kindred/create
1
1
Fork 0
Code Issues 37 Pull Requests Actions Packages Projects 9 Releases 2 Wiki Activity

Merge pull request #9060 from kpemartin/master

Browse Source 
Fix encoding error for import from older DXF files
This commit is contained in:
Yorik van Havre
2023-04-13 13:03:13 +02:00
committed by GitHub
parent 32650526b7 9703853ee1
commit 00b8693ac0
2 changed files with 208 additions and 48 deletions
Download Patch File Download Diff File Expand all files Collapse all files

217
src/Mod/Import/App/dxf/dxf.cpp
View File

@@ -18,7 +18,7 @@
#include <Base/FileInfo.h>
#include <Base/Stream.h>
#include <Base/Vector3D.h>
#include <Base/Interpreter.h>
#include "dxf.h"
@@ -1750,7 +1750,7 @@ CDxfRead::CDxfRead(const char* filepath)
memset( m_str, '\0', sizeof(m_str) );
memset( m_unused_line, '\0', sizeof(m_unused_line) );
m_fail = false;
m_aci = 0;
m_ColorIndex = 0;
m_eUnits = eMillimeters;
m_measurement_inch = false;
strcpy(m_layer_name, "0"); // Default layer name
@@ -1766,11 +1766,16 @@ CDxfRead::CDxfRead(const char* filepath)
}
m_ifs->imbue(std::locale("C"));
m_version = RUnknown;
m_CodePage = NULL;
m_encoding = NULL;
}
CDxfRead::~CDxfRead()
{
delete m_ifs;
delete m_CodePage;
delete m_encoding;
}
double CDxfRead::mm( double value ) const
@@ -1836,7 +1841,7 @@ bool CDxfRead::ReadLine()
switch(n){
case 0:
// next item found, so finish with line
DerefACI();
ResolveColorIndex();
OnReadLine(s, e, hidden);
hidden = false;
return true;
@@ -1884,7 +1889,7 @@ bool CDxfRead::ReadLine()
case 62:
// color index
get_line();
ss.str(m_str); ss >> m_aci; if(ss.fail()) return false;
ss.str(m_str); ss >> m_ColorIndex; if(ss.fail()) return false;
break;
case 100:
@@ -1903,7 +1908,7 @@ bool CDxfRead::ReadLine()
}
try {
DerefACI();
ResolveColorIndex();
OnReadLine(s, e, false);
}
catch(...)
@@ -1934,7 +1939,7 @@ bool CDxfRead::ReadPoint()
switch(n){
case 0:
// next item found, so finish with line
DerefACI();
ResolveColorIndex();
OnReadPoint(s);
return true;
@@ -1962,7 +1967,7 @@ bool CDxfRead::ReadPoint()
case 62:
// color index
get_line();
ss.str(m_str); ss >> m_aci; if(ss.fail()) return false;
ss.str(m_str); ss >> m_ColorIndex; if(ss.fail()) return false;
break;
case 100:
@@ -1982,7 +1987,7 @@ bool CDxfRead::ReadPoint()
}
try {
DerefACI();
ResolveColorIndex();
OnReadPoint(s);
}
catch(...)
