lyx_mirror/src/CutAndPaste.cpp
Daniel Ramoeller f5e66ba3c2 Improved layout adoption on paste
The layout from the paste content is adopted only if
- the paragraph is empty (ie. empty after the selection has been
  erased) and Standard or Plain Layout, or
- empty and multiple lines are pasted, or
- multiple lines are pasted at the beginning of a paragraph

As for what other applications do:
- MS Word decides the adoption of a layout depending on whether the end of line
character is copied
- Libre Writer adopts the paste content layout only if the paragraph is empty
(otherwise it makes *all* pasted paragraph the target layout)

In LyX, as in Libre Writer, one cannot copy the end of line character by
selecting a single line. However, the Libre Writer solution is problematic
because one always has to make sure that a paragraph is in Standard layout if
one wants to paste several lines with their source layout. The implemented
behaviour differs from Libre Writer in that the source format is kept when end
of line characters are pasted, i.e. several lines are pasted, and hence does
what MS Word does in that situation.

Fix for bug #11023: Copy and paste from one list into another often leads to
undesired result.
2022-07-25 22:56:00 +02:00

1579 lines
46 KiB
C++

/**
* \file CutAndPaste.cpp
* This file is part of LyX, the document processor.
* Licence details can be found in the file COPYING.
*
* \author Jürgen Vigna
* \author Lars Gullik Bjønnes
* \author Alfredo Braunstein
* \author Michael Gerz
*
* Full author contact details are available in file CREDITS.
*/
#include <config.h>
#include "CutAndPaste.h"
#include "Author.h"
#include "BranchList.h"
#include "Buffer.h"
#include "BufferList.h"
#include "BufferParams.h"
#include "BufferView.h"
#include "Changes.h"
#include "Cursor.h"
#include "Encoding.h"
#include "ErrorList.h"
#include "FuncCode.h"
#include "FuncRequest.h"
#include "InsetIterator.h"
#include "Language.h"
#include "LyX.h"
#include "LyXRC.h"
#include "Text.h"
#include "Paragraph.h"
#include "ParagraphParameters.h"
#include "ParIterator.h"
#include "TextClass.h"
#include "insets/InsetBibitem.h"
#include "insets/InsetBranch.h"
#include "insets/InsetCitation.h"
#include "insets/InsetCommand.h"
#include "insets/InsetGraphics.h"
#include "insets/InsetGraphicsParams.h"
#include "insets/InsetInclude.h"
#include "insets/InsetLabel.h"
#include "insets/InsetTabular.h"
#include "mathed/MathData.h"
#include "mathed/InsetMath.h"
#include "mathed/InsetMathHull.h"
#include "mathed/InsetMathRef.h"
#include "mathed/MathSupport.h"
#include "support/debug.h"
#include "support/docstream.h"
#include "support/gettext.h"
#include "support/lassert.h"
#include "support/limited_stack.h"
#include "support/lstrings.h"
#include "support/TempFile.h"
#include "support/unique_ptr.h"
#include "frontends/alert.h"
#include "frontends/Clipboard.h"
#include <string>
#include <tuple>
using namespace std;
using namespace lyx::support;
using lyx::frontend::Clipboard;
namespace lyx {
namespace {
typedef pair<pit_type, int> PitPosPair;
typedef pair<DocumentClassConstPtr, AuthorList > DocInfoPair;
typedef limited_stack<pair<ParagraphList, DocInfoPair > > CutStack;
CutStack theCuts(10);
// persistent selection, cleared until the next selection
CutStack selectionBuffer(1);
// temporary scratch area
CutStack tempCut(1);
// store whether the tabular stack is newer than the normal copy stack
// FIXME: this is a workaround for bug 1919. Should be removed for 1.5,
// when we (hopefully) have a one-for-all paste mechanism.
bool dirty_tabular_stack_ = false;
bool checkPastePossible(int index)
{
return size_t(index) < theCuts.size() && !theCuts[index].first.empty();
}
struct PasteReturnValue {
PasteReturnValue(pit_type r_pit, pos_type r_pos, bool r_nu) :
pit(r_pit), pos(r_pos), needupdate(r_nu)
{}
pit_type pit;
pos_type pos;
bool needupdate;
};
PasteReturnValue
pasteSelectionHelper(DocIterator const & cur, ParagraphList const & parlist,
DocumentClassConstPtr oldDocClass, cap::BranchAction branchAction,
ErrorList & errorlist)
{
Buffer const & buffer = *cur.buffer();
pit_type pit = cur.pit();
pos_type pos = cur.pos();
bool need_update = false;
if (parlist.empty())
return PasteReturnValue(pit, pos, need_update);
// Check whether we paste into an inset that does not
// produce output (needed for label duplicate check)
bool in_active_inset = cur.paragraph().inInset().producesOutput();
if (in_active_inset) {
for (size_type sl = 0 ; sl < cur.depth() ; ++sl) {
Paragraph const & outer_par = cur[sl].paragraph();
if (!outer_par.inInset().producesOutput()) {
in_active_inset = false;
break;
}
}
}
InsetText * target_inset = cur.inset().asInsetText();
if (!target_inset) {
InsetTabular * it = cur.inset().asInsetTabular();
target_inset = it ? it->cell(cur.idx())->asInsetText() : nullptr;
}
LASSERT(target_inset, return PasteReturnValue(pit, pos, need_update));
ParagraphList & pars = target_inset->paragraphs();
LASSERT(pos <= pars[pit].size(),
return PasteReturnValue(pit, pos, need_update));
// Make a copy of the CaP paragraphs.
ParagraphList insertion = parlist;
// Now remove all out of the pars which is NOT allowed in the
// new environment and set also another font if that is required.
// Merge paragraphs that are to be pasted into a text inset
// that does not allow multiple pars.
InsetText * inset_text = target_inset->asInsetText();
if (inset_text && !inset_text->allowMultiPar()) {
while (insertion.size() > 1)
mergeParagraph(buffer.params(), insertion, 0);
}
// Convert newline to paragraph break in ParbreakIsNewline
if (target_inset->getLayout().parbreakIsNewline()
|| pars[pit].layout().parbreak_is_newline) {
for (size_t i = 0; i != insertion.size(); ++i) {
for (pos_type j = 0; j != insertion[i].size(); ++j) {
if (insertion[i].isNewline(j)) {
// do not track deletion of newline
insertion[i].eraseChar(j, false);
insertion[i].setInsetOwner(target_inset);
breakParagraphConservative(
buffer.params(),
insertion, i, j);
break;
}
}
}
}
// Prevent to paste uncodable characters in verbatim and ERT.
