aboutsummaryrefslogtreecommitdiffstats
path: root/camel/camel-mime-part-utils.c
Commit message (Collapse)AuthorAgeFilesLines
* And same here ...NotZed2000-05-051-1/+0
| | | | | | | | | | | | | | | | | | | | | | 2000-05-05 NotZed <NotZed@HelixCode.com> * camel-folder-summary.c: And same here ... * camel-mime-utils.c: Defined out some memory profiling stuff I left there by mistake. * camel-mime-parser.c (folder_scan_content): Apply the fix from the header scanner to here too. (folder_scan_header): Only check for end of header if we have space for it (didn't end the read with a newline) (folder_scan_header): inptr is the only real thing we need registerised for performance. Try to help the compiler be smart about it .. (folder_scan_header): Simplified the save header case a tad. Commented out some memory profiling stuff. svn path=/trunk/; revision=2811
* No, we're not going to have g_strcasecmp for no good reason,NotZed2000-05-041-4/+4
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | not even if its slipped in with no changelog. 2000-05-04 NotZed <NotZed@HelixCode.com> * providers/mbox/camel-mbox-summary.c: Yes, and anotherone. * camel-mime-utils.c: And another one. * camel-mime-part.c: And another one. * camel-mime-part-utils.c: And another one. * camel-folder-search.c: And another one. * camel-mime-parser.c: Reverted a change wihtout a ChangeLog entry. 2000-05-04 NotZed <NotZed@HelixCode.com> * camel-folder-summary.[hc]: Yes, CamelFolderSummary is back ... ... re-usable class to summarise and index any stream or message and to manage/load/save the created summaries. * camel-folder.c: Include string.h to kill a warning. 2000-05-03 NotZed <NotZed@HelixCode.com> * camel-folder.h: Added pos/bodypos/endpos to the basic message content info object. Size to be removed? Moved the messageconentinfo and messageinfo back to camel-folder-summary.h. * camel-mime-filter-index.c (camel_mime_filter_index_set_ibex): New function to (re)set the index to use on a filter. * camel-mime-parser.c (camel_mime_parser_scan_from): Whole bunch of inline docs. (camel_mime_parser_drop_step): New function to drop a state from the parser. Needs more testing. svn path=/trunk/; revision=2789
* s/strcasecmp/g_strcasecamp/ everywhere except intl/, per michael'sJacob Leach2000-05-031-4/+5
| | | | | | request. svn path=/trunk/; revision=2776
* > * gmime-utils.[ch]: What the hell, remove it. This will break theMichael Zucci2000-05-031-3/+2
| | | | | | | | | | | | | | | | | | | | | | | | | | | > nntp provider (but its broken anyway). The mime parser can be > used instead though. > Removed from all code including it (but none were using it). > > * gmime-utils.c (_store_header_pair_from_string): Removed bizarre > string_dichotomy version of this. This code is somewhat redundant > now, and is headed for death anyway. > > * gstring-util.c (g_string_dichotomy): Same with this one. > (g_string_clone): Removed a memory leak, g_string_new() allocates > its own memory. > (g_string_append_g_string): Allow to append an empty gstring onto > another gstring, dont abort()! > > * string-utils.c (string_dichotomy): Removed this incredibly weird > function. > > * camel-folder.c (_create): Replaced the rather obtuse use of > "string_dichotomy" function with a simple strrchr(). Still not > sure it'll work. > > * camel-folder-summary.c: cvs removed a long-removed file. svn path=/trunk/; revision=2753
* Dont use autofill on these fucking long function anmes!!!!!!22000-05-031-5/+3
| | | | | | | | | | 2000-05-02 <notzed@helixcode.com> * camel-mime-part-utils.c (simple_data_wrapper_construct_from_parser): Dont use autofill on these fucking long function anmes!!!!!! svn path=/trunk/; revision=2750
* camel_mime_parser_tell() returns an offset from where it started parsing,Dan Winship2000-04-291-12/+21
| | | | | | | | | | | | | | | | | | | * camel-mime-part-utils.c (simple_data_wrapper_construct_from_parser): camel_mime_parser_tell() returns an offset from where it started parsing, not necessarily from the start of data. Since we're parsing a bounded seekable_stream, we need to add the stream's starting bound to camel_mime_parser_tell's return value to create the substream in the right place. * camel-seekable-substream.c (camel_seekable_substream_new_with_seekable_stream_and_bounds): say CAMEL_STREAM_UNBOUND rather than -1 in doc. * camel-seekable-stream.c (camel_seekable_stream_seek): Add more info to docs. svn path=/trunk/; revision=2680