@@ -2017,7 +2022,7 @@ bool CDxfRead::ReadArc()
switch(n){
case 0:
// next item found, so finish with arc
DerefACI();
ResolveColorIndex();
OnReadArc(start_angle, end_angle, radius, c,z_extrusion_dir, hidden);
hidden = false;
return true;
@@ -2065,7 +2070,7 @@ bool CDxfRead::ReadArc()
case 62:
// color index
get_line();
ss.str(m_str); ss >> m_aci; if(ss.fail()) return false;
ss.str(m_str); ss >> m_ColorIndex; if(ss.fail()) return false;
break;
@@ -2089,7 +2094,7 @@ bool CDxfRead::ReadArc()
break;
}
}
DerefACI();
ResolveColorIndex();
OnReadArc(start_angle, end_angle, radius, c, z_extrusion_dir, false);
return false;
}
@@ -2122,7 +2127,7 @@ bool CDxfRead::ReadSpline()
switch(n){
case 0:
// next item found, so finish with Spline
DerefACI();
ResolveColorIndex();
OnReadSpline(sd);
return true;
case 8: // Layer name follows
@@ -2132,7 +2137,7 @@ bool CDxfRead::ReadSpline()
case 62:
// color index
get_line();
ss.str(m_str); ss >> m_aci; if(ss.fail()) return false;
ss.str(m_str); ss >> m_ColorIndex; if(ss.fail()) return false;
break;
case 210:
// normal x
@@ -2270,7 +2275,7 @@ bool CDxfRead::ReadSpline()
break;
}
}
DerefACI();
ResolveColorIndex();
OnReadSpline(sd);
return false;
}
@@ -2296,7 +2301,7 @@ bool CDxfRead::ReadCircle()
switch(n){
case 0:
// next item found, so finish with Circle
DerefACI();
ResolveColorIndex();
OnReadCircle(c, radius, hidden);
hidden = false;
return true;
@@ -2334,7 +2339,7 @@ bool CDxfRead::ReadCircle()
case 62:
// color index
get_line();
ss.str(m_str); ss >> m_aci; if(ss.fail()) return false;
ss.str(m_str); ss >> m_ColorIndex; if(ss.fail()) return false;
break;
case 100:
@@ -2351,7 +2356,7 @@ bool CDxfRead::ReadCircle()
break;
}
}
DerefACI();
ResolveColorIndex();
OnReadCircle(c, radius, false);
return false;
}
@@ -2417,15 +2422,19 @@ bool CDxfRead::ReadText()
// all code 0 records. Changing this would require either some sort of peek/pushback ability or the understanding
// that ReadText() and all the other Read... methods return having already read a code 0.
get_line();
DerefACI();
textPrefix.append(m_str);
OnReadText(c, height * 25.4 / 72.0, textPrefix.c_str());
ResolveColorIndex();
{
const char* utfStr = (this->*stringToUTF8)(m_str);
OnReadText(c, height * 25.4 / 72.0, utfStr);
if (utfStr != m_str)
delete utfStr;
}
return(true);
case 62:
// color index
get_line();
ss.str(m_str); ss >> m_aci; if(ss.fail()) return false;
ss.str(m_str); ss >> m_ColorIndex; if(ss.fail()) return false;
break;
case 100:
@@ -2469,7 +2478,7 @@ bool CDxfRead::ReadEllipse()
switch(n){
case 0:
// next item found, so finish with Ellipse
DerefACI();
ResolveColorIndex();
OnReadEllipse(c, m, ratio, start, end);
return true;
case 8: // Layer name follows
@@ -2525,7 +2534,7 @@ bool CDxfRead::ReadEllipse()
case 62:
// color index
get_line();
ss.str(m_str); ss >> m_aci; if(ss.fail()) return false;
ss.str(m_str); ss >> m_ColorIndex; if(ss.fail()) return false;
break;
case 100:
case 210:
@@ -2540,7 +2549,7 @@ bool CDxfRead::ReadEllipse()
break;
}
}
DerefACI();
ResolveColorIndex();
OnReadEllipse(c, m, ratio, start, end);
return false;
}
@@ -2640,7 +2649,7 @@ bool CDxfRead::ReadLwPolyLine()
case 0:
// next item found
DerefACI();
ResolveColorIndex();
if(x_found && y_found){
// add point
AddPolyLinePoint(this, x, y, z, bulge_found, bulge);
@@ -2694,7 +2703,7 @@ bool CDxfRead::ReadLwPolyLine()
case 62:
// color index
get_line();
ss.str(m_str); ss >> m_aci; if(ss.fail()) return false;
ss.str(m_str); ss >> m_ColorIndex; if(ss.fail()) return false;
break;
default:
// skip the next line
@@ -2708,7 +2717,7 @@ bool CDxfRead::ReadLwPolyLine()
if(closed && poly_first_found)
{
// repeat the first point
DerefACI();
ResolveColorIndex();
AddPolyLinePoint(this, poly_first_x, poly_first_y, poly_first_z, false, 0.0);
}
return true;
@@ -2744,7 +2753,7 @@ bool CDxfRead::ReadVertex(double *pVertex, bool *bulge_found, double *bulge)
ss.imbue(std::locale("C"));
switch(n){
case 0:
DerefACI();
ResolveColorIndex();
put_line(m_str); // read one line too many. put it back.