// The encoding is inherited from the context here.
docstring uncodable_content;
if (target_inset->getLayout().isPassThru() && cur.getEncoding()) {
odocstringstream res;
Encoding const * e = cur.getEncoding();
for (size_t i = 0; i != insertion.size(); ++i) {
pos_type end = insertion[i].size();
for (pos_type j = 0; j != end; ++j) {
// skip insets
if (insertion[i].isInset(j))
continue;
char_type const c = insertion[i].getChar(j);
if (!e->encodable(c)) {
// do not track deletion
res.put(c);
insertion[i].eraseChar(j, false);
--end;
--j;
}
}
}
docstring const uncodable = res.str();
if (!uncodable.empty()) {
if (uncodable.size() == 1)
uncodable_content = bformat(_("The character \"%1$s\" is uncodable in this verbatim context "
"and thus has not been pasted."),
uncodable);
else
uncodable_content = bformat(_("The characters \"%1$s\" are uncodable in this verbatim context "
"and thus have not been pasted."),
uncodable);
}
}
// set the paragraphs to plain layout if necessary
DocumentClassConstPtr newDocClass = buffer.params().documentClassPtr();
Layout const & plainLayout = newDocClass->plainLayout();
Layout const & defaultLayout = newDocClass->defaultLayout();
if (cur.inset().usePlainLayout()) {
bool forcePlainLayout = target_inset->forcePlainLayout();
for (auto & par : insertion) {
Layout const & parLayout = par.layout();
if (forcePlainLayout || parLayout == defaultLayout)
par.setLayout(plainLayout);
}
} else {
// check if we need to reset from plain layout
for (auto & par : insertion) {
Layout const & parLayout = par.layout();
if (parLayout == plainLayout)
par.setLayout(defaultLayout);
}
}
InsetText in(cur.buffer());
// Make sure there is no class difference.
in.paragraphs().clear();
// This works without copying any paragraph data because we have
// a specialized swap method for ParagraphList. This is important
// since we store pointers to insets at some places and we don't
// want to invalidate them.
insertion.swap(in.paragraphs());
cap::switchBetweenClasses(oldDocClass, newDocClass, in, errorlist);
// Do this here since switchBetweenClasses clears the errorlist
if (!uncodable_content.empty())
errorlist.push_back(ErrorItem(_("Uncodable content"), uncodable_content));
insertion.swap(in.paragraphs());
ParagraphList::iterator tmpbuf = insertion.begin();
int depth_delta = pars[pit].params().depth() - tmpbuf->params().depth();
depth_type max_depth = pars[pit].getMaxDepthAfter();
for (; tmpbuf != insertion.end(); ++tmpbuf) {
// If we have a negative jump so that the depth would
// go below 0 depth then we have to redo the delta to
// this new max depth level so that subsequent
// paragraphs are aligned correctly to this paragraph
// at level 0.
if (int(tmpbuf->params().depth()) + depth_delta < 0)
depth_delta = 0;
// Set the right depth so that we are not too deep or shallow.
tmpbuf->params().depth(tmpbuf->params().depth() + depth_delta);
if (tmpbuf->params().depth() > max_depth)
tmpbuf->params().depth(max_depth);
// Set max_depth for the next paragraph
max_depth = tmpbuf->getMaxDepthAfter();
// Set the inset owner of this paragraph.
tmpbuf->setInsetOwner(target_inset);
for (pos_type i = 0; i < tmpbuf->size(); ++i) {
// do not track deletion of invalid insets
if (Inset * inset = tmpbuf->getInset(i))
if (!target_inset->insetAllowed(inset->lyxCode()))
tmpbuf->eraseChar(i--, false);
}
if (lyxrc.ct_markup_copied) {
// Only change to inserted if ct is active,
// otherwise leave markup as is
if (buffer.params().track_changes)
tmpbuf->setChange(Change(Change::INSERTED));
} else
// Resolve all markup to inserted or unchanged
tmpbuf->setChange(Change(buffer.params().track_changes ?
Change::INSERTED : Change::UNCHANGED));
}
bool const target_empty = pars[pit].empty();
// Use the paste content's layout, if...
bool const paste_layout =
// if target paragraph is empty
(target_empty
// ... and its layout is default
&& (pars[pit].layout() == defaultLayout
// ... or plain
|| pars[pit].layout() == plainLayout
// ... or the paste content spans several paragraphs
|| insertion.size() > 1))
// or if pasting is done at the beginning of paragraph
|| (pos == 0
// and the paste content spans several paragraphs
&& insertion.size() > 1);
if (!paste_layout)
// Give the first paragraph to insert the same layout as the
// target paragraph.
insertion.begin()->makeSameLayout(pars[pit]);
// Prepare the paragraphs and insets for insertion.
insertion.swap(in.paragraphs());
InsetIterator const i_end = end(in);
for (InsetIterator it = begin(in); it != i_end; ++it) {
// Even though this will also be done later, it has to be done here
// since some inset might try to access the buffer() member.
it->setBuffer(const_cast<Buffer &>(buffer));
switch (it->lyxCode()) {
case MATH_HULL_CODE: {
// check for equation labels and resolve duplicates
InsetMathHull * ins = it->asInsetMath()->asHullInset();
std::vector<InsetLabel *> labels = ins->getLabels();
for (size_t i = 0; i != labels.size(); ++i) {
if (!labels[i])
continue;
InsetLabel * lab = labels[i];
docstring const oldname = lab->getParam("name");
lab->updateLabel(oldname, in_active_inset);
// We need to update the buffer reference cache.
need_update = true;
docstring const newname = lab->getParam("name");
if (oldname == newname)
continue;
// adapt the references
for (InsetIterator itt = begin(in);
itt != i_end; ++itt) {
if (itt->lyxCode() == REF_CODE) {
InsetCommand * ref = itt->asInsetCommand();
if (ref->getParam("reference") == oldname)
ref->setParam("reference", newname);
} else if (itt->lyxCode() == MATH_REF_CODE) {
InsetMathRef * mi = itt->asInsetMath()->asRefInset();
// this is necessary to prevent an uninitialized
// buffer when the RefInset is in a MathBox.
