SIZE];
int show_lineno = 1, found_newline = 0, is_printed = 0;
if (file->fp == NULL) {
return;
}
while (fgets(buf, sizeof(buf), file->fp) != NULL) {
found_newline = remove_newlines(buf, file->fp);
if (line_min <= file->lineno) {
if (show_lineno) {
printf("%-8d", file->lineno);
}
show_lineno = found_newline;
printf(found_newline ? "%s\n" : "%s", buf);
is_printed = 1;
}
if (found_newline) {
if (line_max < ++file->lineno) {
break;
}
}
}
if (is_printed && !found_newline) {
printf("\n");
}
}
void parse_scenario(FILE *input, struct scenario *scn) {
int r;
struct cmd c;
char linebuf[LINE_LEN + 1];
char cmdbuf[LINE_LEN + 1];
scn->cmd_count = 0;
while (scn->cmd_count < MAX_CMDS && !feof(input)) {
c.cmd = CMD_NOOP;
c.arg = 0;
if (!fgets(linebuf, LINE_LEN, input)) {
linebuf[0] = '\0';
}
remove_newlines(linebuf);
if (linebuf[0] != '#' && linebuf[0] != '\0') {
r = sscanf(linebuf, "%s %d", cmdbuf, &c.arg);
if (r != 2) c.cmd = CMD_SYNTAX_ERROR;
else if (strcmp(cmdbuf, "send") == 0) c.cmd = CMD_SEND;
else if (strcmp(cmdbuf, "wait") == 0) c.cmd = CMD_WAIT;
else if (strcmp(cmdbuf, "tcp_cork") == 0) c.cmd = CMD_CORK;
else if (strcmp(cmdbuf, "tcp_nodelay") == 0) c.cmd = CMD_NODELAY;
else if (strcmp(cmdbuf, "/etc/delay") == 0) c.cmd = CMD_DELAY;
else if (strcmp(cmdbuf, "goto") == 0) c.cmd = CMD_GOTO;
}
scn->cmds[scn->cmd_count++] = c;
}
}
static PangoLayout*
create_layout(HippoCanvasText *text,
int allocation_width)
{
HippoCanvasBox *box = HIPPO_CANVAS_BOX(text);
PangoLayout *layout;
HippoCanvasStyle *style = hippo_canvas_context_get_style(HIPPO_CANVAS_CONTEXT(text));
g_return_val_if_fail(box->context != NULL, NULL);
layout = hippo_canvas_context_create_layout(box->context);
if (box->font_desc) {
PangoFontDescription *merged = pango_font_description_copy(hippo_canvas_style_get_font(style));
pango_font_description_merge(merged, box->font_desc, TRUE);
pango_layout_set_font_description(layout, merged);
pango_font_description_free(merged);
} else {
pango_layout_set_font_description(layout, hippo_canvas_style_get_font(style));
}
{
PangoAttrList *attrs;
HippoTextDecoration decoration = hippo_canvas_style_get_text_decoration(style);
if (text->attributes)
attrs = pango_attr_list_copy(text->attributes);
else
attrs = pango_attr_list_new();
if (ABS(1.0 - text->font_scale) > .000001) {
PangoAttribute *attr = pango_attr_scale_new(text->font_scale);
attr->start_index = 0;
attr->end_index = G_MAXUINT;
pango_attr_list_insert(attrs, attr);
}
if ((decoration & HIPPO_TEXT_DECORATION_UNDERLINE) != 0) {
PangoAttribute *attr = pango_attr_underline_new(TRUE);
attr->start_index = 0;
attr->end_index = G_MAXUINT;
pango_attr_list_insert(attrs, attr);
}
if ((decoration & HIPPO_TEXT_DECORATION_LINE_THROUGH) != 0) {
PangoAttribute *attr = pango_attr_strikethrough_new(TRUE);
attr->start_index = 0;
attr->end_index = G_MAXUINT;
pango_attr_list_insert(attrs, attr);
}
pango_layout_set_attributes(layout, attrs);
pango_attr_list_unref(attrs);
}
if (text->text != NULL) {
pango_layout_set_text(layout, text->text, -1);
}
if (allocation_width >= 0) {
int layout_width, layout_height;
pango_layout_get_size(layout, &layout_width, &layout_height);
layout_width /= PANGO_SCALE;
layout_height /= PANGO_SCALE;
/* Force layout smaller if required, but we don't want to make
* the layout _wider_ because it breaks alignment, so only do
* this if required.