* Remove. We weren't using the fields that made this different fromDan Winship2000-04-271-3/+2
| | | | | | | | | | | | | * camel-mime-body-part.[ch]: Remove. We weren't using the fields that made this different from camel-mime-part, so it basically just forced us to do lots of gratuitous typecasting. * camel-multipart.[ch]: Use CamelMimePart. Remove the multipart parent stuff, since we weren't using that either. * etc: update for CamelMimeBodyPart -> CamelMimePart svn path=/trunk/; revision=2645
* Big cleanup of camel-stream-*, got rid of 3 classes, improved the interfaces,NotZed2000-04-261-7/+3
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | and fixed at least one problem (end of stream never happening in certain cases). Things that can fail now have a way of saying they failed too. So much for taking ANZAC day off to get drunk! 2000-04-26 NotZed <NotZed@HelixCode.com> * camel-seekable-substream.c (stream_seek): Changed to have absolute seek semantics, not relative to the bounds. * camel-seekable-stream.c (reset): When we reset, seek to the start of the bound, if there is one. (stream_tell): Make tell virtual. * camel-stream-filter.c (do_available): Removed. * camel-stream-buffer.c: Remove leading _'s from static functions. (stream_read): Renamed from read(). Fancy that conflicting! (my boo!) Others too. * providers/pop3/camel-pop3-folder.c (get_message_by_number): Changed to stream_mem interface. * providers/mbox/camel-mbox-folder.c (_get_message_by_uid): Fixed for streamfs interface changes, and implement a failure case. (_append_message): Changed for fs stream interface change. * camel-multipart.c (print_part): Iterate rahter than callback. I hate glists's interface (hence, move this to write_to_stream). (write_to_stream): Return an error (yuck, this is a royal PITA to do with the stream write interface). * camel-mime-message.c: Removed leading _ from static names. * camel-mime-part.h: construct_from_parser() now returns an error code. * camel-mime-part-utils.c (camel_mime_part_construct_content_from_parser): Changed to use a camel-data-wrapper instead of a camel-simple-data-wrapper (no change needed elsewhere?). (simple_data_wrapper_construct_from_parser): Fixes for stream-mem interface changes. * camel-simple-data-wrapper.[ch], camel-simple-data-wrapper-stream.[ch], camel-stream-data-wrapper.[ch], removed. Fixed including of these files. * camel-mime-part.c (camel_mime_part_set_text): Remove the use of the camel-simple-data-wrapper-stream, just use a mem stream. (write_to_stream): Renamed from my_* (construct_from_stream): Return an error on error. * camel-stream-mem.c (camel_stream_mem_new*): Remove mode parameter. * camel-stream-mem.h (enum CamelStreamMemMode): Removed. It wasn't used at all. * camel-data-wrapper.h: Add camel_data_wrapper_new() to create these. (write_to_stream, construct_from_stream): Return an error indicator for success. Fixed all methods to match (ICK). * Makefile.am (libcamel_la_SOURCES): Remove camel-simple-data-wrapper.c, camel-simple-data-wrapper-stream.c, camel-stream-data-wrapper.c. Obsoleted by code re-use! * camel-data-wrapper.c (construct_from_stream): Change the default implementation to just set the output stream == construction stream. Well, this lets me get rid of both simple-data-wrapper and stream-data-wrapper (unused anyway), and simple-data-wrapper-stream in one hit. CamelDataWrapper is now also a concrete class. (write_to_stream): Use camel_stream_write_to_stream() to calculate/return values (and save code). Include <errno.h> for obvious reasons. * camel-stream.c (eos): Provide a default implementation of .eos(). (camel_stream_write_to_stream): Make it return an error code on error. (camel_stream_printf): Changed to return the number of bytes written/error. (camel_stream_available): Removed. * camel-stream-fs.h (enum CamelStreamFsMode): Removed. Changed to use unix modes and so forth (wasn't used for anything but new file creation and didn't work well either). * camel-stream-fs.c: Removed leading _'s for