Re-add camel-simple-data-wrapper.h, which was removed for some reason.

* camel.h: Re-add camel-simple-data-wrapper.h, which was removed
for some reason.

Added more files to the imap provider to use as an outline - will likely need heavy editing later in development

This is mostly meant as guidance to both me and any volunteers that start working on imap

Fixes to make the SMTP transport thread-safe

Wed May 24 18:09:26 EDT 2000  Jeffrey Stedfast  <fejj@helixcode.com>

* providers/smtp/camel-smtp-transport.[c,h]: Fixes to make the
SMTP transport thread-safe

Added debug fprintfs, tested with a few messages (smtp_data): Fixed to use

Tue May 23 17:49:21 EDT 2000  Jeffrey Stedfast  <fejj@helixcode.com>

* providers/smtp/camel-smtp-transport.c: Added debug fprintfs, tested with a few messages
(smtp_data): Fixed to use data_wrapper_write_to_stream()

* camel-mime-filter-smtp.c (filter): Modified to escape all lines
beginning with a '.' and to place \r before each \n if one did
not previously exist. Removed code to escape "From " as it was found to not be needed

* providers/imap/.cvsignore: added file

Shut up.

stuff

stuff

stuff dude

junk

changelog

more stuff

changelog

stuff

dude, just stuff

added camel-mime-filter-smtp.[c,h] and made mods to camel-mime-filter-from.c

>  searchpart = strchr(namepart, '?');
2000-05-19  NotZed  <NotZed@HelixCode.com>

        * camel-simple-data-wrapper.c (construct_from_stream): If we
        already have been constructed, unref our content.
        (write_to_stream): Check we've been constructued, and change for
        stream api changes.

        * camel-mime-parser.c: Removed exception stuff.

        * md5-utils.c (md5_get_digest_from_stream): repaired.

        * camel-mime-message.c: Remove exception from write_to_stream, and
        fix, and fix formatting.

        * providers/sendmail/camel-sendmail-transport.c (_send_internal):
        Fix for stream changes.

        * providers/pop3/camel-pop3-store.c (camel_pop3_command): Fixes
        for stream changes.

        * providers/mbox/camel-mbox-folder.c, and elsewhere, fix all
        stream api changes.
        (mbox_append_message): Use stream_close() now its back.
        (mbox_append_message): unref the from filter.

        * camel-stream-mem.c: And here.

        * camel-stream-fs.[ch]: Here too.

        * camel-stream-filter.c: Likewise.  This is getting tedious.

        * camel-stream-buffer.c (stream_write): Fix a few little problems.
        (stream_close): Reimplmeent.
        (camel_stream_buffer_read_line): Slightly more efficient version,
        that also only allocates the right amount of memory for strings.

        * camel-seekable-substream.c: Likewise.

        * camel-seekable-stream.[ch]: Remove exceptions, fix formatting,
        changes for stream (re)fixes.  set_bounds returns an error.

        * camel-stream.[ch]: Remove exceptions.  Make flush and reset return
        an error code, repair all the screwed up formatting, and put back
        close.

        * camel-mime-part-utils.c
        (camel_mime_part_construct_content_from_parser): And here.

        * camel-mime-part.c (camel_mime_part_set_content): And this too.
        (write_to_stream): Fixed for stream changes.

        * camel.h: Fixed.

        * providers/vee/camel-vee-folder.c (vee_search_by_expression):
        Implement.  Performs an intersection of the two searches.
        (camel_vee_folder_finalise): Unref search folders.
        (vee_append_message): Implement append.

remove message_number_capability and require uid capatibility.

* camel-folder.c: remove message_number_capability and require uid
capatibility.
(camel_folder_list_subfolders, camel_folder_get_uid_list,
camel_folder_get_subfolder_info, camel_folder_get_message_info):
removed
(camel_folder_get_subfolder_names,
camel_folder_free_subfolder_names): new subfolder interfaces.
(camel_folder_get_uids, camel_folder_free_uids): new uid
interfaces
(camel_folder_get_summary, camel_folder_free_summary): new summary
interfaces

* providers/mbox/camel-mbox-folder.c,
* providers/nntp/camel-nntp-folder.c:
* providers/vee/camel-vee-folder.c: Update for changes

* providers/pop3/camel-pop3-folder.c: Implement get_uids, update
for other changes.

add urls file

Guess!