// FIXME audit setBuffer calls
mi->setBuffer(const_cast<Buffer &>(buffer));
if (mi->getTarget() == oldname)
mi->changeTarget(newname);
}
}
}
break;
}
case LABEL_CODE: {
// check for duplicates
InsetLabel & lab = static_cast<InsetLabel &>(*it);
docstring const oldname = lab.getParam("name");
lab.updateLabel(oldname, in_active_inset);
// We need to update the buffer reference cache.
need_update = true;
docstring const newname = lab.getParam("name");
if (oldname == newname)
break;
// adapt the references
for (InsetIterator itt = begin(in); itt != i_end; ++itt) {
if (itt->lyxCode() == REF_CODE) {
InsetCommand & ref = static_cast<InsetCommand &>(*itt);
if (ref.getParam("reference") == oldname)
ref.setParam("reference", newname);
} else if (itt->lyxCode() == MATH_REF_CODE) {
InsetMathRef * mi = itt->asInsetMath()->asRefInset();
// this is necessary to prevent an uninitialized
// buffer when the RefInset is in a MathBox.
// FIXME audit setBuffer calls
mi->setBuffer(const_cast<Buffer &>(buffer));
if (mi->getTarget() == oldname)
mi->changeTarget(newname);
}
}
break;
}
case INCLUDE_CODE: {
InsetInclude & inc = static_cast<InsetInclude &>(*it);
inc.updateCommand();
// We need to update the list of included files.
need_update = true;
break;
}
case CITE_CODE: {
InsetCitation & cit = static_cast<InsetCitation &>(*it);
// This actually only needs to be done if the cite engine
// differs, but we do it in general.
cit.redoLabel();
// We need to update the list of citations.
need_update = true;
break;
}
case BIBITEM_CODE: {
// check for duplicates
InsetBibitem & bib = static_cast<InsetBibitem &>(*it);
docstring const oldkey = bib.getParam("key");
bib.updateCommand(oldkey, false);
// We need to update the buffer reference cache.
need_update = true;
docstring const newkey = bib.getParam("key");
if (oldkey == newkey)
break;
// adapt the references
for (InsetIterator itt = begin(in);
itt != i_end; ++itt) {
if (itt->lyxCode() == CITE_CODE) {
InsetCommand * ref = itt->asInsetCommand();
if (ref->getParam("key") == oldkey)
ref->setParam("key", newkey);
}
}
break;
}
case BRANCH_CODE: {
// check if branch is known to target buffer
// or its master
InsetBranch & br = static_cast<InsetBranch &>(*it);
docstring const name = br.branch();
if (name.empty())
break;
bool const is_child = (&buffer != buffer.masterBuffer());
BranchList branchlist = buffer.params().branchlist();
if ((!is_child && branchlist.find(name))
|| (is_child && (branchlist.find(name)
|| buffer.masterBuffer()->params().branchlist().find(name))))
break;
switch(branchAction) {
case cap::BRANCH_ADD: {
// This is for a temporary buffer, so simply create the branch.
// Must not use lyx::dispatch(), since tmpbuffer has no view.
DispatchResult dr;
const_cast<Buffer&>(buffer).dispatch(FuncRequest(LFUN_BRANCH_ADD, name), dr);
break;
}
case cap::BRANCH_ASK: {
docstring text = bformat(
_("The pasted branch \"%1$s\" is undefined.\n"
"Do you want to add it to the document's branch list?"),
name);
if (frontend::Alert::prompt(_("Unknown branch"),
text, 0, 1, _("&Add"), _("&Don't Add")) != 0)
break;
lyx::dispatch(FuncRequest(LFUN_BRANCH_ADD, name));
break;
}
case cap::BRANCH_IGNORE:
break;
}
// We need to update the list of branches.
need_update = true;
break;
}
default:
break; // nothing
}
}
insertion.swap(in.paragraphs());
// Split the paragraph for inserting the buf if necessary.
if (!target_empty)
breakParagraphConservative(buffer.params(), pars, pit, pos);
// If multiple paragraphs are inserted before the target paragraph,
// the cursor is now in a new paragraph before it,
// so use layout from paste content in that case.
if (pos == 0 && insertion.size() > 1)
pars[pit].makeSameLayout(* insertion.begin());
// Paste it!
if (target_empty) {
pars.insert(pars.iterator_at(pit),
insertion.begin(), insertion.end());
// merge the empty par with the last par of the insertion
mergeParagraph(buffer.params(), pars,
pit + insertion.size() - 1);
} else {
pars.insert(pars.iterator_at(pit + 1),
insertion.begin(), insertion.end());
// merge the first par of the insertion with the current par
mergeParagraph(buffer.params(), pars, pit);
}
// Store the new cursor position.
pit_type last_paste = pit + insertion.size() - 1;
pit_type startpit = pit;
pit = last_paste;
pos = pars[last_paste].size();
// FIXME Should we do it here, or should we let updateBuffer() do it?
// Set paragraph buffers. It's important to do this right away
// before something calls Inset::buffer() and causes a crash.
for (pit_type p = startpit; p <= pit; ++p)
pars[p].setInsetBuffers(const_cast<Buffer &>(buffer));
// Join (conditionally) last pasted paragraph with next one, i.e.,
// the tail of the spliced document paragraph
if (!target_empty && last_paste + 1 != pit_type(pars.size())) {
if (pars[last_paste + 1].hasSameLayout(pars[last_paste])) {
mergeParagraph(buffer.params(), pars, last_paste);
} else if (pars[last_paste + 1].empty()) {
pars[last_paste + 1].makeSameLayout(pars[last_paste]);
mergeParagraph(buffer.params(), pars, last_paste);
} else if (pars[last_paste].empty()) {
pars[last_paste].makeSameLayout(pars[last_paste + 1]);
mergeParagraph(buffer.params(), pars, last_paste);
} else {
pars[last_paste + 1].stripLeadingSpaces(buffer.params().track_changes);
++last_paste;
}
}
return PasteReturnValue(pit, pos, need_update);
}
PitPosPair eraseSelectionHelper(BufferParams const & params,
ParagraphList & pars,
pit_type startpit, pit_type endpit,
int startpos, int endpos)
{
// Start of selection is really invalid.
