// Callers delete the ToC tree, so we re-create it from prerecorded
// values (which is faster than re-creating it from html every time)
DocTocItem *ChmModel::GetTocTree()
{
DocTocItem *root = NULL, **nextChild = &root;
Vec<DocTocItem *> levels;
int idCounter = 0;
for (ChmTocTraceItem *ti = tocTrace->IterStart(); ti; ti = tocTrace->IterNext()) {
ChmTocItem *item = new ChmTocItem(ti->title, ti->pageNo, ti->url);
item->id = ++idCounter;
// append the item at the correct level
CrashIf(ti->level < 1);
if ((size_t)ti->level <= levels.Count()) {
levels.RemoveAt(ti->level, levels.Count() - ti->level);
levels.Last()->AddSibling(item);
}
else {
(*nextChild) = item;
levels.Append(item);
}
nextChild = &item->child;
}
if (root)
root->OpenSingleNode();
return root;
}
virtual const unsigned char *GetDataForUrl(const WCHAR *url, size_t *len) {
ScopedCritSec scope(&docAccess);
ScopedMem<WCHAR> plainUrl(url::GetFullPath(url));
ScopedMem<char> urlUtf8(str::conv::ToUtf8(plainUrl));
data.Append(doc->GetData(urlUtf8, len));
return data.Last();
}
static WCHAR *ExtractHtmlText(EpubDoc *doc)
{
size_t len;
const char *data = doc->GetTextData(&len);
str::Str<char> text(len / 2);
HtmlPullParser p(data, len);
HtmlToken *t;
Vec<HtmlTag> tagNesting;
while ((t = p.Next()) != NULL && !t->IsError()) {
if (t->IsText() && !tagNesting.Contains(Tag_Head) && !tagNesting.Contains(Tag_Script) && !tagNesting.Contains(Tag_Style)) {
// trim whitespace (TODO: also normalize within text?)
while (t->sLen > 0 && str::IsWs(t->s[0])) {
t->s++;
t->sLen--;
}
while (t->sLen > 0 && str::IsWs(t->s[t->sLen-1]))
t->sLen--;
if (t->sLen > 0) {
text.AppendAndFree(ResolveHtmlEntities(t->s, t->sLen));
text.Append(' ');
}
}
else if (t->IsStartTag()) {
// TODO: force-close tags similar to HtmlFormatter.cpp's AutoCloseOnOpen?
if (!IsTagSelfClosing(t->tag))
tagNesting.Append(t->tag);
}
else if (t->IsEndTag()) {
if (!IsInlineTag(t->tag) && text.Size() > 0 && text.Last() == ' ') {
text.Pop();
text.Append("\r\n");
}
// when closing a tag, if the top tag doesn't match but
// there are only potentially self-closing tags on the
// stack between the matching tag, we pop all of them
if (tagNesting.Contains(t->tag)) {
while (tagNesting.Last() != t->tag)
tagNesting.Pop();
}
if (tagNesting.Count() > 0 && tagNesting.Last() == t->tag)
tagNesting.Pop();
}
}
return str::conv::FromUtf8(text.Get());
}
bool DjVuEngineImpl::ExtractPageText(miniexp_t item, const WCHAR *lineSep, str::Str<WCHAR>& extracted, Vec<RectI>& coords)
{
miniexp_t type = miniexp_car(item);
if (!miniexp_symbolp(type))
return false;
item = miniexp_cdr(item);
if (!miniexp_numberp(miniexp_car(item))) return false;
int x0 = miniexp_to_int(miniexp_car(item)); item = miniexp_cdr(item);
if (!miniexp_numberp(miniexp_car(item))) return false;
int y0 = miniexp_to_int(miniexp_car(item)); item = miniexp_cdr(item);
if (!miniexp_numberp(miniexp_car(item))) return false;
int x1 = miniexp_to_int(miniexp_car(item)); item = miniexp_cdr(item);
if (!miniexp_numberp(miniexp_car(item))) return false;
int y1 = miniexp_to_int(miniexp_car(item)); item = miniexp_cdr(item);
RectI rect = RectI::FromXY(x0, y0, x1, y1);
miniexp_t str = miniexp_car(item);
if (miniexp_stringp(str) && !miniexp_cdr(item)) {
if (type != miniexp_symbol("char") && type != miniexp_symbol("word") ||
coords.Count() > 0 && rect.y < coords.Last().y - coords.Last().dy * 0.8) {
AppendNewline(extracted, coords, lineSep);
}
const char *content = miniexp_to_str(str);
WCHAR *value = str::conv::FromUtf8(content);
if (value) {
size_t len = str::Len(value);
// TODO: split the rectangle into individual parts per glyph
for (size_t i = 0; i < len; i++)
coords.Append(RectI(rect.x, rect.y, rect.dx, rect.dy));
extracted.AppendAndFree(value);
}