2000-05-18  NotZed  <NotZed@HelixCode.com>

        * providers/vee/camel-vee-folder.c: Guess!

        * camel-folder-search.c (search_user_flag): Implement user_flag
        search term.

        * camel-folder-search.h: Added user_flag search capability
        (user-flag "blah")

        * providers/mbox/camel-mbox-folder.c (mbox_init): Set USER flag in
        permanent flags for the folder.

Remove unused async open/close and copy_message_to functions. Rename

* camel-folder.c: Remove unused async open/close and
copy_message_to functions.
Rename functions without initial _. Fix glib preconditions and
gtk-doc comments.

remove get/set_output_stream operations. They're redundant with

* camel-data-wrapper.c: remove get/set_output_stream operations.
They're redundant with write_to_stream, and CamelMimePart and
CamelMimeMessage only implement the latter, meaning that trying to
get_output_stream on a CamelMimeMessage that was built from pieces
rather than being parsed from a stream doesn't work. Anything that
uses get_output_stream can be rewritten to use write_to_stream, so
we'll standardize on that.
(camel_data_wrapper_new): remove this: CamelDataWrapper is
supposed to be an abstract class.
(write_to_stream): remove default implementation. (Moved to
CamelSimpleDataWrapper)

* camel-simple-data-wrapper.c: resurrect, although it's not really
the same thing it was before. A simple data wrapper, which is
backed by a CamelStream.

* camel-mime-part-utils.c
(simple_data_wrapper_construct_from_parser): Use
construct_from_stream rather than set_output_stream.
(camel_mime_part_construct_content_from_parser): Change
camel_data_wrapper_new to camel_simple_data_wrapper_new.

* camel-mime-part.c (camel_mime_part_set_content): Change
camel_data_wrapper_new to camel_simple_data_wrapper_new.

Quick fix to get it to compile. I hope I don't get into trouble.

* camel-folder-summary.c: (message_info_load):
Quick fix to get it to compile. I hope I don't get into trouble.

Don't include the no-longer-distributed possibly-to-be-removed headers.

* camel.h: Don't include the no-longer-distributed
possibly-to-be-removed headers.

fix an off-by-one error in address parsing (smtp_data): use

* providers/smtp/camel-smtp-transport.c
(smtp_get_email_addr_from_text): fix an off-by-one error in
address parsing
(smtp_data): use camel_data_wrapper_get_output_stream rather than
data_wrapper->output_stream

All this basically to support user flags in the summary. They are not yet

All this basically to support user flags in the summary.  They
are not yet saved to the message headers (complicates things a bit).

2000-05-17  NotZed  <NotZed@HelixCode.com>

        * providers/mbox/camel-mbox-folder.c (message_changed): Snoop
        changes to user flags on the message into the summary as well.

        * providers/mbox/camel-mbox-summary.c (camel_mbox_summary_init):
        Changed version init to include the parent class version info
        (i.e. add it not overwrite it).

        * camel-folder-summary.c (message_info_new): Initialise user_flags
        to empty.
        (message_info_load): And load user flags.
        (message_info_save): And save user flags.
        (message_info_free): And free them.
        (CAMEL_FOLDER_SUMMARY_VERSION): Bumped file revision.

        * camel-folder-summary.h: Added user-flags to summary.

        * camel-mime-message.c (camel_mime_message_set_user_flag): Dont
        use a hashtable for user flags.
        (camel_mime_message_get_user_flag): And changed here too.
        (camel_flag_get): New interface to get a flag from a flag
        list.  Flag lists are easier to work with than hash tables, and
        save memory too.
        (camel_flag_set): And set.
        (camel_flag_list_free): And free.
        (free_key_only): Discard.
        (finalize): Remove the flag list.

error checking on gethostbyaddr() in providers/smtp/camel-smtp-transport.c

Implement.

2000-05-16  NotZed  <NotZed@HelixCode.com>

        * providers/mbox/camel-mbox-folder.c (mbox_delete_message_by_uid):
        Implement.