return(x_found && y_found);
break;
@@ -2780,7 +2789,7 @@ bool CDxfRead::ReadVertex(double *pVertex, bool *bulge_found, double *bulge)
case 62:
// color index
get_line();
ss.str(m_str); ss >> m_aci; if(ss.fail()) return false;
ss.str(m_str); ss >> m_ColorIndex; if(ss.fail()) return false;
break;
default:
@@ -2820,7 +2829,7 @@ bool CDxfRead::ReadPolyLine()
switch(n){
case 0:
// next item found
DerefACI();
ResolveColorIndex();
get_line();
if (! strcmp(m_str,"VERTEX"))
{
@@ -2855,7 +2864,7 @@ bool CDxfRead::ReadPolyLine()
case 62:
// color index
get_line();
ss.str(m_str); ss >> m_aci; if(ss.fail()) return false;
ss.str(m_str); ss >> m_ColorIndex; if(ss.fail()) return false;
break;
default:
// skip the next line
@@ -2943,7 +2952,7 @@ bool CDxfRead::ReadInsert()
switch(n){
case 0:
// next item found
DerefACI();
ResolveColorIndex();
OnReadInsert(c, s, name, rot * M_PI/180);
return(true);
case 8:
@@ -2994,7 +3003,7 @@ bool CDxfRead::ReadInsert()
case 62:
// color index
get_line();
ss.str(m_str); ss >> m_aci; if(ss.fail()) return false;
ss.str(m_str); ss >> m_ColorIndex; if(ss.fail()) return false;
break;
case 100:
case 39:
@@ -3035,7 +3044,7 @@ bool CDxfRead::ReadDimension()
switch(n){
case 0:
// next item found
DerefACI();
ResolveColorIndex();
OnReadDimension(s, e, p, rot * M_PI/180);
return(true);
case 8:
@@ -3096,7 +3105,7 @@ bool CDxfRead::ReadDimension()
case 62:
// color index
get_line();
ss.str(m_str); ss >> m_aci; if(ss.fail()) return false;
ss.str(m_str); ss >> m_ColorIndex; if(ss.fail()) return false;
break;
case 100:
case 39:
@@ -3216,7 +3225,7 @@ bool CDxfRead::ReadUnits()
bool CDxfRead::ReadLayer()
{
std::string layername;
int aci = -1;
ColorIndex_t colorIndex = -1;
while(!((*m_ifs).eof()))
{
@@ -3238,7 +3247,7 @@ bool CDxfRead::ReadLayer()
printf("CDxfRead::ReadLayer() - no layer name\n");
return false;
}
m_layer_aci[layername] = aci;
m_layer_ColorIndex_map[layername] = colorIndex;
return true;
case 2: // Layer name follows
@@ -3249,7 +3258,7 @@ bool CDxfRead::ReadLayer()
case 62:
// layer color ; if negative, layer is off
get_line();
if(sscanf(m_str, "%d", &aci) != 1)
if(sscanf(m_str, "%d", &colorIndex) != 1)
return false;
break;
@@ -3271,7 +3280,117 @@ bool CDxfRead::ReadLayer()
return false;
}
void CDxfRead::DoRead(const bool ignore_errors /* = false */ )
bool CDxfRead::ReadVersion()
{
static const std::vector<std::string> VersionNames = {
// This table is indexed by eDXFVersion_t - (ROlder+1)
"AC1006",
"AC1009",
"AC1012",
"AC1014",
"AC1015",
"AC1018",
"AC1021",
"AC1024",
"AC1027",
"AC1032"};
assert(VersionNames.size() == RNewer - ROlder - 1);
get_line();
get_line();
std::vector<std::string>::const_iterator first = VersionNames.cbegin();
std::vector<std::string>::const_iterator last = VersionNames.cend();