if (startpit == pit_type(pars.size()) ||
(startpos > pars[startpit].size()))
return PitPosPair(endpit, endpos);
// Start and end is inside same paragraph
if (endpit == pit_type(pars.size()) || startpit == endpit) {
endpos -= pars[startpit].eraseChars(startpos, endpos, params.track_changes);
return PitPosPair(endpit, endpos);
}
for (pit_type pit = startpit; pit != endpit + 1;) {
pos_type const left = (pit == startpit ? startpos : 0);
pos_type right = (pit == endpit ? endpos : pars[pit].size() + 1);
bool const merge = pars[pit].isMergedOnEndOfParDeletion(params.track_changes);
// Logically erase only, including the end-of-paragraph character
pars[pit].eraseChars(left, right, params.track_changes);
// Separate handling of paragraph break:
if (merge && pit != endpit &&
(pit + 1 != endpit
|| pars[pit].hasSameLayout(pars[endpit])
|| pars[endpit].size() == endpos)) {
if (pit + 1 == endpit)
endpos += pars[pit].size();
mergeParagraph(params, pars, pit);
--endpit;
} else
++pit;
}
// Ensure legal cursor pos:
endpit = startpit;
endpos = startpos;
return PitPosPair(endpit, endpos);
}
Buffer * copyToTempBuffer(ParagraphList const & paragraphs, DocumentClassConstPtr docclass)
{
// This used to need to be static to avoid a memory leak. It no longer needs
// to be so, but the alternative is to construct a new one of these (with a
// new temporary directory, etc) every time, and then to destroy it. So maybe
// it's worth just keeping this one around.
static TempFile tempfile("clipboard.internal");
tempfile.setAutoRemove(false);
// The initialization of staticbuffer is thread-safe. Using a lambda
// guarantees that the properties are set only once.
static Buffer * staticbuffer = [&](){
Buffer * b =
theBufferList().newInternalBuffer(tempfile.name().absFileName());
b->setUnnamed(true);
b->inset().setBuffer(*b);
//initialize staticbuffer with b
return b;
}();
// Use a clone for the complicated stuff so that we do not need to clean
// up in order to avoid a crash.
Buffer * buffer = staticbuffer->cloneBufferOnly();
LASSERT(buffer, return nullptr);
// This needs doing every time.
// Since setDocumentClass() causes deletion of the old document class
// we need to reset all layout pointers in paragraphs (otherwise they
// would be dangling).
ParIterator const end = buffer->par_iterator_end();
for (ParIterator it = buffer->par_iterator_begin(); it != end; ++it) {
docstring const name = it->layout().name();
if (docclass->hasLayout(name))
it->setLayout((*docclass)[name]);
else
it->setPlainOrDefaultLayout(*docclass);
}
buffer->params().setDocumentClass(docclass);
// we will use pasteSelectionHelper to copy the paragraphs into the
// temporary Buffer, since it does a lot of things to fix them up.
DocIterator dit = doc_iterator_begin(buffer, &buffer->inset());
ErrorList el;
pasteSelectionHelper(dit, paragraphs, docclass, cap::BRANCH_ADD, el);
return buffer;
}
void putClipboard(ParagraphList const & paragraphs,
DocInfoPair docinfo, docstring const & plaintext,
BufferParams const & bp)
{
Buffer * buffer = copyToTempBuffer(paragraphs, docinfo.first);
if (!buffer) // already asserted in copyToTempBuffer()
return;
// We don't want to produce images that are not used. Therefore,
// output formulas as MathML. Even if this is not understood by all
// applications, the number that can parse it should go up in the future.
buffer->params().html_math_output = BufferParams::MathML;
// Copy authors to the params. We need those pointers.
for (Author const & a : bp.authors())
buffer->params().authors().record(a);
// Make sure MarkAsExporting is deleted before buffer is
{
// The Buffer is being used to export. This is necessary so that the
// updateMacros call will record the needed information.
MarkAsExporting mex(buffer);
buffer->updateBuffer(Buffer::UpdateMaster, OutputUpdate);
buffer->updateMacros();
buffer->updateMacroInstances(OutputUpdate);
// LyX's own format
string lyx;
ostringstream oslyx;
if (buffer->write(oslyx))
lyx = oslyx.str();
// XHTML format
odocstringstream oshtml;
OutputParams runparams(encodings.fromLyXName("utf8"));
// We do not need to produce images, etc.
runparams.dryrun = true;
// We are not interested in errors (bug 8866)
runparams.silent = true;
buffer->writeLyXHTMLSource(oshtml, runparams, Buffer::FullSource);
theClipboard().put(lyx, oshtml.str(), plaintext);
}
// Save that memory
delete buffer;
}
/// return true if the whole ParagraphList is deleted
static bool isFullyDeleted(ParagraphList const & pars)
{
pit_type const pars_size = static_cast<pit_type>(pars.size());
// check all paragraphs
for (pit_type pit = 0; pit < pars_size; ++pit) {
if (!pars[pit].empty()) // prevent assertion failure
if (!pars[pit].isDeleted(0, pars[pit].size()))
return false;
}
return true;
}
void copySelectionHelper(Buffer const & buf, Text const & text,
pit_type startpit, pit_type endpit,
int start, int end, DocumentClassConstPtr dc, CutStack & cutstack)
{
ParagraphList const & pars = text.paragraphs();
// In most of these cases, we can try to recover.
LASSERT(0 <= start, start = 0);
LASSERT(start <= pars[startpit].size(), start = pars[startpit].size());
LASSERT(0 <= end, end = 0);
LASSERT(end <= pars[endpit].size(), end = pars[endpit].size());
LASSERT(startpit != endpit || start <= end, return);
// Clone the paragraphs within the selection.