2000-05-12  NotZed  <NotZed@HelixCode.com>

        * camel-movemail.c (camel_movemail): Open the destination with
        O_APPEND, so we dont blow away a partially transferred mbox.
        (camel_movemail): Loop if we get errno=INTR, and not fail.

get things working with new camel summary stuff.

* providers/nntp/camel-nntp-folder.c:
* providers/nntp/camel-nntp-folder.h:
* providers/nntp/camel-nntp-provider.c:
* providers/nntp/camel-nntp-store.c:
* providers/nntp/camel-nntp-utils.c:
* providers/nntp/camel-nntp-utils.h:
get things working with new camel summary stuff.

* providers/nntp/camel-nntp-summary.c:
* providers/nntp/camel-nntp-summary.h:
removed files since camel-folder-summary does all we need.

o Added some preliminary ESMTP AUTH support

Mon May 15 17:19:31 EDT 2000  Jeffrey Stedfast  <fejj@stampede.org>

  o Added some preliminary ESMTP AUTH support

Remove camel_folder_get_summary, which no longer exists.

* camel-folder.h: Remove camel_folder_get_summary, which no longer
exists.

remove some cruft that we're not currently using.

* Makefile.am: remove some cruft that we're not currently using.

Change to match prototype (size_t vs unsigned int) so it works on 64-bit

* camel-stream-mem.c (camel_stream_mem_new_with_buffer): Change
to match prototype (size_t vs unsigned int) so it works on 64-bit
machines. Noted by msw.

now it deff compiles

>  * providers/mbox/camel-mbox-folder.c (message_changed): Indicate
>  the summary changed also.
>

>  (camel_mbox_summary_update): Also save summary when done.
>  (camel_mbox_summary_expunge): Unindex items when deleting them.
>  (camel_mbox_summary_expunge): Save the index as well as the
>  summary.

>  (camel_folder_summary_touch): New function, indicate the summary
>  info changed.
>  (camel_folder_summary_remove): Dirty here.

camel-smtp-transport.c will now compile :)

providers/smtp/camel-smtp-transport.c: updated smtp_helo to more closely comply with RFC 821 standards

Big bunch o memleaks fixed.

2000-05-11  NotZed  <NotZed@HelixCode.com>

        * camel-mime-part.c (write_to_stream): Unref the filter after
        adding it to the filtering stream.

        * providers/mbox/camel-mbox-summary.c
        (camel_mbox_summary_finalise): Free the folder path.

        * camel-folder-summary.c (camel_folder_summary_finalise): Free the
        summary path.

        * camel-internet-address.c (internet_decode): Free multiple entry
        addresses properly.

        * camel-mime-utils.c (header_decode_mailbox): Plugged another
        memleak, free text after converting it.
        (header_decode_addrspec): More leaks plugged.

        * camel-mime-message.c (finalize): Free message_uid.
        (finalize): Free the recipients hashtable.

Free summary items and charset filters.

2000-05-11    <notzed@helixcode.com>

        * camel-folder-summary.c (camel_folder_summary_finalise): Free
        summary items and charset filters.

2000-05-10    <notzed@helixcode.com>

        * camel-folder-summary.c (camel_folder_summary_finalise): Don't
        free stuff in p, after we've free'd p.

        * providers/mbox/camel-mbox-folder.c (mbox_append_message): Unref
        the stream we created for appending.

fixed camel_smtp_transport_class_init(): now initializes service_class

fix a stupid typo. Thank you, C.

* camel-multipart.c (write_to_stream): fix a stupid typo. Thank
you, C.

* camel-mime-part.c (write_to_stream): don't ref the stream before
wrapper a filter around it, since nothing will ever unref it.

Removed dist-hook section.

2000-05-10  Christopher James Lahey  <clahey@helixcode.com>

* Makefile.am: Removed dist-hook section.

* configure.in: Set the version number.  Added a check for gnome
window icons.  Removed a bunch of unused Makefiles.

* tools/Makefile.am: Created a proper EXTRA_DIST section.

* widgets/e-text/Makefile.am: Added
e-text-event-processor-types.h.