if (miniexp_symbol("word") == type) {
extracted.Append(' ');
coords.Append(RectI(rect.x + rect.dx, rect.y, 2, rect.dy));
}
item = miniexp_cdr(item);
}
while (miniexp_consp(str)) {
ExtractPageText(str, lineSep, extracted, coords);
item = miniexp_cdr(item);
str = miniexp_car(item);
}
return !item;
}
virtual void Visit(const WCHAR *name, const WCHAR *url, int level) {
int pageNo = CreatePageNoForURL(url);
ChmTocItem *item = new ChmTocItem(str::Dup(name), pageNo, str::Dup(url));
item->id = ++idCounter;
item->open = level == 1;
// append the item at the correct level
CrashIf(level < 1);
if (!*root) {
*root = item;
lastItems.Append(*root);
} else if ((size_t)level <= lastItems.Count()) {
lastItems.RemoveAt(level, lastItems.Count() - level);
lastItems.Last() = lastItems.Last()->next = item;
} else {
lastItems.Last()->child = item;
lastItems.Append(item);
}
}
// don't emit multiple spaces and don't emit spaces
// at the beginning of the line
static bool CanEmitElasticSpace(float currX, float NewLineX, float maxCurrX, Vec<DrawInstr>& currLineInstr) {
if (NewLineX == currX || 0 == currLineInstr.size())
return false;
// prevent elastic spaces from being flushed to the
// beginning of the next line
if (currX > maxCurrX)
return false;
DrawInstr& di = currLineInstr.Last();
// don't add a space if only an anchor would be in between them
if (InstrAnchor == di.type && currLineInstr.size() > 1)
di = currLineInstr.at(currLineInstr.size() - 2);
return (InstrElasticSpace != di.type) && (InstrFixedSpace != di.type);
}
// TODO: quite possibly the real logic for generating "click" events is
// more complicated
// (x, y) is in the coordinates of the root window
LRESULT EventMgr::OnLButtonUp(WPARAM keys, int x, int y, bool& wasHandled)
{
Vec<CtrlAndOffset> controls;
uint16 wantedInputMask = bit::FromBit<uint16>(Control::WantsMouseClickBit);
size_t count = CollectWindowsAt(wndRoot, x, y, wantedInputMask, &controls);
if (0 == count)
return 0;
// TODO: should this take z-order into account?
Control *c = controls.Last().c;
c->MapRootToMyPos(x, y);
NotifyClicked(c, x, y);
NotifyNamedEventClicked(c, x, y);
return 0;
}
// adapted from DisplayModel::NextZoomStep
float ChmModel::GetNextZoomStep(float towardsLevel) const
{
float currZoom = GetZoomVirtual();
if (gGlobalPrefs->zoomIncrement > 0) {
if (currZoom < towardsLevel)
return min(currZoom * (gGlobalPrefs->zoomIncrement / 100 + 1), towardsLevel);
if (currZoom > towardsLevel)
return max(currZoom / (gGlobalPrefs->zoomIncrement / 100 + 1), towardsLevel);
return currZoom;
}
Vec<float> *zoomLevels = gGlobalPrefs->zoomLevels;
CrashIf(zoomLevels->Count() != 0 && (zoomLevels->At(0) < ZOOM_MIN || zoomLevels->Last() > ZOOM_MAX));
CrashIf(zoomLevels->Count() != 0 && zoomLevels->At(0) > zoomLevels->Last());
const float FUZZ = 0.01f;
float newZoom = towardsLevel;
if (currZoom < towardsLevel) {
for (size_t i = 0; i < zoomLevels->Count(); i++) {
if (zoomLevels->At(i) - FUZZ > currZoom) {
newZoom = zoomLevels->At(i);
break;
}
}
}
else if (currZoom > towardsLevel) {
for (size_t i = zoomLevels->Count(); i > 0; i--) {
if (zoomLevels->At(i - 1) + FUZZ < currZoom) {
newZoom = zoomLevels->At(i - 1);
break;
}
}
}
return newZoom;
}
// TODO: optimize by getting both mouse over and mouse move windows in one call
// x, y is a position in the root window
LRESULT EventMgr::OnMouseMove(WPARAM keys, int x, int y, bool& wasHandled)
{
Vec<CtrlAndOffset> windows;
Control *c;
uint16 wantedInputMask = bit::FromBit<uint16>(Control::WantsMouseOverBit);
size_t count = CollectWindowsAt(wndRoot, x, y, wantedInputMask, &windows);
if (0 == count) {
if (currOver) {
currOver->SetIsMouseOver(false);
currOver->NotifyMouseLeave();
currOver = nullptr;
}
} else {
// TODO: should this take z-order into account ?