names. And removed some virtual method 'documentation'. (destroy): Dont try and close a closed/error fd. Only report error if close returns -1. Moved all the code to finalise(), and killed this function. (init_with_fd): Properly setup the seek offset, if it is a valid and seekable file descriptor. (init_with_fd_and_bounds): Use off_t for bounds, set bounds on the seekable stream. (init_with_name): Return error codes. (init_with_name_and_bounds): Ditto. (camel_stream_fs_new_with_name): REturn NULL object if it failed. (camel_stream_fs_new_with_name_and_bounds): Return NULL object on failure. Changed with_name* api's to take unix open style args and flags. (read): The bounded stream bounds checking seemed off, simplified code a bit. (write): Implement bounds checking for writing, the comment was wrong, it could make sense to bound writing. Cleaned up a little. (available): Gone. (eos): Removed. Use CamelStream's implementation now. (close): Reset the fd to -1, provide a warning for bad usage. (seek): Cleaned up. Changed the behaviour a little, the returned offset is the absolute position in the file, even in bounded streams. (seek): Seek from end mirrors lseek() behaviour (reverse seeking). 2000-04-25 NotZed <NotZed@HelixCode.com> * camel-stream-fs.h (struct _CamelStreamFs): Moved bounds and eof indicator to other parent classes. * camel-stream.c (camel_stream_printf): New utility function. Obvious use. * camel-stream-mem.c: Removed leading _'s from static func's. (camel_stream_mem_new_with_byte_array): Fixed for api changes, set the owner for the byte array to us. : Removed A bunch of gtk doc stuff for static (implementation) functions. (available): Removed. (write): Fixed the write implementation so that seek() works on a seekable memory stream, as expected. Seeking past the end of the buffer has unix semantics (filling with 0). (available): Removed. (write): Implement seekable stream bounded stream. (read): Implement seekable stream bounded stream. (close): Dont free the stream_mem if we're not the owner. (seek): Allow to seek beyond the end of memory area, implement bounds checking. (seek): Set errno on bad policy. * camel-stream-mem.h (struct _CamelStreamMem): Changed position to off_t. (new_with_buffer): Changed len to be a size_t. (set_buffer, set_byte_array): New interface functions. (struct _CamelStreamMem): Removed position, it is stored in the superclass. * camel-stream.h: Removed some of the seemingly random whitespace. Removed the available method (its not impelemented/useful enough). * camel-seekable-substream.c (init_with_seekable_stream_and_bounds): Remove the data_available stuff, it hasn't been properly implemented/finished, and may never work (unfortunately *sigh). (reemit_parent_signal): Removed part of the above change. (set_bounds): Removed (moved to seekable-stream). : Fixed up some of the generally unreadable indenting (sorry, wrapping at 80 characters with camels_really_long_function_names() just_doesnt_work_very_well_does_it(). (available): Removed. (stream_seek): Fixup for object changes. Make sure we return -1 if the parent stream can't seek. * camel-seekable-stream.c (ccamel_seekable_stream_set_bounds): New function to bound any seekable stream. : Removed _'s. (camel_seekable_stream_class_init): Implement an init function, to setup the stream bounds to unbound. * camel-seekable-stream.h (CamelSeekableStreamClass): New virtual method set_bounds for seekable streams. (CAMEL_STREAM_UNBOUND): New define for no bound. * camel-seekable-substream.h (struct _CamelSeekableSubstream): Removed sup_bound and inf_bound, moved to CamelSeekableStream (and renamed, and changed to off_t's). (new_with_seekable_stream_and_bounds): Use off_t as the bounds. (CamelSeekableSubstreamClass): Uh, why was the intialiser virtual? Removed. * camel-seekable-stream.[ch] (CamelSeekableStreamClass): Changed seek to accept an off_t as the offset. (struct _CamelSeekableStream): Renamed cur_pos to position and changed it to an off_t type. (enum CamelStreamSeekPolicy): Set to match the SEEK_* constants from lseek(). (get_current_position): Renamed to tell(). * camel-stream-buffer.h: Commented out set_vbuf - never implemented. svn path=/trunk/; revision=2624