From addressbook/ChangeLog:

2000-05-10  Christopher James Lahey  <clahey@helixcode.com>

* backend/ebook/Makefile.am: Added e-book-types.h, e-card-pairs.h,
e-card-types.h.

* backend/pas/Makefile.am: Added pas-backend-ldap.h.

* contact-editor/Makefile.am: Added a proper EXTRA_DIST section.
Removed some old defines.

* ename/Makefile.am: Added e-name-western-tables.h.

* gui/component/Makefile.am: Added e-ldap-server-dialog.h.  Added
a proper EXTRA_DIST section.

* gui/minicard/e-reflow.c: Added a missed cast.

* printing/Makefile.am: Added a proper EXTRA_DIST section.

From calendar/ChangeLog:

2000-05-10  Christopher James Lahey  <clahey@helixcode.com>

* gui/Makefile.am: Added main.h.  Combined the two EXTRA_DIST
sections.

From camel/ChangeLog:

2000-05-10  Christopher James Lahey  <clahey@helixcode.com>

* Makefile.am: Added camel-types.h, camel-folder-pt-proxy.h, and
camel-thread-proxy.h.

From default_user/ChangeLog:

2000-05-10  Christopher James Lahey  <clahey@helixcode.com>

* Makefile.am, local/Calendar/Makefile.am,
local/Contacts/Makefile.am, local/Directories/Makefile.am,
local/Directories/subfolders/Bigfoot/Makefile.am,
local/Directories/subfolders/Netcenter/Makefile.am,
local/Drafts/Makefile.am, local/Inbox/Makefile.am,
local/Outbox/Makefile.am, local/Trash/Makefile.am: Created a
proper EXTRA_DIST section.

From filter/ChangeLog:

2000-05-10  Christopher James Lahey  <clahey@helixcode.com>

* filter-arg.h: Changed tree.h to gnome-xml/tree.h.

* Makefile.am: Added filter-arg-types.h, filter-arg.h,
filter-xml.h, filter-format.h, filter-druid.h, filter-editor.h.

From libical/ChangeLog:

2000-05-10  Christopher James Lahey  <clahey@helixcode.com>

* configure.in: Reorder Makefiles so that it will build.

From mail/ChangeLog:

2000-05-09  Christopher James Lahey  <clahey@helixcode.com>

* Makefile.am: Removed folder-browser-factory.h since it doesn't
exist.  Added mail-display.h, mail-types.h, pixmaps.h.

From widgets/e-table/ChangeLog:

2000-05-10  Christopher James Lahey  <clahey@helixcode.com>

* Makefile.am: Added e-table-col-dnd.h and table-test.h.

From wombat/ChangeLog:

2000-05-10  Christopher James Lahey  <clahey@helixcode.com>

* Makefile.am: Added an include for the pcs directory.

Bleah. Can't fsync a pipe. As a quick kludge, just don't stream_flush it.

* providers/sendmail/camel-sendmail-transport.c (_send_internal):
Bleah. Can't fsync a pipe. As a quick kludge, just don't
stream_flush it. The right fix will require bringing back
stream_close though.

o Undid changes to camel-internet-address.[c,h] o Fixed

Tue May  9 09:32:33 EDT 2000  Jeffrey Stedfast  <fejj@stampede.org>
  o Undid changes to camel-internet-address.[c,h]
  o Fixed providers/smtp/camel-smtp-transport.c _send to use camel_internet_address_get()
  o formatting changes to providers/smtp/camel-smtp-transport.c

o Moved struct_address from camel-internet-address.c to

Tue May  9 09:03:18 EDT 2000  Jeffrey Stedfast  <fejj@stampede.org>
  o Moved struct_address from camel-internet-address.c to camel-internet-address.h (so it could be used in providers/smtp/camel-smtp-transport.h)
  o providers/smtp/camel-smtp-transport.c _send should now work like it was meant to

CVS

one more refcounting fix I missed before.