std::vector<std::string>::const_iterator found = std::lower_bound(first, last, m_str);
if (found == last)
m_version = RNewer;
else if (*found == m_str)
m_version = (eDXFVersion_t)(std::distance(first, found) + (ROlder + 1));
else if (found == first)
m_version = ROlder;
else
m_version = RUnknown;
return ResolveEncoding();
}
bool CDxfRead::ReadDWGCodePage()
{
get_line();
get_line();
assert(m_CodePage == NULL); // If not, we have found two DWGCODEPAGE variables or DoRead was called twice on the same CDxfRead object.
m_CodePage = new std::string(m_str);
return ResolveEncoding();
}
bool CDxfRead::ResolveEncoding()
{
if (m_encoding != NULL) {
delete m_encoding;
m_encoding = NULL;
}
if (m_version >= R2007) { // Note this does not include RUnknown, but does include RLater
m_encoding = new std::string("utf_8");
stringToUTF8 = &CDxfRead::UTF8ToUTF8;
}
else if (m_CodePage == NULL) {
// cp1252
m_encoding = new std::string("cp1252");
stringToUTF8 = &CDxfRead::GeneralToUTF8;
}
else {
// Codepage names may be of the form "ansi_1252" which we map to "cp1252" but we don't map "ansi_x3xxxx" (which happens to mean "ascii")
std::string* p = new std::string(*m_CodePage);
if (strncmp(p->c_str(), "ansi_", 5) == 0
&& strncmp(p->c_str(), "ansi_x3", 7) != 0)
p->replace(0, 5, "cp");
m_encoding = p;
// At this point we want to recognize synonyms for "utf_8" and use the custom decoder function.
// This is because this is one of the common cases and our decoder function is a fast no-op.
// We don't actually use the decoder function we get from PyCodec_Decoder because to call it we have to convert the (char *) text into
// a 'bytes' object first so we can pass it to the function using PyObject_Callxxx(), getting the PYObject containing the
// Python string, which we then decode back to UTF-8. It is simpler to call PyUnicode_DecodeXxxx which takes a (const char *)
// and is just a direct c++ callable.
Base::PyGILStateLocker lock;
PyObject* pyDecoder = PyCodec_Decoder(m_encoding->c_str());
if (pyDecoder == NULL)
return false; // A key error exception will have been placed.
PyObject* pyUTF8Decoder = PyCodec_Decoder("utf_8");
assert(pyUTF8Decoder != NULL);
if (pyDecoder == pyUTF8Decoder)
stringToUTF8 = &CDxfRead::UTF8ToUTF8;
else
stringToUTF8 = &CDxfRead::GeneralToUTF8;
Py_DECREF(pyDecoder);
Py_DECREF(pyUTF8Decoder);
}
return m_encoding != NULL;
}
const char* CDxfRead::UTF8ToUTF8(const char* encoded) const
{
return encoded;
}
const char* CDxfRead::GeneralToUTF8(const char* encoded) const
{
Base::PyGILStateLocker lock;
PyObject* decoded = PyUnicode_Decode(encoded, strlen(encoded), m_encoding->c_str(), "strict");
if (decoded == NULL)
return NULL;
Py_ssize_t len;
const char* converted = PyUnicode_AsUTF8AndSize(decoded, &len);
char* result = NULL;
if (converted != NULL) {
// converted only has lifetime of decoded so we must save a copy.
