--- parser3/src/classes/image.C 2017/02/06 16:17:12 1.159 +++ parser3/src/classes/image.C 2019/12/03 15:09:49 1.166 @@ -1,7 +1,7 @@ /** @file Parser: @b image parser class. - Copyright (c) 2001-2015 Art. Lebedev Studio (http://www.artlebedev.com) + Copyright (c) 2001-2017 Art. Lebedev Studio (http://www.artlebedev.com) Author: Alexandr Petrosian (http://paf.design.ru) */ @@ -24,8 +24,9 @@ #include "pa_vimage.h" #include "pa_vdate.h" #include "pa_table.h" +#include "pa_charsets.h" -volatile const char * IDENT_IMAGE_C="$Id: image.C,v 1.159 2017/02/06 16:17:12 moko Exp $"; +volatile const char * IDENT_IMAGE_C="$Id: image.C,v 1.166 2019/12/03 15:09:49 moko Exp $"; // defines @@ -290,6 +291,7 @@ public: switch(whence) { case SEEK_CUR: new_offset=offset+value; break; case SEEK_SET: new_offset=(size_t)value; break; + case SEEK_END: new_offset=size; break; default: throw Exception(0, 0, "whence #%d not supported", 0, whence); break; // never @@ -309,23 +311,36 @@ public: /// PNG file header struct PNG_Header { - char dummy[12]; - char signature[4]; //< must be "IHDR" - uchar high_width[2]; //< image width high bytes [we ignore for now] - uchar width[2]; //< image width low bytes - uchar high_height[2]; //< image height high bytes [we ignore for now] - uchar height[4]; //< image height + char dummy[12]; + char signature[4]; //< must be "IHDR" + uchar high_width[2]; //< image width high bytes [we ignore for now] + uchar width[2]; //< image width low bytes + uchar high_height[2]; //< image height high bytes [we ignore for now] + uchar height[4]; //< image height }; /// GIF file header struct GIF_Header { - char signature[3]; // 'GIF' - char version[3]; - uchar width[2]; - uchar height[2]; - char dif; - char fonColor; - char nulls; + char signature[3]; // 'GIF' + char version[3]; + uchar width[2]; + uchar height[2]; + char dif; + char fonColor; + char nulls; +}; + +/// BMP file header + DIB header part +struct BMP_Header { + char signature[2]; // 'BM' + uchar file_size[4]; + uchar reserved[4]; + uchar bitmap_offset[4]; + uchar header_size[4]; + uchar width[2]; + uchar high_width[2]; //< image width high bytes [we ignore for now] + uchar height[2]; + uchar high_height[2]; //< image height high bytes [we ignore for now] }; /// JPEG record head @@ -336,21 +351,16 @@ struct JPG_Segment_head { }; /// JPEG frame header struct JPG_Size_segment_body { - char data; //< data precision of bits/sample - uchar height[2]; //< image height - uchar width[2]; //< image width - char numComponents; //< number of color components -}; - -/// JPEG frame header -struct JPG_Exif_segment_begin { - char signature[6]; // Exif\0\0 + char data; //< data precision of bits/sample + uchar height[2]; //< image height + uchar width[2]; //< image width + char numComponents; //< number of color components }; /// JPEG Exif TIFF Header struct JPG_Exif_TIFF_header { - uchar byte_align_identifier[2]; - char dummy[2]; // always 000A [or 0A00] + char byte_align_identifier[2]; + uchar signature[2]; // always 000A [or 0A00] uchar first_IFD_offset[4]; // Usually the first IFD starts immediately next to TIFF header, so this offset has value '0x00000008'. }; @@ -373,6 +383,17 @@ struct JPG_Exif_IFD_entry { #define JPEG_EXIF_DATE_CHARS 20 + +#ifndef DOXYGEN +struct Measure_info { + ushort width; + ushort height; + Value** exif; + Value** xmp; + Charset* xmp_charset; +}; +#endif + // inline ushort x_endian_to_ushort(uchar b0, uchar b1) { @@ -393,21 +414,6 @@ inline uint endian_to_uint(bool is_big, x_endian_to_uint(b[0], b[1], b[2], b[3]); } -static void measure_gif(const String& origin_string, Measure_reader& reader, ushort& width, ushort& height) { - - const char* buf; - const size_t head_size=sizeof(GIF_Header); - if(reader.read(buf, head_size)signature, "GIF", 3)!