c = windows.Last().c;
if (c != currOver) {
if (currOver) {
currOver->SetIsMouseOver(false);
currOver->NotifyMouseLeave();
}
currOver = c;
currOver->SetIsMouseOver(true);
currOver->NotifyMouseEnter();
}
}
wantedInputMask = bit::FromBit<uint16>(Control::WantsMouseMoveBit);
count = CollectWindowsAt(wndRoot, x, y, wantedInputMask, &windows);
if (0 == count)
return 0;
c = windows.Last().c;
c->MapRootToMyPos(x, y);
c->NotifyMouseMove(x, y);
return 0;
}
ImageData *GetImageData(const char *id, const char *pagePath) {
ScopedMem<char> url(NormalizeURL(id, pagePath));
str::UrlDecodeInPlace(url);
for (size_t i = 0; i < images.Count(); i++) {
if (str::Eq(images.At(i).id, url))
return &images.At(i).base;
}
ImageData2 data = { 0 };
data.base.data = (char *)doc->GetData(url, &data.base.len);
if (!data.base.data)
return NULL;
data.id = url.StealData();
images.Append(data);
return &images.Last().base;
}
static void VecTest()
{
Vec<int> ints;
assert(ints.Count() == 0);
ints.Append(1);
ints.Push(2);
ints.InsertAt(0, -1);
assert(ints.Count() == 3);
assert(ints.At(0) == -1 && ints.At(1) == 1 && ints.At(2) == 2);
assert(ints.At(0) == -1 && ints.Last() == 2);
int last = ints.Pop();
assert(last == 2);
assert(ints.Count() == 2);
ints.Push(3);
ints.RemoveAt(0);
assert(ints.Count() == 2);
assert(ints.At(0) == 1 && ints.At(1) == 3);
ints.Reset();
assert(ints.Count() == 0);
for (int i = 0; i < 1000; i++) {
ints.Push(i);
}
assert(ints.Count() == 1000 && ints.At(500) == 500);
ints.Remove(500);
assert(ints.Count() == 999 && ints.At(500) == 501);
last = ints.Pop();
assert(last == 999);
ints.Append(last);
assert(ints.AtPtr(501) == &ints.At(501));
{
Vec<int> ints2(ints);
assert(ints2.Count() == 999);
assert(ints.LendData() != ints2.LendData());
ints.Remove(600);
assert(ints.Count() < ints2.Count());
ints2 = ints;
assert(ints2.Count() == 998);
}
{
char buf[2] = {'a', '\0'};
str::Str<char> v(0);
for (int i = 0; i < 7; i++) {
v.Append(buf, 1);
buf[0] = buf[0] + 1;
}
char *s = v.LendData();
assert(str::Eq("abcdefg", s));
assert(7 == v.Count());
v.Set("helo");
assert(4 == v.Count());
assert(str::Eq("helo", v.LendData()));
}
{
str::Str<char> v(128);
v.Append("boo", 3);
assert(str::Eq("boo", v.LendData()));
assert(v.Count() == 3);
v.Append("fop");
assert(str::Eq("boofop", v.LendData()));
assert(v.Count() == 6);
v.RemoveAt(2, 3);
assert(v.Count() == 3);
assert(str::Eq("bop", v.LendData()));
v.Append('a');
assert(v.Count() == 4);
assert(str::Eq("bopa", v.LendData()));
char *s = v.StealData();
assert(str::Eq("bopa", s));
free(s);
assert(v.Count() == 0);
}
{
str::Str<char> v(0);
for (int i = 0; i < 32; i++) {
assert(v.Count() == i * 6);
v.Append("lambd", 5);
if (i % 2 == 0)
v.Append('a');
else
v.Push('a');
}
for (int i=1; i<=16; i++) {
v.RemoveAt((16 - i) * 6, 6);
assert(v.Count() == (32 - i) * 6);
}
v.RemoveAt(0, 6 * 15);
assert(v.Count() == 6);
char *s = v.LendData();
assert(str::Eq(s, "lambda"));
s = v.StealData();
assert(str::Eq(s, "lambda"));
free(s);
assert(v.Count() == 0);
v.Append("lambda");
assert(str::Eq(v.LendData(), "lambda"));
char c = v.Pop();
assert(c == 'a');
assert(str::Eq(v.LendData(), "lambd"));
}
VecTestAppendFmt();
{
Vec<PointI *> v;
srand((unsigned int)time(NULL));