result = (char *)malloc(len + 1);
if (result == NULL)
PyErr_SetString(PyExc_MemoryError, "Out of memory");
else
memcpy(result, converted, len + 1);
}
Py_DECREF(decoded);
return result;
}
void CDxfRead::DoRead(const bool ignore_errors /* = false */)
{
m_ignore_errors = ignore_errors;
if(m_fail)
@@ -3298,7 +3417,19 @@ void CDxfRead::DoRead(const bool ignore_errors /* = false */ )
continue;
} // End if - then
else if(!strcmp(m_str, "0"))
if (!strcmp(m_str, "$ACADVER")) {
if (!ReadVersion())
return;
continue;
}// End if - then
if (!strcmp(m_str, "$DWGCODEPAGE")) {
if (!ReadDWGCodePage())
return;
continue;
}// End if - then
if (!strcmp(m_str, "0"))
{
get_line();
if (!strcmp( m_str, "SECTION" )){
@@ -3443,12 +3574,12 @@ void CDxfRead::DoRead(const bool ignore_errors /* = false */ )
}
void CDxfRead::DerefACI()
void CDxfRead::ResolveColorIndex()
{
if (m_aci == 256) // if color = layer color, replace by color from layer
if (m_ColorIndex == ColorBylayer) // if color = layer color, replace by color from layer
{
m_aci = m_layer_aci[std::string(m_layer_name)];
m_ColorIndex = m_layer_ColorIndex_map[std::string(m_layer_name)];
}
}

39
src/Mod/Import/App/dxf/dxf.h
View File

@@ -26,7 +26,7 @@
#define HAVE_IOSTREAM
#endif
typedef int Aci_t; // AutoCAD color index
typedef int ColorIndex_t; // DXF color index
typedef enum
{
@@ -118,6 +118,22 @@ struct LWPolyDataOut
std::vector<double> Bulge;
point3D Extr;
};
typedef enum
{
RUnknown,
ROlder,
R10,
R11_12,
R13,
R14,
R2000,
R2004,
R2007,
R2010,
R2013,
R2018,
RNewer,
} eDXFVersion_t;
//********************
class CDxfWrite{
@@ -256,8 +272,9 @@ private:
bool m_ignore_errors;
typedef std::map< std::string,Aci_t > LayerAciMap_t;
LayerAciMap_t m_layer_aci; // layer names -> layer color aci map
std::map<std::string,ColorIndex_t> m_layer_ColorIndex_map; // Mapping from layer name -> layer color index
const ColorIndex_t ColorBylayer = 256;
const ColorIndex_t ColorByBlock = 0;
bool ReadUnits();
bool ReadLayer();
@@ -277,13 +294,25 @@ private:
bool ReadInsert();
bool ReadDimension();
bool ReadBlockInfo();
bool ReadVersion();
bool ReadDWGCodePage();
bool ResolveEncoding();
void get_line();
void put_line(const char *value);
void DerefACI();
void ResolveColorIndex();
protected:
Aci_t m_aci; // manifest color name or 256 for layer color
ColorIndex_t m_ColorIndex;
eDXFVersion_t m_version;// Version from $ACADVER variable in DXF
const char* (CDxfRead::*stringToUTF8)(const char*) const;
private:
const std::string* m_CodePage; // Code Page name from $DWGCODEPAGE or null if none/not read yet
// The following was going to be python's canonical name for the encoding, but this is (a) not easily found and (b) does not speed up finding the encoding object.
const std::string* m_encoding;// A name for the encoding implied by m_version and m_CodePage
const char* UTF8ToUTF8(const char* encoded) const;
const char* GeneralToUTF8(const char* encoded) const;
public:
ImportExport CDxfRead(const char* filepath); // this opens the file