* fix a cut-and-pasto.Dan Winship2000-04-251-1/+1
| | | | | | | * camel-mime-part-utils.c (simple_data_wrapper_construct_from_parser): fix a cut-and-pasto. svn path=/trunk/; revision=2587
* Big fixes for the last commit which broke the message creation entirely.Michael Zucci2000-04-241-3/+5
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Plus an address decoder/formatter, etc. * camel-mime-message.c (construct_from_parser): Allow MESSAGE_END _or_ EOF as valid termination conditions. * providers/mbox/camel-mbox-summary.c (message_struct_new): Decode and then re-encode the addresses, so they are consistently formatted. * camel-mime-utils.c (header_decode_mailbox): Store the address in a _header_address. And try to get a comment-stored name if there is one. (header_decode_address): Actually return an address. (header_to_decode): Renamed to header_address_decode() (header_mailbox_decode): New function to get a single mailbox. (header_mime_decode): Return the major/minor value, as appropriate. (header_address_new, and friends): Whole bunch of utility functions for working with the address thingies. (header_decode_domain): Free the string header, and dont expand '.' into ' . '. (camel_mime_part_construct_content_from_parser): Oops, this was totally screwed up, try creating the right cotnent on the right object. svn path=/trunk/; revision=2579
* No longer call data_wrapper_repository_init.NotZed2000-04-241-117/+162
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | 2000-04-24 NotZed <NotZed@HelixCode.com> * camel.c (camel_init): No longer call data_wrapper_repository_init. * camel-medium.c (write_to_stream): Moved (back) to camel-mime-part. (add_header): (set_header): (remove_header): (get_header): Make all these abstract, and spit warnings if called. I guess it could manage the list, but well, it doesn't. * camel-medium.h (struct _CamelMedium): Dont store headers here, the implementor is the only one who knows their format. (CamelMediumClass): Changed header values to be void *'s. They need not be strings? * camel-simple-data-wrapper.c (construct_from_stream): And we're back. Set the output stream. (construct_from_parser): Moved to camel-mime-part-utils. * camel-mime-part-utils.c (camel_mime_part_construct_content_from_parser): Create the contents of multipart and simple messages. * camel-multipart.c (construct_from_parser): Moved to camel-mime-part-utils. (separate_part): Removed. * camel-mime-part.c (construct_from_stream): Back again! This now switches over to using a mime parser for any mime parts, only. (my_write_to_stream): Write our headers and so forth here. (add_header): Add header directly, parent class is abstract. (remove_header): Ditto. (set_header): Ditto. * camel-data-wrapper.c (camel_data_wrapper_construct_from_stream): Remade abstract. (camel_data_wrapper_construct_from_parser): Moved to camel_mime_part. * camel-data-wrapper.h: Put back construct_from_stream. * camel-mime-part.h: Put construct_from_parser in here, the data-wrapper shouldn't know about mime. Ok, so now to undo half of the last hours changes ... duh. 2a51,93 * providers/pop3/camel-pop3-folder.c (get_message_by_number): Use construct_from_stream instead of set_input_stream(). * camel-simple-data-wrapper-stream.c (camel_simple_data_wrapper_stream_construct): REmoved the destroy callback code. (wrapper_destroy_cb): Removed. * camel-simple-data-wrapper.h: Add prototype for _construct() method. * camel.c: Include unicode.h to kill a warning. * camel-data-wrapper.h (CameldataWrapperClass): Removed construct_from_stream virtual method. Removed get/set input stream. * data-wrapper-repository.[ch]: Removed&from build. Obsoleted? The justification as is follows: It is mixing storage protocol/format with message architecture. It really just doesn't serve any purpose, as each medium implementor will have to have its own type->handler mapping, and the only current implementor, mimepart has a very simple structure and no need for this. * camel-medium.c (write_to_stream): Moved here from most of the stuff in camel-mime-part. Well, the MEDIUM is the one that knows what the headers are, and the content is, let it write it out. * camel-mime-part-utils.c (camel_mime_part_construct_content): Copied from camel-mime-part.c, removed handling of message followon state (moved to camel-mime-message). (camel_mime_part_construct_content_from_parser): Renamed from construct_content. (camel_mime_part_construct_headers_from_stream): (camel_mime_part_construct_content_from_stream): (camel_mime_part_store_stream_in_buffer): Removed. Replaced by the new construct from parser stuff. * camel-mime-message.c (construct_from_parser): Do construct_from_parser for mime-message. (_write_to_stream): Set the mime-version header for medium to write out, rather than writing it out ourselves. 