for (int i = 0; i < 128; i++) {
v.Append(new PointI(i, i));
size_t pos = rand() % v.Count();
v.InsertAt(pos, new PointI(i, i));
}
assert(v.Count() == 128 * 2);
size_t idx = 0;
for (PointI **p = v.IterStart(); p; p = v.IterNext()) {
assert(idx == v.IterIdx());
++idx;
}
while (v.Count() > 64) {
size_t pos = rand() % v.Count();
PointI *f = v.At(pos);
v.Remove(f);
delete f;
}
DeleteVecMembers(v);
}
{
Vec<int> v;
v.Append(2);
for (int i = 0; i < 500; i++)
v.Append(4);
v.At(250) = 5;
v.Reverse();
assert(v.Count() == 501 && v.At(0) == 4 && v.At(249) == v.At(251) && v.At(250) == 5 && v.At(500) == 2);
v.Remove(4);
v.Reverse();
assert(v.Count() == 500 && v.At(0) == 2 && v.At(249) == v.At(251) && v.At(250) == 5 && v.At(499) == 4);
}
}
// see http://itexmac.sourceforge.net/pdfsync.html for the specification
int Pdfsync::RebuildIndex()
{
size_t len;
ScopedMem<char> data(file::ReadAll(syncfilepath, &len));
if (!data)
return PDFSYNCERR_SYNCFILE_CANNOT_BE_OPENED;
// convert the file data into a list of zero-terminated strings
str::TransChars(data, "\r\n", "\0\0");
// parse preamble (jobname and version marker)
char *line = data;
char *dataEnd = data + len;
// replace star by spaces (TeX uses stars instead of spaces in filenames)
str::TransChars(line, "*/", " \\");
ScopedMem<WCHAR> jobName(str::conv::FromAnsi(line));
jobName.Set(str::Join(jobName, L".tex"));
jobName.Set(PrependDir(jobName));
line = Advance0Line(line, dataEnd);
UINT versionNumber = 0;
if (!line || !str::Parse(line, "version %u", &versionNumber) || versionNumber != 1)
return PDFSYNCERR_SYNCFILE_CANNOT_BE_OPENED;
// reset synchronizer database
srcfiles.Reset();
lines.Reset();
points.Reset();
fileIndex.Reset();
sheetIndex.Reset();
Vec<size_t> filestack;
UINT page = 1;
sheetIndex.Append(0);
// add the initial tex file to the source file stack
filestack.Push(srcfiles.Count());
srcfiles.Append(jobName.StealData());
PdfsyncFileIndex findex = { 0 };
fileIndex.Append(findex);
PdfsyncLine psline;
PdfsyncPoint pspoint;
// parse data
UINT maxPageNo = engine->PageCount();
while ((line = Advance0Line(line, dataEnd)) != NULL) {
if (!line)
break;
switch (*line) {
case 'l':
psline.file = filestack.Last();
if (str::Parse(line, "l %u %u %u", &psline.record, &psline.line, &psline.column))
lines.Append(psline);
else if (str::Parse(line, "l %u %u", &psline.record, &psline.line)) {
psline.column = 0;
lines.Append(psline);
}
// else dbg("Bad 'l' line in the pdfsync file");
break;
case 's':
if (str::Parse(line, "s %u", &page))
sheetIndex.Append(points.Count());
// else dbg("Bad 's' line in the pdfsync file");
// if (0 == page || page > maxPageNo)
// dbg("'s' line with invalid page number in the pdfsync file");
break;
case 'p':
pspoint.page = page;
if (0 == page || page > maxPageNo)
/* ignore point for invalid page number */;
else if (str::Parse(line, "p %u %u %u", &pspoint.record, &pspoint.x, &pspoint.y))
points.Append(pspoint);
else if (str::Parse(line, "p* %u %u %u", &pspoint.record, &pspoint.x, &pspoint.y))
points.Append(pspoint);
// else dbg("Bad 'p' line in the pdfsync file");
break;
case '(':
{
ScopedMem<WCHAR> filename(str::conv::FromAnsi(line + 1));
// if the filename contains quotes then remove them
// TODO: this should never happen!?