*/
if (layout_width > allocation_width) {
pango_layout_set_width(layout, allocation_width * PANGO_SCALE);
/* If we set ellipsize, then it overrides wrapping. If we get
* too-small allocation for HIPPO_CANVAS_SIZE_FULL_WIDTH, then
* we want to ellipsize instead of wrapping.
*/
if (text->size_mode == HIPPO_CANVAS_SIZE_WRAP_WORD) {
pango_layout_set_ellipsize(layout, PANGO_ELLIPSIZE_NONE);
} else {
pango_layout_set_ellipsize(layout, PANGO_ELLIPSIZE_END);
}
/* For now if we say ellipsize end, we always just want one line.
* Maybe this should be an orthogonal property?
*/
if (text->size_mode == HIPPO_CANVAS_SIZE_ELLIPSIZE_END) {
pango_layout_set_single_paragraph_mode(layout, TRUE);
/* Pango's line separator character in this case is ugly, so we
* fix it. Not a very efficient approach, but oh well.
*/
if (text->text != NULL) {
char *new_text = remove_newlines(text->text);
/* avoid making the layout recompute everything
* if we didn't have newlines anyhow
*/
if (strcmp(text->text, new_text) != 0) {
pango_layout_set_text(layout, new_text, -1);
}
g_free(new_text);
}
}
}
}
return layout;
}
int countwords_HTML(SWISH *sw, FileProp *fprop, FileRec *fi, char *buffer)
{
int ftotalwords;
int *metaID;
int metaIDlen;
int position; /* Position of word in file */
int currentmetanames;
char *p,
*newp,
*tag,
*endtag;
int structure;
FileRec *thisFileEntry = fi;
struct metaEntry *metaNameEntry;
IndexFILE *indexf = sw->indexlist;
struct MOD_Index *idx = sw->Index;
char *Content = NULL,
*Name = NULL,
*summary = NULL;
char *title = (char *)sw_ConvHTMLEntities2ISO(sw, (unsigned char *)parseHTMLtitle(sw, buffer));
if (!isoktitle(sw, title))
return -2;
if (fprop->stordesc)
summary = parseHtmlSummary(buffer, fprop->stordesc->field, fprop->stordesc->size, sw);
addCommonProperties( sw, fprop, fi, title, summary, 0 );
/* Init meta info */
metaID = (int *) Mem_ZoneAlloc(sw->Index->perDocTmpZone,(metaIDlen = 16) * sizeof(int));
currentmetanames = 0;
ftotalwords = 0;
structure = IN_FILE;
metaID[0] = 1;
position = 1;
for (p = buffer; p && *p;)
{
/* Look for non escaped '<' */
if ((tag = strchr(p, '<')) && ((tag == p) || (*(tag - 1) != '\\')))
{
/* Index up to the tag */
*tag++ = '\0';
newp = (char *)sw_ConvHTMLEntities2ISO(sw, (unsigned char *)p);
if ( ! currentmetanames )
currentmetanames++;
ftotalwords += indexstring(sw, newp, idx->filenum, structure, currentmetanames, metaID, &position);
/* Now let us look for a not escaped '>' */
for (endtag = tag;;)
if ((endtag = strchr(endtag, '>')))
{
if (*(endtag - 1) != '\\')
break;
else
endtag++;
}
else
break;
if (endtag)
{
*endtag++ = '\0';
if ((tag[0] == '!') && lstrstr(tag, "META") && (lstrstr(tag, "START") || lstrstr(tag, "END")))
{
/* Check for META TAG TYPE 1 */
structure |= IN_META;
if (lstrstr(tag, "START"))
{
char *parsed_tag;
if (
(metaNameEntry =
getHTMLMeta(indexf, tag, sw, NULL, &parsed_tag, fprop->real_path)))
{
/* realloc memory if needed */
if (currentmetanames == metaIDlen)
{
int *newbuf = (int *)Mem_ZoneAlloc(sw->Index->perDocTmpZone, metaIDlen * 2 * sizeof(int));
memcpy((char *)newbuf,(char *)metaID,metaIDlen * sizeof(int));
metaID = newbuf;
metaIDlen *= 2;
}
/* add metaname to array of current metanames */
metaID[currentmetanames] = metaNameEntry->metaID;
/* Bump position for all metanames unless metaname in dontbumppositionOnmetatags */
if (!isDontBumpMetaName(sw->dontbumpstarttagslist, metaNameEntry->metaName))
position++;
currentmetanames++;
p = endtag;
/* If it is also a property store it until a < is found */
if ((metaNameEntry = getPropNameByName(&indexf->header, parsed_tag)))
{