* providers/mbox/camel-mbox-folder.c (mbox_get_message_by_uid):
one more refcounting fix I missed before.

o updated to use CamelException (as needed by new camel-stream.c) o no

Mon May  8 22:12:55 EDT 2000  Jeffrey Stedfast  <fejj@stampede.org>
  o updated to use CamelException (as needed by new camel-stream.c)
  o no longer frees memory it shouldn't touch

Use CamelException to signal failure. (camel_stream_write_strings):

* camel-stream.c (camel_stream_read, camel_stream_write,
camel_stream_flush, camel_stream_reset, camel_stream_printf,
camel_stream_write_to_stream): Use CamelException to signal
failure.
(camel_stream_write_strings): Remove. camel_stream_printf is more
useful in most of the places that used this.
(camel_stream_write_string): Change from macro to function to
prevent problems with double-evaluation.

* camel-seekable-stream.c (camel_seekable_stream_seek,
camel_seekable_stream_set_bounds): Use CamelException.
(reset): Update.

* camel-seekable-substream.c, camel-stream-buffer.c,
camel-stream-filter.c, camel-stream-fs.c, camel-stream-mem.c:
Update.

* camel-stream-fs.c: Remove the virtual init functions and move
the code into the creator functions. Add CamelExceptions to
creation functions that could fail.

* camel-data-wrapper.c (camel_data_wrapper_write_to_stream): Use
CamelException.
* camel-mime-message.c, camel-mime-part.c, camel-multipart.c
(write_to_stream): Update.

* camel-mime-parser.c: add an exception to the mime parser private
data and pass that to stream functions as needed.

* gmime-content-field.c, md5-utils.c: Update (badly) for stream
changes.

* camel-exception.h (camel_exception_is_set): convenience macro.

* providers/Makefile.am: disable SMTP for now

* providers/mbox/camel-mbox-folder.c (mbox_append_message): Pass
CamelException to the functions that now need it. Check the
exception after calling camel_stream_flush, and fail if it fails.
(mbox_get_message_by_uid): More updates.

* providers/pop/camel-pop3-folder.c,
providers/pop/camel-pop3-store.c,
providers/sendmail/camel-sendmail/transport.c: Update.

Format From and Reply-To to at least a decoded string. Should probably

2000-05-08  NotZed  <NotZed@HelixCode.com>

        * camel-mime-message.c (process_header): Format From and Reply-To
        to at least a decoded string.  Should probably store them as an
        camelinternetaddress.

blah

Merged NEW_SUMMARY branch back to trunk, and resolved conflicts.

2000-05-08  NotZed  <NotZed@HelixCode.com>

* Merged NEW_SUMMARY branch back to trunk, and resolved conflicts.

* providers/mbox/camel-mbox-summary.c (camel_mbox_summary_update):
Return status.

* camel-stream-filter.c (do_close): We NEED a stream close.

Renamed from camel_mbox_summary_expunge. Takes a gboolean saying whether

* providers/mbox/camel-mbox-summary.c (camel_mbox_summary_sync):
Renamed from camel_mbox_summary_expunge. Takes a gboolean saying
whether to expunge or just sync the mbox file. Change some
g_errors to g_warning so we don't abort. Make the quick
X-Evolution updating code lseek around correctly. Update the
mbox mtime in the summary file even in the quick case.

* providers/mbox/camel-mbox-summary.h: make
CAMEL_MESSAGE_FOLDER_NOXEV not conflict with
CAMEL_MESSAGE_FOLDER_FLAGGED defined in camel-mime-message.h

* providers/mbox/camel-mbox-folder.c (mbox_close): call
camel_mbox_summary_sync to save flag state if not expunging.
(mbox_expunge): Update for camel_mbox_summary_expunge rename.

Merged NEW_SUMMARY branch back to trunk, and resolved conflicts.

2000-05-08  NotZed  <NotZed@HelixCode.com>

* Merged NEW_SUMMARY branch back to trunk, and resolved conflicts.

* providers/mbox/camel-mbox-summary.c (camel_mbox_summary_update):
Return status.

* camel-stream-filter.c (do_close): We NEED a stream close.

Make camel not leak like a sieve.

* camel-object.c: New subclass of GtkObject which is now the base
of the Camel object hierarchy. Currently the only difference
between CamelObject and GtkObject is that CamelObjects don't start
out floating.

* *.h: Move a bunch of typedefs to camel-types.h. Standardize on
using <camel/foo.h> in header files rather than <foo.h>, "foo.h",
or "camel/foo.h". Remove some unneeded includes.