ParagraphList copy_pars(pars.iterator_at(startpit),
pars.iterator_at(endpit + 1));
// Remove the end of the last paragraph; afterwards, remove the
// beginning of the first paragraph. Keep this order - there may only
// be one paragraph! Do not track deletions here; this is an internal
// action not visible to the user
Paragraph & back = copy_pars.back();
back.eraseChars(end, back.size(), false);
Paragraph & front = copy_pars.front();
front.eraseChars(0, start, false);
for (auto & par : copy_pars) {
// Since we have a copy of the paragraphs, the insets
// do not have a proper buffer reference. It makes
// sense to add them temporarily, because the
// operations below depend on that (acceptChanges included).
par.setInsetBuffers(const_cast<Buffer &>(buf));
// PassThru paragraphs have the Language
// latex_language. This is invalid for others, so we
// need to change it to the buffer language.
if (par.isPassThru())
par.changeLanguage(buf.params(),
latex_language, buf.language());
}
// do not copy text (also nested in insets) which is marked as
// deleted, unless the whole selection was deleted
if (!lyxrc.ct_markup_copied) {
if (!isFullyDeleted(copy_pars))
acceptChanges(copy_pars, buf.params());
else
rejectChanges(copy_pars, buf.params());
}
// do some final cleanup now, to make sure that the paragraphs
// are not linked to something else.
for (auto & par : copy_pars) {
par.resetBuffer();
par.setInsetOwner(nullptr);
}
cutstack.push(make_pair(copy_pars, make_pair(dc, buf.params().authors())));
}
} // namespace
namespace cap {
void region(CursorSlice const & i1, CursorSlice const & i2,
row_type & r1, row_type & r2,
col_type & c1, col_type & c2)
{
Inset const & p = i1.inset();
c1 = p.col(i1.idx());
c2 = p.col(i2.idx());
if (c1 > c2)
swap(c1, c2);
r1 = p.row(i1.idx());
r2 = p.row(i2.idx());
if (r1 > r2)
swap(r1, r2);
}
docstring grabAndEraseSelection(Cursor & cur)
{
if (!cur.selection())
return docstring();
docstring res = grabSelection(cur);
eraseSelection(cur);
return res;
}
bool reduceSelectionToOneCell(CursorData & cur)
{
if (!cur.selection() || !cur.inMathed())
return false;
CursorSlice i1 = cur.selBegin();
CursorSlice i2 = cur.selEnd();
if (!i1.inset().asInsetMath())
return false;
// the easy case: do nothing if only one cell is selected
if (i1.idx() == i2.idx())
return true;
cur.top().pos() = 0;
cur.resetAnchor();
cur.top().pos() = cur.top().lastpos();
return true;
}
bool multipleCellsSelected(CursorData const & cur)
{
if (!cur.selection() || !cur.inMathed())
return false;
CursorSlice i1 = cur.selBegin();
CursorSlice i2 = cur.selEnd();
if (!i1.inset().asInsetMath())
return false;
if (i1.idx() == i2.idx())
return false;
return true;
}
void switchBetweenClasses(DocumentClassConstPtr oldone,
DocumentClassConstPtr newone, InsetText & in) {
ErrorList el = {};
switchBetweenClasses(oldone, newone, in, el);
}
void switchBetweenClasses(DocumentClassConstPtr oldone,
DocumentClassConstPtr newone, InsetText & in, ErrorList & errorlist)
{
errorlist.clear();
LBUFERR(!in.paragraphs().empty());
if (oldone == newone)
return;
DocumentClass const & oldtc = *oldone;
DocumentClass const & newtc = *newone;
// layouts
ParIterator it = par_iterator_begin(in);
ParIterator pend = par_iterator_end(in);
// for remembering which layouts we've had to add
set<docstring> newlayouts;
for (; it != pend; ++it) {
docstring const name = it->layout().name();
// the pasted text will keep their own layout name. If this layout does
// not exist in the new document, it will behave like a standard layout.
bool const added_one = newtc.addLayoutIfNeeded(name);
if (added_one)
newlayouts.insert(name);
if (added_one || newlayouts.find(name) != newlayouts.end()) {
// Warn the user.
docstring const s = bformat(_("Layout `%1$s' was not found."), name);
errorlist.push_back(ErrorItem(_("Layout Not Found"), s,
{it->id(), 0}, {it->id(), -1}));
}
if (in.usePlainLayout())
it->setLayout(newtc.plainLayout());
else
it->setLayout(newtc[name]);
}
// character styles and hidden table cells
InsetIterator const i_end = end(in);
for (InsetIterator iit = begin(in); iit != i_end; ++iit) {
InsetCode const code = iit->lyxCode();
if (code == FLEX_CODE) {
// FIXME: Should we verify all InsetCollapsible?
docstring const layoutName = iit->layoutName();
docstring const & n = newone->insetLayout(layoutName).name();
bool const is_undefined = n.empty() ||
n == DocumentClass::plainInsetLayout().name();
docstring const & oldn = oldone->insetLayout(layoutName).name();
bool const was_undefined = oldn.empty() ||
oldn == DocumentClass::plainInsetLayout().name();
if (!is_undefined || was_undefined)
continue;
// The flex inset is undefined in newtc
docstring const oldname = from_utf8(oldtc.name());
docstring const newname = from_utf8(newtc.name());
docstring s;
if (oldname == newname)
s = bformat(_("Flex inset %1$s is undefined after "
"reloading `%2$s' layout."), layoutName, oldname);
else
s = bformat(_("Flex inset %1$s is undefined because of "
"conversion from `%2$s' layout to `%3$s'."),
layoutName, oldname, newname);
// To warn the user that something had to be done.
errorlist.push_back(ErrorItem(
_("Undefined flex inset"), s,
{iit.paragraph().id(), iit.pos()},
{iit.paragraph().id(), iit.pos() + 1}));
} else if (code == TABULAR_CODE) {
// The recursion above does not catch paragraphs in "hidden" cells,
// i.e., ones that are part of a multirow or multicolum. So we need
// to handle those separately.
// This is the cause of bug #9049.