if (filename[0] == '"' && filename[str::Len(filename) - 1] == '"')
filename.Set(str::DupN(filename + 1, str::Len(filename) - 2));
// undecorate the filepath: replace * by space and / by \
str::TransChars(filename, L"*/", L" \\");
// if the file name extension is not specified then add the suffix '.tex'
if (str::IsEmpty(path::GetExt(filename)))
filename.Set(str::Join(filename, L".tex"));
// ensure that the path is absolute
if (PathIsRelative(filename))
filename.Set(PrependDir(filename));
filestack.Push(srcfiles.Count());
srcfiles.Append(filename.StealData());
findex.start = findex.end = lines.Count();
fileIndex.Append(findex);
}
break;
case ')':
if (filestack.Count() > 1)
fileIndex.At(filestack.Pop()).end = lines.Count();
// else dbg("Unbalanced ')' line in the pdfsync file");
break;
default:
// dbg("Ignoring invalid pdfsync line starting with '%c'", *line);
break;
}
}
fileIndex.At(0).end = lines.Count();
assert(filestack.Count() == 1);
return Synchronizer::RebuildIndex();
}
WCHAR *EbookEngine::ExtractPageText(int pageNo, WCHAR *lineSep, RectI **coords_out, RenderTarget target)
{
ScopedCritSec scope(&pagesAccess);
str::Str<WCHAR> content;
Vec<RectI> coords;
bool insertSpace = false;
Vec<DrawInstr> *pageInstrs = GetHtmlPage(pageNo);
for (DrawInstr *i = pageInstrs->IterStart(); i; i = pageInstrs->IterNext()) {
RectI bbox = GetInstrBbox(i, pageBorder);
switch (i->type) {
case InstrString:
if (coords.Count() > 0 && bbox.x < coords.Last().BR().x) {
content.Append(lineSep);
coords.AppendBlanks(str::Len(lineSep));
CrashIf(*lineSep && !coords.Last().IsEmpty());
}
else if (insertSpace && coords.Count() > 0) {
int swidth = bbox.x - coords.Last().BR().x;
if (swidth > 0) {
content.Append(' ');
coords.Append(RectI(bbox.x - swidth, bbox.y, swidth, bbox.dy));
}
}
insertSpace = false;
{
ScopedMem<WCHAR> s(str::conv::FromHtmlUtf8(i->str.s, i->str.len));
content.Append(s);
size_t len = str::Len(s);
double cwidth = 1.0 * bbox.dx / len;
for (size_t k = 0; k < len; k++)
coords.Append(RectI((int)(bbox.x + k * cwidth), bbox.y, (int)cwidth, bbox.dy));
}
break;
case InstrRtlString:
if (coords.Count() > 0 && bbox.BR().x > coords.Last().x) {
content.Append(lineSep);
coords.AppendBlanks(str::Len(lineSep));
CrashIf(*lineSep && !coords.Last().IsEmpty());
}
else if (insertSpace && coords.Count() > 0) {
int swidth = coords.Last().x - bbox.BR().x;
if (swidth > 0) {
content.Append(' ');
coords.Append(RectI(bbox.BR().x, bbox.y, swidth, bbox.dy));
}
}
insertSpace = false;
{
ScopedMem<WCHAR> s(str::conv::FromHtmlUtf8(i->str.s, i->str.len));
content.Append(s);
size_t len = str::Len(s);
double cwidth = 1.0 * bbox.dx / len;
for (size_t k = 0; k < len; k++)
coords.Append(RectI((int)(bbox.x + (len - k - 1) * cwidth), bbox.y, (int)cwidth, bbox.dy));
}
break;
case InstrElasticSpace:
case InstrFixedSpace:
insertSpace = true;
break;
}
}
if (coords_out) {
CrashIf(coords.Count() != content.Count());
*coords_out = new RectI[coords.Count()];
memcpy(*coords_out, coords.LendData(), coords.Count() * sizeof(RectI));
}
return content.StealData();
}