4a96,102 (construct_from_stream): Removed. (camel_data_wrapper_construct_from_stream): Changed to a helper function, creates a mime_parser, and constructs from that. (set_input_stream): Removed. (camel_data_wrapper_set_input_stream): Removed. (get_input_stream): Removed. (camel_data_wrapper_get_input_stream): Removed. 11a110,111 (_get_message_by_uid): Use construct_from_stream() instead of creating our own parser. 16a117,131 (camel_mime_part_construct_content): Removed to camel-mime-part-utils.c (my_get_output_stream): Removed. The streeam is in the data-wrapper. (my_get_content_object): Removed. The content object is stored in the medium. If none is there, the object wasn't created properly. (my_write_content_to_stream): Removed. The content object is the one that knows how to write itself out!!!!!!!! (my_write_to_stream): Remove the base header writing stuff - has been moved to camel-medium, where it belongs. This can just be used to check for mandatory headers. (my_construct_from_stream): Removed. (my_set_input_stream): What the hell, i'll remove this too. Nobody seems to understand how it differs from create from stream, and they both seem to serve the same purpose ... 19a135,136 (construct_from_stream): Removed! Job taken over by construct_from_parser. 24a142 (set_input_stream): REmoved. Replaced by construct_from_parser. svn path=/trunk/; revision=2577
* kill camel-logDan Winship2000-04-191-25/+0
| | | | svn path=/trunk/; revision=2487
* fix the name of "Helix Code, Inc." in all the copyrightsDan Winship2000-03-101-1/+1
| | | | svn path=/trunk/; revision=2091
* Make sure we open with create with a creation mask.NotZed2000-03-041-2/+0
| | | | | | | | | | | | | | | 2000-03-03 NotZed <NotZed@HelixCode.com> * providers/mbox/camel-mbox-utils.c (camel_mbox_write_xev): Make sure we open with create with a creation mask. 2000-03-01 NotZed <NotZed@HelixCode.com> * camel-mime-part-utils.c (camel_mime_part_construct_content_from_stream): DO NOT assert on content type, we have fallback code 4 lines below it ... *sigh* svn path=/trunk/; revision=2026
* Free the data in the headers hash table. (_add_header): g_strdup theDan Winship2000-02-241-4/+5
| | | | | | | | | | | | | * camel-medium.c (_finalize): Free the data in the headers hash table. (_add_header): g_strdup the header name and value when adding it. * camel-mime-part-utils.c (camel_mime_part_construct_headers_from_stream): Free the header data after calling camel_medium_add_header, since it will have g_strdup()ed it itself. svn path=/trunk/; revision=1913
* don't seek the begining of the substream. (_eos): fix eos conditionbertrand2000-02-171-0/+1
| | | | | | | | | | | | | | | | | | | | | | 2000-02-16 bertrand <Bertrand.Guiheneuf@aful.org> * camel/camel-seekable-substream.c (_set_bounds): don't seek the begining of the substream. (_eos): fix eos condition testing. (_finalize): unref parent stream (_init_with_seekable_stream_and_bounds): ref parent stream * camel/gstring-util.c (g_string_equal_for_hash): (g_string_equal_for_glist): return type is int. * camel/camel.h: * camel/camel.c (camel_init): use (void) instead of (). A lot of other small changes to make the set_input_stream scheme work. It actually works. svn path=/trunk/; revision=1798
* this routine replaces the _read_part routine and does not store the partbertrand2000-02-161-4/+20
| | | | | | | | | | | | | | | | | | | | | | | | 2000-02-15 bertrand <bertrand@helixcode.com> * camel/camel-multipart.c (_localize_part): this routine replaces the _read_part routine and does not store the part in a buffer. (_set_input_stream): use the set_input_stream instead of the construct_from_stream. each bodypart is given an input stream. * camel/camel-mime-part-utils.c: include the data-wrapper-repository header. (camel_mime_part_construct_content_from_stream): use the set_input_stream instead of the construct_from_stream method. * camel/camel-seekable-substream.c (_set_bounds): cur position is set to 0 not to inf_bound. Sync svn path=/trunk/; revision=1790