InsetTabular * table = iit->asInsetTabular();
table->setLayoutForHiddenCells(newtc);
}
}
}
vector<docstring> availableSelections(Buffer const * buf)
{
vector<docstring> selList;
if (!buf)
return selList;
for (auto const & cut : theCuts) {
ParagraphList const & pars = cut.first;
docstring textSel;
for (auto const & para : pars) {
Paragraph par(para, 0, 46);
// adapt paragraph to current buffer.
par.setInsetBuffers(const_cast<Buffer &>(*buf));
textSel += par.asString(AS_STR_INSETS);
if (textSel.size() > 45) {
support::truncateWithEllipsis(textSel,45);
break;
}
}
selList.push_back(textSel);
}
return selList;
}
size_type numberOfSelections()
{
return theCuts.size();
}
namespace {
void cutSelectionHelper(Cursor & cur, CutStack & cuts, bool realcut, bool putclip)
{
// This doesn't make sense, if there is no selection
if (!cur.selection())
return;
// OK, we have a selection. This is always between cur.selBegin()
// and cur.selEnd()
if (cur.inTexted()) {
Text * text = cur.text();
LBUFERR(text);
saveSelection(cur);
// make sure that the depth behind the selection are restored, too
cur.recordUndoSelection();
pit_type begpit = cur.selBegin().pit();
pit_type endpit = cur.selEnd().pit();
int endpos = cur.selEnd().pos();
BufferParams const & bp = cur.buffer()->params();
if (realcut) {
copySelectionHelper(*cur.buffer(),
*text,
begpit, endpit,
cur.selBegin().pos(), endpos,
bp.documentClassPtr(), cuts);
// Stuff what we got on the clipboard.
// Even if there is no selection.
if (putclip)
putClipboard(cuts[0].first, cuts[0].second,
cur.selectionAsString(true, true), bp);
}
if (begpit != endpit)
cur.screenUpdateFlags(Update::Force | Update::FitCursor);
tie(endpit, endpos) =
eraseSelectionHelper(bp, text->paragraphs(), begpit, endpit,
cur.selBegin().pos(), endpos);
// cutSelection can invalidate the cursor so we need to set
// it anew. (Lgb)
// we prefer the end for when tracking changes
cur.pos() = endpos;
cur.pit() = endpit;
// need a valid cursor. (Lgb)
cur.clearSelection();
// After a cut operation, we must make sure that the Buffer is updated
// because some further operation might need updated label information for
// example. So we cannot just use "cur.forceBufferUpdate()" here.
// This fixes #7071.
cur.buffer()->updateBuffer();
// tell tabular that a recent copy happened
dirtyTabularStack(false);
}
if (cur.inMathed()) {
if (cur.selBegin().idx() != cur.selEnd().idx()) {
// The current selection spans more than one cell.
// Record all cells
cur.recordUndoInset();
} else {
// Record only the current cell to avoid a jumping
// cursor after undo
cur.recordUndo();
}
if (realcut)
copySelection(cur);
eraseSelection(cur);
}
}
} // namespace
void cutSelection(Cursor & cur, bool realcut)
{
cutSelectionHelper(cur, theCuts, realcut, true);
}
void cutSelectionToTemp(Cursor & cur, bool realcut)
{
cutSelectionHelper(cur, tempCut, realcut, false);
}
void copySelection(Cursor const & cur)
{
copySelection(cur, cur.selectionAsString(true, true));
}
void copyInset(Cursor const & cur, Inset * inset, docstring const & plaintext)
{
ParagraphList pars;
Paragraph par;
BufferParams const & bp = cur.buffer()->params();
par.setLayout(bp.documentClass().plainLayout());
Font font(inherit_font, bp.language);
par.insertInset(0, inset, font, Change(Change::UNCHANGED));
pars.push_back(par);
theCuts.push(make_pair(pars, make_pair(bp.documentClassPtr(), bp.authors())));
// stuff the selection onto the X clipboard, from an explicit copy request
putClipboard(theCuts[0].first, theCuts[0].second, plaintext, bp);
}
namespace {
void copySelectionToStack(CursorData const & cur, CutStack & cutstack)
{
// this doesn't make sense, if there is no selection
if (!cur.selection())
return;
// copySelection can not yet handle the case of cross idx selection
if (cur.selBegin().idx() != cur.selEnd().idx())
return;
if (cur.inTexted()) {
Text * text = cur.text();
LBUFERR(text);
// ok we have a selection. This is always between cur.selBegin()
// and sel_end cursor
copySelectionHelper(*cur.buffer(), *text,
cur.selBegin().pit(), cur.selEnd().pit(),
cur.selBegin().pos(), cur.selEnd().pos(),
cur.buffer()->params().documentClassPtr(),
cutstack);
// Reset the dirty_tabular_stack_ flag only when something
// is copied to the clipboard (not to the selectionBuffer).
if (&cutstack == &theCuts)
dirtyTabularStack(false);
}
if (cur.inMathed()) {
//lyxerr << "copySelection in mathed" << endl;
ParagraphList pars;
Paragraph par;
BufferParams const & bp = cur.buffer()->params();
// FIXME This should be the plain layout...right?
par.setLayout(bp.documentClass().plainLayout());
// For pasting into text, we set the language to the paragraph language
// (rather than the default_language which is always English; see #2596)
par.insert(0, grabSelection(cur), Font(sane_font, par.getParLanguage(bp)),
Change(Change::UNCHANGED));
pars.push_back(par);
cutstack.push(make_pair(pars, make_pair(bp.documentClassPtr(), bp.authors())));
}
}
} // namespace
void copySelectionToStack()
{
if (!selectionBuffer.empty())
theCuts.push(selectionBuffer[0]);
}
void copySelectionToTemp(Cursor const & cur)
{
copySelectionToStack(cur, tempCut);
}
void copySelection(Cursor const & cur, docstring const & plaintext)
{
// In tablemode, because copy and paste actually use a special table stack,
// we need to go through the cells and collect the paragraphs.
