|
|
/*
* This program source code file is part of KiCad, a free EDA CAD application. * * Copyright (C) 2010 SoftPLC Corporation, Dick Hollenbeck <dick@softplc.com> * Copyright (C) 2012 Wayne Stambaugh <stambaughw@gmail.com> * Copyright (C) 2010-2020 KiCad Developers, see change_log.txt for contributors. * * 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, you may find one here: * http://www.gnu.org/licenses/old-licenses/gpl-2.0.html
* or you may search the http://www.gnu.org website for the version 2 license,
* or you may write to the Free Software Foundation, Inc., * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA */
#include <cstring>
#include <memory>
#include <wx/translation.h>
#include <macros.h> // TO_UTF8()
#include <lib_id.h>
static inline int okLogical( const UTF8& aField ){ // std::string::npos is largest positive number, casting to int makes it -1.
// Returning that means success.
return int( aField.find_first_of( ":" ) );}
void LIB_ID::clear(){ m_libraryName.clear(); m_itemName.clear(); m_subLibraryName.clear();}
int LIB_ID::Parse( const UTF8& aId, bool aFix ){ clear();
size_t partNdx; int offset = -1;
//=====<library nickname>=============================
if( ( partNdx = aId.find( ':' ) ) != aId.npos ) { offset = SetLibNickname( aId.substr( 0, partNdx ) );
if( offset > -1 ) return offset;
++partNdx; // skip ':'
} else { partNdx = 0; }
//=====<item name>====================================
UTF8 fpname = aId.substr( partNdx );
// Be sure the item name is valid.
// Some chars can be found in legacy files converted files from other EDA tools.
if( aFix ) fpname = FixIllegalChars( fpname, false ); else offset = HasIllegalChars( fpname );
if( offset > -1 ) return offset;
SetLibItemName( fpname );
return -1;}
LIB_ID::LIB_ID( const wxString& aLibraryName, const wxString& aItemName ) : m_libraryName( aLibraryName ), m_itemName( aItemName ){}
int LIB_ID::SetLibNickname( const UTF8& aLogical ){ int offset = okLogical( aLogical );
if( offset == -1 ) m_libraryName = aLogical;
return offset;}
int LIB_ID::SetLibItemName( const UTF8& aLibItemName ){ m_itemName = aLibItemName;
return -1;}
UTF8 LIB_ID::Format() const{ UTF8 ret;
if( m_libraryName.size() ) { ret += m_libraryName; ret += ':'; }
ret += m_itemName;
return ret;}
UTF8 LIB_ID::Format( const UTF8& aLibraryName, const UTF8& aLibItemName ){ UTF8 ret; int offset;
if( aLibraryName.size() ) { offset = okLogical( aLibraryName );
if( offset != -1 ) { THROW_PARSE_ERROR( _( "Illegal character found in logical library name" ), wxString::FromUTF8( aLibraryName.c_str() ), aLibraryName.c_str(), 0, offset ); }
ret += aLibraryName; ret += ':'; }
ret += aLibItemName; // TODO: Add validity test.
return ret;}
int LIB_ID::compare( const LIB_ID& aLibId ) const{ // Don't bother comparing the same object.
if( this == &aLibId ) return 0;
int retv = m_libraryName.compare( aLibId.m_libraryName );
if( retv != 0 ) return retv;
return m_itemName.compare( aLibId.m_itemName );}
int LIB_ID::HasIllegalChars( const UTF8& aLibItemName ){ int offset = 0;
for( auto& ch : aLibItemName ) { if( !isLegalChar( ch ) ) return offset; else ++offset; }
return -1;}
UTF8 LIB_ID::FixIllegalChars( const UTF8& aLibItemName, bool aLib ){ UTF8 fixedName;
for( UTF8::uni_iter chIt = aLibItemName.ubegin(); chIt < aLibItemName.uend(); ++chIt ) { auto ch = *chIt; if( aLib ) fixedName += isLegalLibraryNameChar( ch ) ? ch : '_'; else fixedName += isLegalChar( ch ) ? ch : '_'; }
return fixedName;}
bool LIB_ID::isLegalChar( unsigned aUniChar ){ bool const space_allowed = true; bool const illegal_filename_chars_allowed = false;
if( aUniChar < ' ' ) return false;
// This list of characters is also duplicated in validators.cpp and footprint.cpp
// TODO: Unify forbidden character lists - Warning, invalid filename characters are not the same
// as invalid LIB_ID characters. We will need to separate the FP filenames from FP names before this
// can be unified
switch( aUniChar ) { case ':': case '\t': case '\n': case '\r': return false;
case '\\': case '<': case '>': case '"': return illegal_filename_chars_allowed;
case ' ': return space_allowed;
default: return true; }}
unsigned LIB_ID::FindIllegalLibraryNameChar( const UTF8& aLibraryName ){ for( unsigned ch : aLibraryName ) { if( !isLegalLibraryNameChar( ch ) ) return ch; }
return 0;}
bool LIB_ID::isLegalLibraryNameChar( unsigned aUniChar ){ bool const space_allowed = true;
if( aUniChar < ' ' ) return false;
switch( aUniChar ) { case '\\': case ':': return false;
case ' ': return space_allowed;
default: return true; }}
const wxString LIB_ID::GetFullLibraryName() const{ wxString suffix = m_subLibraryName.wx_str().IsEmpty() ? wxString( wxS( "" ) ) : wxString::Format( wxT( " - %s" ), m_subLibraryName.wx_str() ); return wxString::Format( wxT( "%s%s" ), m_libraryName.wx_str(), suffix );}
|