* camel-address.c, camel-data-wrapper.c, camel-folder-search.c,
camel-folder-summary.c, camel-folder.c, camel-mime-filter.c,
camel-mime-parser.c, camel-service.c, camel-session.c,
camel-stream.c: These are now subclasses of CamelObject.

* camel-data-wrapper.c (set_output_stream):
* camel-medium.c (set_content_object):
* camel-seekable-substream.c
(init_with_seekable_stream_and_bounds):
* providers/mbox/camel-mbox-folder.c (mbox_get_message_by_uid):
remove gtk_object_sink calls.

* camel-stream-buffer.c (init_vbuf):
* camel-stream-filter.c (camel_stream_filter_new_with_stream):
ref the original stream.

* camel-folder-summary.c (camel_folder_summary_finalise): unref
the filters when finalizing.

* camel-mime-part-utils.c
(simple_data_wrapper_construct_from_parser,
camel_mime_part_construct_content_from_parser):
* camel-mime-part.c (camel_mime_part_set_content): Unref objects
that are created only to be handed off to other objects. If
they're going to be needed later, they will have been additionally
ref'ed by the object that needs them.

* providers/pop3/camel-pop3-folder.c (get_message_by_number):
unref the message stream after creating the data from it.

* camel-stream.c, camel-stream-buffer.c, camel-stream-filter.c,
camel-stream-fs.c, camel-stream-mem.c: Remove camel_stream_close,
since its semantics are dubious (what happens when you close a
stream other people still have references on?).

* providers/nntp/camel-nntp-store.c:
* providers/smtp/camel-smtp-transport.c:
* providers/pop3/camel-pop3-store.c:
replace camel_stream_close calls with gtk_object_unref.

* providers/mbox/camel-mbox-folder.c:
* providers/nntp/camel-nntp-folder.c:
* providers/sendmail/camel-sendmail-transport.c:
replace camel_stream_close with camel_stream_flush +
gtk_object_unref

A machine which serves neither POP nor KPOP is not a POP server.

* providers/pop3/camel-pop3-store.c (query_auth_types): A machine
which serves neither POP nor KPOP is not a POP server.

* providers/smtp/camel-smtp-provider.c: Note in the description
that this provider is not yet tested.

And same here ... (camel_folder_summary_encode_fixed_int32): Ugh, fwrite

2000-05-05  NotZed  <NotZed@HelixCode.com>

        * camel-folder-summary.c: And same here ...
        (camel_folder_summary_encode_fixed_int32): Ugh, fwrite doesn't
        return -1 on error ..
        (camel_folder_summary_decode_fixed_int32): Neither deos fread.
        (camel_folder_summary_encode_token): Fix here too.

And same here ...

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.

Plug a memory leak. (header_decode_text): Fixed memory leaks with

2000-05-05    <notzed@helixcode.com>

        * camel-mime-utils.c (header_decode_mailbox): Plug a memory leak.
        (header_decode_text): Fixed memory leaks with g_string_append().
        (header_encode_string): And here too, and a few other places.  The
        glib api is so awful ...
        (header_content_type_decode): More memory leaks, more left ...

2000-05-05    <notzed@helixcode.com>

        * camel-mime-parser.c (folder_scan_init_with_fd): Make sure we
        init the end of buffer sentinal!
        (folder_scan_init_with_stream): And here too ...

Maxcount is minimum of the max and the requested count, not the maximum :)

2000-05-04  NotZed  <NotZed@HelixCode.com>

        * providers/mbox/camel-mbox-folder.c (summary_get_message_info):
        Maxcount is minimum of the max and the requested count, not the
        maximum :)

        * camel-mime-parser.c (folder_scan_content): Properly set midline,
        so we dont falsely catch offset boundary markers (i.e. From inside
        content).
        (folder_read): Set a sentinal on the end of the read data (\n) so
        we dont have to check the buffer boundary in the inner loop.
        (mempool_*): New experimental memory management routines, speed
        up simple structure parsing by about 25% ... not compiled in by
        default.  Something similar may be needed for camel-mime-utils to
        address performance issues with g_malloc and friends.

