lib-mail: message-parser: Fixed Clang 6.0 compiler warning. Warning was: index-mail.c:1182:3: warning: arithmetic on a null pointer treated as a cast from integer to pointer is a GNU extension [-Wnull-pointer-arithmetic] message_parser_parse_body(data->parser_ctx, ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ../../../src/lib-mail/message-parser.h:95:28: note: expanded from macro 'message_parser_parse_body' (void *)((char *)context + CALLBACK_TYPECHECK(callback, \ ~~~~~~~~~~~~~~~ ^

lib-mail: return void from message_parser_deinit()

lib-mail: Added message_parser_deinit_from_parts() The returned error strings could still be improved to give more details, especially give an approximate offset of where the problem is.

lib-mail: Added comments to message-parser.h

lib-mail: Moved struct message_part to a separate message-part.h

Removed CONTEXT_TYPE_SAFETY macro and reimplemented its functionality better. gcc/clang now gives a compiler error in many places if callback isn't exactly what was expected. It's also now much easier to add more of these checks.

Marked functions parameters that are allowed to be NULL. Some APIs were also changed. The non-obvious APIs where NULL parameter was changed to "" are master_service_init() and auth_master_user_list_init(). These checks can currently be enabled only on a patched clang: http://llvm.org/bugs/show_bug.cgi?id=6786

message parser: Added MESSAGE_PARSER_FLAG_INCLUDE_BOUNDARIES flag.

message parser: Added MESSAGE_PARSER_FLAG_INCLUDE_MULTIPART_BLOCKS. Patch by Stephan Bosch.

Changed message_parser_deinit() to return -1 if the parser was using preparsed broken message parts. Callers catch the error and mark the cache file corrupted. --HG-- branch : HEAD

Changed message_parser_deinit() to return -1 if the parser was using preparsed broken message parts. Callers catch the error and mark the cache file corrupted. --HG-- branch : HEAD

Removed message_parser_set_crlfs(). It's no longer needed and it might not work as wanted anyway. --HG-- branch : HEAD

Changed .h ifdef/defines to use <NAME>_H format. --HG-- branch : HEAD

By default assume MIME message if Content-Type: exists even if Mime-Version: doesn't. MESSAGE_PARSER_FLAG_MIME_VERSION_STRICT disables this. --HG-- branch : HEAD

Header parser has now flags parameter to tell it how to handle linefeeds. Changed message parser's boolean parameter to enum as well. --HG-- branch : HEAD

Removed message_parse_from_parts(). Added message_parser_init_from_parts() instead. Searching code now uses it whenever possible. --HG-- branch : HEAD

Type safe callbacks weren't as easy as I thought. Only callback(void *context) can be handled generically. Others can be handled specially, but only if all the parameters are pointers, otherwise eg. int parameter can be replaced with long without compiler giving any warnings. --HG-- branch : HEAD

Added context parameter type safety checks for most callback APIs. --HG-- branch : HEAD

Removed unused flag and added comment. --HG-- branch : HEAD

Some fixes and added message_parser_set_crlfs(). --HG-- branch : HEAD

Added message_part to struct message_block and some cleanups. --HG-- branch : HEAD

Rewrote the message bodystructure parser to allow parsing from non-blocking streams. Also did a couple of API changes and cleanups. --HG-- branch : HEAD

deinit, unref, destroy, close, free, etc. functions now take a pointer to their data pointer, and set it to NULL. This makes double-frees less likely to cause security holes. --HG-- branch : HEAD

Added "bool" type and changed all ints that were used as booleans to bool. --HG-- branch : HEAD

s/occured/occurred/ --HG-- branch : HEAD

message_parse_header_next() can now return "need more data" with nonblocking input streams. --HG-- branch : HEAD

Added struct message_header_line.middle and middle_len to contain the ':' in header and whitespace around it. --HG-- branch : HEAD