=0) - throw Exception(IMAGE_FORMAT, &origin_string, "not GIF file - wrong signature"); - - width=endian_to_ushort(false, head->width); - height=endian_to_ushort(false, head->height); -} - static Value* parse_IFD_entry_formatted_one_value(bool is_big, ushort format, size_t component_size, const uchar *value) { switch(format) { case 1: // unsigned byte @@ -576,14 +582,8 @@ static void parse_IFD(HashStringValue& h // then goes: LLLLLLLL Offset to next IFD [not going there] } -static Value* parse_exif(Measure_reader& reader, const String& origin_string) { +static Value* parse_exif(Measure_reader& reader) { const char* buf; - if(reader.read(buf, sizeof(JPG_Exif_segment_begin))signature, "Exif\0\0", 4+2)!=0) //signature invalid? - return 0; // ignore invalid block uint tiff_base=reader.tell(); if(reader.read(buf, sizeof(JPG_Exif_TIFF_header))marker!=MARKER) throw Exception(IMAGE_FORMAT, &origin_string, "not JPEG file - marker not found"); + ushort segment_length=endian_to_ushort(true, head->length); + switch(head->code) { - // http://park2.wakwak.com/~tsuruzoh/Computer/Digicams/exif-e.html - case CODE_EXIF: - if(exif && !*exif) // seen .jpg with some xml under EXIF tag, after real exif block :) - *exif=parse_exif(reader, origin_string); + // http://dev.exiv2.org/projects/exiv2/wiki/The_Metadata_in_JPEG_files + case CODE_APP1: + { + const size_t EXIF_SIG_LEN=6; // Exif\0\0 + const size_t XMP_SIG_LEN=29; // http://ns.adobe.com/xap/1.0/\0 + + if(segment_lengthwidth); - height=endian_to_ushort(true, body->height); - } + info.width=endian_to_ushort(true, body->width); + info.height=endian_to_ushort(true, body->height); + } return; }; - reader.seek(segment_base+endian_to_ushort(true, head->length), SEEK_SET); + reader.seek(segment_base + segment_length, SEEK_SET); } throw Exception(IMAGE_FORMAT, &origin_string, "broken JPEG file - size frame not found"); } +static bool parse_tiff_IFD(bool is_big, Measure_reader& reader, Measure_info &info) { + const char* buf; + if(reader.read(buf, sizeof(JPG_Exif_IFD_begin))directory_entry_count); + for(int i=0; itag); + + if(entry_tag == 256 || entry_tag == 257){ + ushort entry_format=endian_to_ushort(is_big, entry->format); + if(entry_format != 3 && entry_format != 4 || endian_to_uint(is_big, entry->components_count) != 1) + return false; + uint value = (entry_format == 3) ? endian_to_ushort(is_big, entry->value_or_offset_to_it) : endian_to_uint(is_big, entry->value_or_offset_to_it); + (entry_tag == 256) ? info.width=value : info.height=value; + if(info.width && info.height) + return true; + } + } + + return false; + // then goes: LLLLLLLL Offset to next IFD [not going there] +} + +static void measure_tiff(const String& origin_string, Measure_reader& reader, Measure_info &info) { + const char* buf; + + if(reader.read(buf, sizeof(JPG_Exif_TIFF_header))byte_align_identifier, "II", 2)!=0 && strncmp(head->byte_align_identifier, "MM", 2)!