if ((endtag = strchr(p, '<')))
*endtag = '\0';
p = (char *)sw_ConvHTMLEntities2ISO(sw, (unsigned char *)p);
remove_newlines(p); /** why isn't this just done for the entire doc? */
if (!addDocProperty(&thisFileEntry->docProperties, metaNameEntry, (unsigned char *)p, strlen(p), 0))
progwarn("property '%s' not added for document '%s'\n", metaNameEntry->metaName, fprop->real_path);
if (endtag)
*endtag = '<';
continue;
}
}
}
else if (lstrstr(tag, "END"))
{
/* this will close the last metaname */
if (currentmetanames)
{
currentmetanames--;
if (!currentmetanames)
metaID[0] = 1;
}
}
p = endtag;
}
/* Check for META TAG TYPE 2 */
else if ((tag[0] != '!') && lstrstr(tag, "META") && (Name = lstrstr(tag, "NAME")) && (Content = lstrstr(tag, "CONTENT")))
{
ftotalwords += parseMetaData(sw, indexf, tag, idx->filenum, structure, Name, Content, thisFileEntry, &position, fprop->real_path);
p = endtag;
} /* Check for COMMENT */
else if ((tag[0] == '!') && sw->indexComments)
{
ftotalwords += parsecomment(sw, tag, idx->filenum, structure, 1, &position);
p = endtag;
} /* Default: Continue */
else
{
structure = getstructure(tag, structure);
p = endtag;
}
}
else
p = tag; /* tag not closed: continue */
}
else
{ /* No more '<' */
newp = (char *)sw_ConvHTMLEntities2ISO(sw, (unsigned char *)p);
if ( ! currentmetanames )
currentmetanames++;
ftotalwords += indexstring(sw, newp, idx->filenum, structure, currentmetanames, metaID, &position);
p = NULL;
}
}
return ftotalwords;
}
static char *parseHtmlSummary(char *buffer, char *field, int size, SWISH * sw)
{
char *p,
*q,
*tag,
*endtag,
c = '\0';
char *summary,
*beginsum,
*endsum,
*tmp,
*tmp2,
*tmp3;
int found,
lensummary;
/* Get the summary if no metaname/field is given */
if (!field && size)
{
/* Jump title if it exists */
if ((p = lstrstr(buffer, "</title>")))
{
p += 8;
}
else
p = buffer;
/* Let us try to find <body> */
if ((q = lstrstr(p, "<body")))
{
q = strchr(q, '>');
}
else
q = p;
summary = (char *) Mem_ZoneAlloc(sw->Index->perDocTmpZone,strlen(p)+1);
strcpy(summary,p);
remove_newlines(summary);
//$$$$ Todo: remove tag and content of scripts, css, java, embeddedobjects, comments, etc
remove_tags(summary);
summary = (char *)sw_ConvHTMLEntities2ISO(sw, (unsigned char *)summary);
/* use only the required memory -save those not used */
/* 2001-03-13 rasc copy only <size> bytes of string */
if((int) strlen(summary) > size)
summary[size]='\0';
return summary;
}
for (p = buffer, summary = NULL, found = 0, beginsum = NULL, endsum = NULL; p && *p;)
{
if ((tag = strchr(p, '<')) && ((tag == p) || (*(tag - 1) != '\\')))
{ /* Look for non escaped '<' */
tag++;
for (endtag = tag;;)
if ((endtag = strchr(endtag, '>')))
{
if (*(endtag - 1) != '\\')
break;
else
endtag++;
}
else
break;
if (endtag)
{
c = *endtag;
*endtag++ = '\0';
if ((tag[0] == '!') && lstrstr(tag, "META") && (lstrstr(tag, "START") || lstrstr(tag, "END")))
{ /* Check for META TAG TYPE 1 */
if (lstrstr(tag, "START"))
{
if ((tmp = lstrstr(tag, "NAME")))
{
tmp += 4;
if (lstrstr(tmp, field))
{
beginsum = endtag;
found = 1;
}
p = endtag;
}
else
p = endtag;
}
else if (lstrstr(tag, "END"))
{
if (!found)
{
p = endtag;
}
else
{
endsum = tag - 1;
*(endtag - 1) = c;
break;
}
}
} /* Check for META TAG TYPE 2 */
else if ((tag[0] != '!') && lstrstr(tag, "META") && (tmp = lstrstr(tag, "NAME")) && (tmp2 = lstrstr(tag, "CONTENT")))
{
tmp += 4;
tmp3 = lstrstr(tmp, field);
if (tmp3 && tmp3 < tmp2)
{
tmp2 += 7;
if ((tmp = strchr(tmp2, '=')))
{
for (++tmp; isspace((int) ((unsigned char) *tmp)); tmp++);
if (*tmp == '\"')
{
beginsum = tmp + 1;
for (tmp = endtag - 1; tmp > beginsum; tmp--)