* If we get a -1 read, DONT update the total bytes, and try and truncate theNotZed2000-02-101-8/+11
| | | | | | | | | | | | | | | | 2000-02-09 NotZed <notzed@zedzone.helixcode.com> * camel/camel-mime-part-utils.c (camel_mime_part_store_stream_in_buffer): If we get a -1 read, DONT update the total bytes, and try and truncate the array in reverse. Eek. * camel/camel-mime-part.c (camel_mime_part_encoding_from_string): This was DEFINETLEY not the right way to do it. strncmp!=strcmp (camel_mime_part_encoding_to_string): Handle the default case. : include string.h for strcmp() etc. svn path=/trunk/; revision=1720
* Updated my e-mail address to reflect my new employmentBertrand Guiheneuf2000-01-041-1/+1
| | | | svn path=/trunk/; revision=1533
* Fix copyright informations once againBertrand Guiheneuf1999-12-151-1/+1
| | | | svn path=/trunk/; revision=1489
* Copyright changeBertrand Guiheneuf1999-10-231-1/+4
| | | | svn path=/trunk/; revision=1348
* Address updateBertrand Guiheneuf1999-08-301-1/+1
| | | | svn path=/trunk/; revision=1147
* text parameter declared constbertrand1999-08-291-10/+13
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | 1999-08-28 bertrand <Bertrand.Guiheneuf@aful.org> * camel/camel-mime-part.c (camel_mime_part_set_text): text parameter declared const * camel/camel-mime-part-utils.c (camel_mime_part_store_stream_in_buffer): actually test correctly nb_bytes_read_chunk is >0 * camel/gstring-util.c: * camel/gmime-content-field.c: * camel/providers/MH/camel-mh-folder.c: * camel/camel-stream-fs.c: include string.h * camel/camel-stream-mem.c (_write): return the numver of written bytes. * camel/camel-stream-buffered-fs.c (_eos): return sthg * camel/camel-stream.c (default_camel_seek): return something. * Started to work on new recipient code. * fixed various leaks. svn path=/trunk/; revision=1146
* contruct the content from the buffer before calling CamelMediumbertrand1999-08-261-0/+31
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | 1999-08-26 bertrand <Bertrand.Guiheneuf@aful.org> * camel/camel-mime-part.c (_get_content_object): contruct the content from the buffer before calling CamelMedium implementation. (_construct_from_stream): Do not construct the content by default, just store the content bytes in a temporary buffer. Content will be constructed only at caller request (when calling CamelMedium::get_content_object) Providers with better access to the messages (mbox/MH ...) will have to provider lighter implementation, that is shall not read content at all unless the caller asks for it (again with get_content). * camel/camel-mime-part-utils.c: new file, groups mime-part related utils. Meant to be used by providers subclassing MimeMessage. (camel_mime_part_construct_headers_from_stream): (camel_mime_part_construct_content_from_stream): no more useless temporary hash table. * camel/camel-mime-part.c (_construct_from_stream): calls mime-part-utils functions now. * camel/gmime-utils.c (_store_header_pair_from_string): do not use hash table to store header, use an array instead. svn path=/trunk/; revision=1145
* new file, groups mime-part related utils. Meant to be used by providersbertrand1999-08-261-0/+134
1999-08-26 bertrand <Bertrand.Guiheneuf@aful.org> * camel/camel-mime-part-utils.c: new file, groups mime-part related utils. Meant to be used by providers subclassing MimeMessage. (camel_mime_part_construct_headers_from_stream): (camel_mime_part_construct_content_from_stream): no more useless temporary hash table. * camel/camel-mime-part.c (_construct_from_stream): calls mime-part-utils functions now. * camel/gmime-utils.c (_store_header_pair_from_string): do not use hash table to store header, use an array instead. * MimePart cleanup. Not complete yet. Content must not be parsed when constructing but only when content object is requested. Default implementation will keep content in a stream. svn path=/trunk/; revision=1144