=0) + throw Exception(IMAGE_FORMAT, &origin_string, "not TIFF file - wrong signature"); + + bool is_big=head->byte_align_identifier[0]=='M'; // [M]otorola vs [I]ntel + + if(endian_to_ushort(is_big, head->signature) != 42) + throw Exception(IMAGE_FORMAT, &origin_string, "not TIFF file - wrong signature"); + + reader.seek(endian_to_uint(is_big, head->first_IFD_offset), SEEK_SET); + if(!parse_tiff_IFD(is_big, reader, info)) + throw Exception(IMAGE_FORMAT, &origin_string, "broken TIFF file - size field entry not found"); +} + +static void measure_gif(const String& origin_string, Measure_reader& reader, ushort& width, ushort& height) { + + const char* buf; + const size_t head_size=sizeof(GIF_Header); + if(reader.read(buf, head_size)signature, "GIF", 3)!=0) + throw Exception(IMAGE_FORMAT, &origin_string, "not GIF file - wrong signature"); + + width=endian_to_ushort(false, head->width); + height=endian_to_ushort(false, head->height); +} + static void measure_png(const String& origin_string, Measure_reader& reader, ushort& width, ushort& height) { const char* buf; @@ -675,18 +771,41 @@ static void measure_png(const String& or height=endian_to_ushort(true, head->height); } +static void measure_bmp(const String& origin_string, Measure_reader& reader, ushort& width, ushort& height) { + + const char* buf; + const size_t head_size=sizeof(BMP_Header); + if(reader.read(buf, head_size)signature, "BM", 2)!=0) + throw Exception(IMAGE_FORMAT, &origin_string, "not BMP file - wrong signature"); + + reader.seek(0, SEEK_END); + if((uint)reader.tell() != endian_to_uint(false, head->file_size)) + throw Exception(IMAGE_FORMAT, &origin_string, "not BMP file - length header and file size do not match"); + + width=endian_to_ushort(false, head->width); + height=endian_to_ushort(false, head->height); +} + // measure center -static void measure(const String& file_name, Measure_reader& reader, ushort& width, ushort& height, Value** exif) { +static void measure(const String& file_name, Measure_reader& reader, Measure_info &info) { const char* file_name_cstr=file_name.taint_cstr(String::L_FILE_SPEC); if(const char* cext=strrchr(file_name_cstr, '.')) { cext++; if(strcasecmp(cext, "GIF")==0) - measure_gif(file_name, reader, width, height); - else if(strcasecmp(cext, "JPG")==0 || strcasecmp(cext, "JPEG")==0) - measure_jpeg(file_name, reader, width, height, exif); + measure_gif(file_name, reader, info.width, info.height); + else if(strcasecmp(cext, "JPG")==0 || strcasecmp(cext, "JPEG")==0) + measure_jpeg(file_name, reader, info); else if(strcasecmp(cext, "PNG")==0) - measure_png(file_name, reader, width, height); + measure_png(file_name, reader, info.width, info.height); + else if(strcasecmp(cext, "BMP")==0) + measure_bmp(file_name, reader, info.width, info.height); + else if(strcasecmp(cext, "TIF")==0 || strcasecmp(cext, "TIFF")==0) + measure_tiff(file_name, reader, info); else throw Exception(IMAGE_FORMAT, &file_name, "unhandled image file name extension '%s'", cext); } else @@ -695,69 +814,69 @@ static void measure(const String& file_n // methods -#ifndef DOXYGEN -struct File_measure_action_info { - ushort* width; - ushort* height; - Value** exif; -}; -#endif static void file_measure_action(struct stat& /*finfo*/, int f, const String& file_spec, void *context) { - File_measure_action_info& info=*static_cast(context); - Measure_file_reader reader(f, file_spec); - measure(file_spec, reader, *info.width, *info.height, info.exif); + measure(file_spec, reader, *static_cast(context)); } static void _measure(Request& r, MethodParams& params) { Value& data=params.as_no_junction(0, "data must not be code"); - ushort width=0; - ushort height=0; Value* exif=0; + Value* xmp=0; + Measure_info info={ 0, 0, 0, 0, &pa_UTF8_charset }; + + if(params.count()>1) + if(HashStringValue* options=params.as_hash(1, "methods options")) { + int valid_options=0; + for(HashStringValue::Iterator i(*options); i; i.next() ){ + String::Body key=i.key(); + Value* value=i.value(); + if(key == "exif") { + if(r.process(*value).as_bool()) + info.exif=&exif; + valid_options++; + } + if(key == "xmp") { + if(r.process(*value).as_bool()) + info.xmp=&xmp; + valid_options++; + } + if(key == "xmp-charset") { + info.xmp_charset=&pa_charsets.get(value->as_string()); + valid_options++; + } + } + if(valid_options!