// In math matrices, we generate a plain text version.
if (cur.selBegin().idx() != cur.selEnd().idx()) {
ParagraphList pars;
BufferParams const & bp = cur.buffer()->params();
if (cur.inMathed()) {
Paragraph par;
par.setLayout(bp.documentClass().plainLayout());
// Replace (column-separating) tabs by space (#4449)
docstring const clean_text = subst(plaintext, '\t', ' ');
// For pasting into text, we set the language to the paragraph language
// (rather than the default_language which is always English; see #11898)
par.insert(0, clean_text, Font(sane_font, par.getParLanguage(bp)),
Change(Change::UNCHANGED));
pars.push_back(par);
} else {
// Get paragraphs from all cells
InsetTabular * table = cur.inset().asInsetTabular();
LASSERT(table, return);
ParagraphList tplist = table->asParList(cur.selBegin().idx(), cur.selEnd().idx());
for (auto & cpar : tplist) {
cpar.setLayout(bp.documentClass().plainLayout());
pars.push_back(cpar);
// since the pars are merged later, we separate them by blank
Paragraph epar;
epar.insert(0, from_ascii(" "), Font(sane_font, epar.getParLanguage(bp)),
Change(Change::UNCHANGED));
pars.push_back(epar);
}
// remove last empty par
pars.pop_back();
// merge all paragraphs to one
while (pars.size() > 1)
mergeParagraph(bp, pars, 0);
}
theCuts.push(make_pair(pars, make_pair(bp.documentClassPtr(), bp.authors())));
} else {
copySelectionToStack(cur, theCuts);
}
// stuff the selection onto the X clipboard, from an explicit copy request
putClipboard(theCuts[0].first, theCuts[0].second, plaintext,
cur.buffer()->params());
}
void saveSelection(Cursor const & cur)
{
// This function is called, not when a selection is formed, but when
// a selection is cleared. Therefore, multiple keyboard selection
// will not repeatively trigger this function (bug 3877).
if (cur.selection()
&& cur.selBegin() == cur.bv().cursor().selBegin()
&& cur.selEnd() == cur.bv().cursor().selEnd()) {
LYXERR(Debug::SELECTION, "saveSelection: '" << cur.selectionAsString(true, true) << "'");
copySelectionToStack(cur, selectionBuffer);
}
}
bool selection()
{
return !selectionBuffer.empty();
}
void clearSelection()
{
selectionBuffer.clear();
}
void clearCutStack()
{
theCuts.clear();
tempCut.clear();
}
docstring selection(size_t sel_index, DocInfoPair docinfo, bool for_math)
{
if (sel_index >= theCuts.size())
return docstring();
unique_ptr<Buffer> buffer(copyToTempBuffer(theCuts[sel_index].first,
docinfo.first));
if (!buffer)
return docstring();
int options = AS_STR_INSETS | AS_STR_NEWLINES;
if (for_math)
options |= AS_STR_MATHED;
return buffer->paragraphs().back().asString(options);
}
void pasteParagraphList(Cursor & cur, ParagraphList const & parlist,
DocumentClassConstPtr docclass, AuthorList const & authors,
ErrorList & errorList,
cap::BranchAction branchAction)
{
// Copy authors to the params. We need those pointers.
for (Author const & a : authors)
cur.buffer()->params().authors().record(a);
if (cur.inTexted()) {
Text * text = cur.text();
LBUFERR(text);
PasteReturnValue prv =
pasteSelectionHelper(cur, parlist, docclass, branchAction, errorList);
cur.forceBufferUpdate();
cur.clearSelection();
text->setCursor(cur, prv.pit, prv.pos);
}
// mathed is handled in InsetMathNest/InsetMathGrid
LATTEST(!cur.inMathed());
}
bool pasteFromStack(Cursor & cur, ErrorList & errorList, size_t sel_index)
{
// this does not make sense, if there is nothing to paste
if (!checkPastePossible(sel_index))
return false;
cur.recordUndo();
pasteParagraphList(cur, theCuts[sel_index].first,
theCuts[sel_index].second.first, theCuts[sel_index].second.second,
errorList, BRANCH_ASK);
return true;
}
bool pasteFromTemp(Cursor & cur, ErrorList & errorList)
{
// this does not make sense, if there is nothing to paste
if (tempCut.empty() || tempCut[0].first.empty())
return false;
cur.recordUndo();
pasteParagraphList(cur, tempCut[0].first,
tempCut[0].second.first, tempCut[0].second.second,
errorList, BRANCH_IGNORE);
return true;
}
bool pasteClipboardText(Cursor & cur, ErrorList & errorList, bool asParagraphs,
Clipboard::TextType type)
{
// Use internal clipboard if it is the most recent one
// This overrides asParagraphs and type on purpose!
if (theClipboard().isInternal())
return pasteFromStack(cur, errorList, 0);
// First try LyX format
if ((type == Clipboard::LyXTextType ||
type == Clipboard::LyXOrPlainTextType ||
type == Clipboard::AnyTextType) &&
theClipboard().hasTextContents(Clipboard::LyXTextType)) {
string lyx = theClipboard().getAsLyX();
if (!lyx.empty()) {
Buffer buffer(string(), false);
buffer.setInternal(true);
buffer.setUnnamed(true);
if (buffer.readString(lyx)) {
cur.recordUndo();
pasteParagraphList(cur, buffer.paragraphs(),
buffer.params().documentClassPtr(),
buffer.params().authors(),
errorList);
return true;
}
}
}
// Then try TeX and HTML
Clipboard::TextType types[2] = {Clipboard::HtmlTextType, Clipboard::LaTeXTextType};
string names[2] = {"html", "latexclipboard"};
for (int i = 0; i < 2; ++i) {
if (type != types[i] && type != Clipboard::AnyTextType)
continue;
bool available = theClipboard().hasTextContents(types[i]);
// If a specific type was explicitly requested, try to
// interpret plain text: The user told us that the clipboard
// contents is in the desired format
if (!available && type == types[i]) {
types[i] = Clipboard::PlainTextType;
available = theClipboard().hasTextContents(types[i]);
}
if (available) {
docstring text = theClipboard().getAsText(types[i]);
available = !text.empty();
if (available) {
Buffer buffer(string(), false);
buffer.setInternal(true);
buffer.setUnnamed(true);
buffer.params() = cur.buffer()->params();
available = buffer.importString(names[i], text, errorList);
if (available)
available = !buffer.paragraphs().empty();
if (available && !buffer.paragraphs()[0].empty()) {
// TeX2lyx (also used in the HTML chain) assumes English as document language
// if no language is explicitly set (as is the case here).