if (*tmp == '\"')
break;
if (tmp == beginsum)
endsum = endtag - 1;
else
endsum = tmp;
}
else
{
beginsum = tmp;
endsum = endtag - 1;
}
found = 1;
*(endtag - 1) = c;
break;
}
}
p = endtag;
} /* Default: Continue */
else
{
p = endtag;
}
}
else
p = NULL; /* tag not closed ->END */
if (endtag)
*(endtag - 1) = c;
}
else
{ /* No more '<' */
p = NULL;
}
}
if (found && beginsum && endsum && endsum > beginsum)
{
lensummary = endsum - beginsum;
summary = (char *)Mem_ZoneAlloc(sw->Index->perDocTmpZone, lensummary + 1);
memcpy(summary, beginsum, lensummary);
summary[lensummary] = '\0';
}
/* If field is set an no metaname is found, let us search */
/* for something like <field>bla bla </field> */
if (!summary && field)
{
summary = parsetag(sw, field, buffer, 0, CASE_SENSITIVE_OFF);
}
/* Finally check for something after title (if exists) and */
/* after <body> (if exists) */
if (!summary)
{
/* Jump title if it exists */
if ((p = lstrstr(buffer, "</title>")))
{
p += 8;
}
else
p = buffer;
/* Let us try to find <body> */
if ((q = lstrstr(p, "<body")))
{
q = strchr(q, '>');
}
else
q = p;
summary = (char *)Mem_ZoneAlloc(sw->Index->perDocTmpZone,strlen(q) + 1);
strcpy(summary,q);
}
if (summary)
{
remove_newlines(summary);
remove_tags(summary);
summary = (char *)sw_ConvHTMLEntities2ISO(sw, (unsigned char *)summary);
}
if (summary && size && ((int) strlen(summary)) > size)
summary[size] = '\0';
return summary;
}
/* Parses the Meta tag */
static int parseMetaData(SWISH * sw, IndexFILE * indexf, char *tag, int filenum, int structure, char *name, char *content, FileRec *thisFileEntry,
int *position, char *filename)
{
int metaName;
struct metaEntry *metaNameEntry;
char *temp,
*start,
*convtag;
int wordcount = 0; /* Word count */
char *parsed_tag;
/* Lookup (or add if "auto") meta name for tag */
metaNameEntry = getHTMLMeta(indexf, tag, sw, name, &parsed_tag, filename);
metaName = metaNameEntry ? metaNameEntry->metaID : 1;
temp = content + 7; /* 7 is strlen("CONTENT") */
/* Get to the " sign disreguarding other characters */
if ((temp = strchr(temp, '\"')))
{
structure |= IN_META;
start = temp + 1;
/* Jump escaped \" */
temp = strchr(start, '\"');
while (temp)
{
if (*(temp - 1) == '\\')
temp = strchr(temp + 1, '\"');
else
break;
}
if (temp)
*temp = '\0'; /* terminate CONTENT, temporarily */
/* Convert entities, if requested, and remove newlines */
convtag = (char *)sw_ConvHTMLEntities2ISO(sw, (unsigned char *)start);
remove_newlines(convtag); /** why isn't this just done for the entire doc? */
/* Index only if a metaEntry was found, or if not not ReqMetaName */
if ( sw->UndefinedMetaTags != UNDEF_META_IGNORE || metaNameEntry)
{
/* Meta tags get bummped */
/* I'm not clear this works as well as I'd like because it always bumps on a new Meta tag,
* but in order to disable this behavior the name MUST be a meta name.
* Probably better to let getHTMLMeta() return the name as a string.
*/
if (!metaNameEntry || !isDontBumpMetaName(sw->dontbumpstarttagslist, metaNameEntry->metaName))
position[0]++;
wordcount = indexstring(sw, convtag, filenum, structure, 1, &metaName, position);
if (!metaNameEntry || !isDontBumpMetaName(sw->dontbumpendtagslist, metaNameEntry->metaName))
position[0]++;
}
/* If it is a property store it */
if ((metaNameEntry = getPropNameByName(&indexf->header, parsed_tag)))
if (!addDocProperty(&thisFileEntry->docProperties, metaNameEntry, (unsigned char*)convtag, strlen(convtag), 0))
progwarn("property '%s' not added for document '%s'\n", metaNameEntry->metaName, filename);
if (temp)
*temp = '\"'; /* restore string */
}
return wordcount;
}