=options->count()) + throw Exception(PARSER_RUNTIME, 0, CALLED_WITH_INVALID_OPTION); + } + const String* file_name; - if((file_name=data.get_string())) { - File_measure_action_info info={ - &width, &height, - &exif - }; + + if(file_name=data.get_string()) { file_read_action_under_lock(r.absolute(*file_name), "measure", file_measure_action, &info); } else { VFile* vfile=data.as_vfile(String::L_AS_IS); file_name=&vfile->fields().get(name_name)->as_string(); - Measure_buf_reader reader( - vfile->value_ptr(), - vfile->value_size(), - *file_name - ); - measure(*file_name, reader, width, height, &exif); + Measure_buf_reader reader(vfile->value_ptr(), vfile->value_size(), *file_name); + measure(*file_name, reader, info); } - GET_SELF(r, VImage).set(file_name, width, height, 0, exif); + GET_SELF(r, VImage).set(file_name, info.width, info.height, 0, exif, xmp); } -#ifndef DOXYGEN -struct Attrib_info { - String* tag; ///< html tag being constructed - HashStringValue* skip; ///< tag attributes not to append to tag string [to skip] -}; -#endif -static void append_attrib_pair(HashStringValue::key_type key, HashStringValue::value_type value, Attrib_info* info) { - // skip user-specified, internal(starting with "line-") attributes and border attribute with empty value - if( - (info->skip && info->skip->get(key)) - || key.pos("line-")==0 - || (key=="border" && !value->is_defined()) - ) +static void append_attrib_pair(String &tag, String::Body key, Value* value){ + // skip border attribute with empty value + if(key=="border" && !value->is_defined()) return; // src="a.gif" width="123" ismap[=-1] - *info->tag << " " << key; + tag << " " << key; if(value->is_string() || value->as_int()>=0) - *info->tag << "=\"" << value->as_string() << "\""; + tag << "=\"" << value->as_string() << "\""; } -static void _html(Request& r, MethodParams& params) { +static void _html(Request& r, MethodParams& params) { String tag; tag << "for_each(append_attrib_pair, &info); + for(HashStringValue::Iterator i(*attribs); i; i.next() ) + append_attrib_pair(tag, i.key(), i.value()); } else throw Exception(PARSER_RUNTIME, 0, "attributes must be hash"); } } - { - Attrib_info info={&tag, attribs}; - fields.for_each(append_attrib_pair, &info); + for(HashStringValue::Iterator i(fields); i; i.next() ){ + String::Body key=i.key(); + // skip user-specified attributes + if(attribs && attribs->get(key)) + continue; + // allow only html attributes (to exclude exif, line-*) + if(key=="src" || key=="width" || key=="height" || key=="border") + append_attrib_pair(tag, key, i.value()); } + tag << " />"; r.write(tag); } @@ -1210,7 +1335,8 @@ static void _pixel(Request& r, MethodPar MImage::MImage(): Methoded("image") { // ^image:measure[DATA] - add_native_method("measure", Method::CT_DYNAMIC, _measure, 1, 1); + // ^image:measure[DATA; $.exif(false) $.xmp(false) $.xmp-charset[UTF-8] ] + add_native_method("measure", Method::CT_DYNAMIC, _measure, 1, 2); // ^image.html[] // ^image.html[hash]