// We thus reset the temp buffer's language to the context language
buffer.changeLanguage(buffer.language(), cur.getFont().language());
cur.recordUndo();
pasteParagraphList(cur, buffer.paragraphs(),
buffer.params().documentClassPtr(),
buffer.params().authors(),
errorList);
return true;
}
}
}
}
// Then try plain text
docstring const text = theClipboard().getAsText(Clipboard::PlainTextType);
if (text.empty())
return false;
cur.recordUndo();
if (asParagraphs)
cur.text()->insertStringAsParagraphs(cur, text, cur.current_font);
else
cur.text()->insertStringAsLines(cur, text, cur.current_font);
cur.forceBufferUpdate();
return true;
}
void pasteSimpleText(Cursor & cur, bool asParagraphs)
{
docstring text;
// Use internal clipboard if it is the most recent one
if (theClipboard().isInternal()) {
if (!checkPastePossible(0))
return;
ParagraphList const & pars = theCuts[0].first;
ParagraphList::const_iterator it = pars.begin();
for (; it != pars.end(); ++it) {
if (it != pars.begin())
text += "\n";
text += (*it).asString();
}
asParagraphs = false;
} else {
// Then try plain text
text = theClipboard().getAsText(Clipboard::PlainTextType);
}
if (text.empty())
return;
cur.recordUndo();
cutSelection(cur, false);
if (asParagraphs)
cur.text()->insertStringAsParagraphs(cur, text, cur.current_font);
else
cur.text()->insertStringAsLines(cur, text, cur.current_font);
}
void pasteClipboardGraphics(Cursor & cur, ErrorList & /* errorList */,
Clipboard::GraphicsType preferedType)
{
LASSERT(theClipboard().hasGraphicsContents(preferedType), return);
// get picture from clipboard
FileName filename = theClipboard().getAsGraphics(cur, preferedType);
if (filename.empty())
return;
// create inset for graphic
InsetGraphics * inset = new InsetGraphics(cur.buffer());
InsetGraphicsParams params;
params.filename = support::DocFileName(filename.absFileName(), false);
inset->setParams(params);
cur.recordUndo();
cur.insert(inset);
}
void pasteSelection(Cursor & cur, ErrorList & errorList)
{
if (selectionBuffer.empty())
return;
cur.recordUndo();
pasteParagraphList(cur, selectionBuffer[0].first,
selectionBuffer[0].second.first,
selectionBuffer[0].second.second,
errorList);
}
void replaceSelectionWithString(Cursor & cur, docstring const & str)
{
cur.recordUndo();
DocIterator selbeg = cur.selectionBegin();
// Get font setting before we cut, we need a copy here, not a bare reference.
Font const font =
selbeg.paragraph().getFontSettings(cur.buffer()->params(), selbeg.pos());
// Insert the new string
pos_type pos = cur.selEnd().pos();
Paragraph & par = cur.selEnd().paragraph();
for (auto const & c : str) {
par.insertChar(pos, c, font, cur.buffer()->params().track_changes);
++pos;
}
// Cut the selection
cutSelection(cur, false);
}
void replaceSelection(Cursor & cur)
{
if (cur.selection())
cutSelection(cur, false);
}
void eraseSelection(Cursor & cur)
{
//lyxerr << "cap::eraseSelection begin: " << cur << endl;
CursorSlice const & i1 = cur.selBegin();
CursorSlice const & i2 = cur.selEnd();
if (!i1.asInsetMath()) {
LYXERR0("Can't erase this selection");
return;
}
saveSelection(cur);
cur.top() = i1;
InsetMath * p = i1.asInsetMath();
if (i1.idx() == i2.idx()) {
i1.cell().erase(i1.pos(), i2.pos());
// We may have deleted i1.cell(cur.pos()).
// Make sure that pos is valid.
if (cur.pos() > cur.lastpos())
cur.pos() = cur.lastpos();
} else if (p->nrows() > 0 && p->ncols() > 0) {
// This is a grid, delete a nice square region
row_type r1, r2;
col_type c1, c2;
region(i1, i2, r1, r2, c1, c2);
for (row_type row = r1; row <= r2; ++row)
for (col_type col = c1; col <= c2; ++col)
p->cell(p->index(row, col)).clear();
// We've deleted the whole cell. Only pos 0 is valid.
cur.pos() = 0;
} else {
idx_type idx1 = i1.idx();
idx_type idx2 = i2.idx();
if (idx1 > idx2)
swap(idx1, idx2);
for (idx_type idx = idx1 ; idx <= idx2; ++idx)
p->cell(idx).clear();
// We've deleted the whole cell. Only pos 0 is valid.
cur.pos() = 0;
}
// need a valid cursor. (Lgb)
cur.clearSelection();
//lyxerr << "cap::eraseSelection end: " << cur << endl;
}
void selDel(Cursor & cur)
{
//lyxerr << "cap::selDel" << endl;
if (cur.selection())
eraseSelection(cur);
}
void selClearOrDel(Cursor & cur)
{
//lyxerr << "cap::selClearOrDel" << endl;
if (lyxrc.auto_region_delete)
selDel(cur);
else
cur.selection(false);
}
docstring grabSelection(CursorData const & cur)
{
if (!cur.selection())
return docstring();
#if 0
// grab selection by glueing multiple cells together. This is not what
// we want because selections spanning multiple cells will get "&" and "\\"
// seperators.
ostringstream os;
for (DocIterator dit = cur.selectionBegin();
dit != cur.selectionEnd(); dit.forwardPos())
os << asString(dit.cell());
return os.str();
#endif
CursorSlice i1 = cur.selBegin();
CursorSlice i2 = cur.selEnd();
if (i1.idx() == i2.idx()) {
if (i1.inset().asInsetMath()) {
MathData::const_iterator it = i1.cell().begin();
Buffer * buf = cur.buffer();
return asString(MathData(buf, it + i1.pos(), it + i2.pos()));
} else {
return from_ascii("unknown selection 1");
}
}
row_type r1, r2;
col_type c1, c2;
region(i1, i2, r1, r2, c1, c2);
docstring data;
if (i1.inset().asInsetMath()) {
for (row_type row = r1; row <= r2; ++row) {
if (row > r1)
data += "\\\\";
for (col_type col = c1; col <= c2; ++col) {
if (col > c1)
data += '&';
data += asString(i1.asInsetMath()->
cell(i1.asInsetMath()->index(row, col)));
}
}
} else {
data = from_ascii("unknown selection 2");
}
return data;
}
void dirtyTabularStack(bool b)
{
dirty_tabular_stack_ = b;
}
bool tabularStackDirty()
{
return dirty_tabular_stack_;
}
} // namespace cap
} // namespace lyx