Added skip_initial_lwsp parameter to message_parse_header_init(). --HG-- branch : HEAD

importing new index code. mbox still broken. --HG-- branch : HEAD

Message parsing can now be done in two parts - header and body. We're now more eager at parsing MIME structure for message. It's done whenever body has to be fully read (eg. rfc822.size). --HG-- branch : HEAD

If BODY/BODYSTRUCTURE is requested with some other headers, parse the headers only once. If body contains multiple MIME parts, cache the internal MIME structure so BODY[part] fetching doesn't need to parse it again. --HG-- branch : HEAD

Ignore Content-* headers if there's no MIME-Version header. Note that this change breaks cached message parts in indexes. --HG-- branch : HEAD

IMAP protocol doesn't allow server to send NULs to client. Send ascii #128 instead. --HG-- branch : HEAD

Better handling for multiline headers. Before we skipped headers larger than input buffer size (8k with read (default), 256k with mmap). The skipping was also a bit buggy. Now we parse the lines one at a time. There's also a way to read the header fully into memory before parsing it, if really needed. --HG-- branch : HEAD

Function typedefs now define them as functions, not function pointers. --HG-- branch : HEAD

mail-storage.h interface changes, affects pretty much everything. FETCH, SEARCH, SORT and THREAD handling were pretty much moved from lib-storage/ to imap/ so adding non-index storages would be much easier now. Also POP3 server can now be easily implemented with lib-storage. Not too well tested, and at least one major problem: partial fetching is _slow_. --HG-- branch : HEAD

Naming change for function typedefs. --HG-- branch : HEAD

Naming style changes, finally got tired of most of the typedefs. Also the previous enum -> macro change reverted so that we don't use the highest bit anymore, that's incompatible with old indexes so they will be rebuilt. --HG-- branch : HEAD

Use unsigned char* when accessing non-NUL terminating strings. Compiler warnings would then notify about accidentally passing them to functions which require them NUL-terminated. Changed a few functions to use void* to avoid unneeded casting. --HG-- branch : HEAD

Rewrote rfc822-tokenize.c to work one token at a time so it won't uselessly take memory, maybe also a bit faster. This caused pretty large changes all around. Also moved all string (un)escaping code to lib/strescape.c. --HG-- branch : HEAD

Renamed IBuffer and OBuffer to IStream and OStream which describes their functionality better. I tried to keep the variable names and comments also sensible. --HG-- branch : HEAD

Split IOBuffer into mmaped IBuffer, file IBuffer, memory data IBuffer and file OBuffer. --HG-- branch : HEAD

changed all "temporary memory pool" references to "data stack" --HG-- branch : HEAD

message_parse_header() now calls the function with empty name/value at end of headers. we use this feature with mbox to allow usage of Content-Length header field. --HG-- branch : HEAD

move size_t fixes --HG-- branch : HEAD

Added support for caching of MessagePart data. This is useful for fetching body[part]s. Also BODY and BODYSTRUCTURE can be generated fast using it. Also fixed index corruption in some situations when more cached data was added to index. --HG-- branch : HEAD

there was no need for MessagePart->pos.virtual_pos, so removed it. --HG-- branch : HEAD

Introduced uoff_t which is the unsigned-equilevant of off_t. This was needed to be able to handle off_t overflows properly. Also changed a few unsigned int fields into uoff_t so we should now support >2G mails if uoff_t is 64bit. Also fixed several potential integer overflows. --HG-- branch : HEAD

s/user_data/context/ and some s/Data/Context/ --HG-- branch : HEAD

fixes, seems to be somewhat working now. --HG-- branch : HEAD

Message data is parsed in blocks (no longer entirely mmap()ed). Several IOBuffer changes. All mixed signed/unsigned comparisions were fixed so code can now be compiled with gcc's -W flag. mbox support is broken currently, and there's most likely several other problems too. --HG-- branch : HEAD

Initial revision --HG-- branch : HEAD