        * camel-mime-utils.c: Added a macro w(x) used to wrap all warnings
        about mime/rfc violations, so they can be turned off.

        * camel-folder-summary.c (summary_build_content_info): Step after
        the end of a message ...
        Turn into a stand-alone program for testing and profiling.

Make this compile again in the !KRB4 case

Don't fall back to plaintext passwords if APOP fails, since it should also

* providers/pop3/camel-pop3-store.c (pop3_connect): Don't fall
back to plaintext passwords if APOP fails, since it should also
fail.

New function to replace camel_provider_scan. Returns a list of either (a)

        * camel-session.c (camel_session_list_providers): New function to
        replace camel_provider_scan. Returns a list of either (a) all
        currently-loaded providers, or (b) all available providers.

        * camel-url.[ch]: Add an "empty" flag to CamelURL (indicating that
        it contains only a protocol).

        * camel-service.c (camel_service_query_auth_types): Make this take
        a CamelException (since it may have to try to connect to the
        server, and it might not able to.)

        * providers/pop3/camel-pop3-store.c: add KPOP (Kerberized POP)
        support. This is mostly so I have two kinds of authmech to play
        with instead of just one. (But it does actually work.)

        * providers/smtp/camel-smtp-transport.c (query_auth_types): update
        for prototype change, but disable the functionality, since it
        doesn't really support any auth types yet.
        (camel_smtp_transport_get_type): add an object init function to
        set the service url_flags.

No, we're not going to have g_strcasecmp for no good reason,
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.

Some small build fixes:

* camel/Makefile.am (INCLUDES): add $(UNICODE_CFLAGS) to the INCLUDES,
people who installed libunicde in non-standard include paths
need this.

        * mail/Makefile.am (evolution_mail_LDADD):
s/-lunicode/$(UNICODE_LIBS)/

s/strcasecmp/g_strcasecamp/ everywhere except intl/, per michael's
request.

If the iconv handle is -1, then dont try and convert (crashes

2000-05-03  NotZed  <NotZed@HelixCode.com>

        * camel-mime-utils.c (rfc2047_decode_word): If the iconv handle is
        -1, then dont try and convert (crashes unicode_iconv?).
        (rfc2047_decode_word): Use alloca for variables instead of
        g_malloc - by the rfc they should always be short.
        (rfc2047_decode_word): If we can't do the charset conversion, undo
        the quoted-printable/base64 at least?  Should probably convert
        unknown characters to the utf-8 unknown character.

fix typo when dereferencing saveoffset.

2000-05-02  Larry Ewing  <lewing@helixcode.com>

* camel-mime-utils.c (header_decode_date): fix typo when
dereferencing saveoffset.

Missed these ...

REmove gmime-utils.h from here.

2000-05-02  NotZed  <NotZed@HelixCode.com>

        * camel.h: REmove gmime-utils.h from here.

        * providers/mbox/camel-mbox-search.[ch]: Removed.  Functionally
        redundant.

        * providers/mbox/camel-mbox-folder.c (mbox_search_by_expression):
        Use the new CamelFolderSearch class to do the actual searching,
        just setup the search here.

        * camel-folder-search.[ch]: A helper class that providers may
        subclass to provide their own search functionality, or they can
        simply use as is, it supports body searches if an ibex is
        supplied, and header searches if a summary is supplied.

Set G_LOG_DOMAIN in following files.

 addressbook/demo/Makefile.am
 addressbook/printing/Makefile.am
 camel/Makefile.am
 camel/providers/MH/Makefile.am
 camel/providers/maildir/Makefile.am
 camel/providers/mbox/Makefile.am
 camel/providers/nntp/Makefile.am
 camel/providers/pop3/Makefile.am
 camel/providers/sendmail/Makefile.am
 camel/providers/smtp/Makefile.am
 composer/Makefile.am
 e-util/Makefile.am
 filter/Makefile.am
 libibex/Makefile.am
 mail/Makefile.am
 shell/Makefile.am
 tests/Makefile.am
 tests/ui-tests/Makefile.am
 widgets/e-table/Makefile.am
 widgets/e-text/Makefile.am
 widgets/meeting-time-sel/Makefile.am
 widgets/shortcut-bar/Makefile.am

>  * providers/mbox/camel-mbox-search.c
>  (camel_mbox_folder_search_by_expression): Dont store/remove
>  current search from the search list.
>
>  * providers/mbox/camel-mbox-folder.h: Removed searches list,
>  searches are all sync now.

>  * gmime-utils.[ch]: What the hell, remove it.  This will break the
>  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.

Removed bizarre string_dichotomy version of this. This code is somewhat

* 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.

Fix the previous overflow problem properly (can happen in 2 places).

2000-05-02  NotZed  <NotZed@HelixCode.com>

        * camel-mime-parser.c (folder_scan_header): Fix the previous
        overflow problem properly (can happen in 2 places).
        (header_append): A new macro to include the code changed above, so
        it only appears in one place.
        (folder_scan_step): Change the content type to text/plain if the
        multipart is broken.  Doesn't actually change the header though.
        (header_append): Also move the header-start tracking stuff here.
        Could be a static function to save code.

Dont use autofill on these fucking long function anmes!!!!!!

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!!!!!!

Fix the offset for the summary when an item is expunged to take account of

2000-05-02  NotZed  <NotZed@HelixCode.com>

        * providers/mbox/camel-mbox-summary.c
        (camel_mbox_summary_expunge): Fix the offset for the summary when
        an item is expunged to take account of the From line.

Deleted messages properly get removed from the index.
A new folder signal for gui/etc's to track changes.

(camel_mbox_summary_expunge): Oops, my wrong, use the string uid
to unindex on.
(mbox_expunge): Emit a folder_changed signal on expunge (uh, even
if it didn't ...)
(camel_folder_class_init): Added a folder_changed signal.

>  (camel_folder_expunge): Changed to only allow expunge on an open
>  folder.  It doesn't make sense for mbox, otherwise (?)

>  (func_header_contains): Debug out some search stuff.
>  (index_folder): Make sure we index using a decimal uid, since
>  thats what everything else indexes off (oops).
>  Upped SUMMARY_VERSION as a result.
>
>  * camel-folder.h: Removed CamelSearchFunc.

Removed search_id.

* providers/mbox/camel-mbox-folder.h (CamelMboxFolder): Removed
search_id.

* providers/mbox/camel-mbox-search.c
(camel_mbox_folder_search_cancel): Remove.d
(camel_mbox_folder_search_complete): Removed.
(camel_mbox_folder_search_by_expression): Changed back to sync
api.
(struct _searchcontext): Removed cancelled flag.
(find_context): Removed.

* providers/mbox/camel-mbox-search.h
(camel_mbox_folder_search_by_expression): Moved back to sync api.

(camel_folder_search_complete): Removed.
(camel_folder_search_cancel): Removed.
(CamelFolderClass): New function summary_get_by_uid() to get a single
summary.
(*search*): Moved back to synchronous search api ... *sigh*

More message cleanup.  Flags now work, are saved, and the delete
flag causes an expunge to expunge the deleted messages.

2000-05-01  NotZed  <NotZed@HelixCode.com>

        * providers/mbox/camel-mbox-summary.c
        (camel_mbox_summary_set_flags_by_uid): New function to update the
        flags in the summary.
        (camel_mbox_summary_expunge): Expunge messages from a folder.
        (offset_content): Re-align offsets of summary when messages
        added/removed to an existing summary.
        (camel_mbox_summary_remove_uid): Remove a message summary entry by
        uid.
        (index_folder): Restore flags from X-Evolution header, if they are set.

        * providers/mbox/camel-mbox-folder.c (_get_message_by_uid):
        Connect to the message_changed signal.
        (_init): Set permanent flags to something reasonable.  No user
        flags yet ...
        (message_changed): If the flags of the message change, update the
        flags in the summary.
        (mbox_expunge): Implement the expunge.
        (camel_mbox_folder_class_init): Renamed all leading _'s to mbox_'s

        * camel-folder.c (_finalize): Uh, dont free permanent_flags
        anymore (this wouldn't failed anyway, it was a GList !!!)

        * camel-folder.h (struct _CamelFolder): Change permanent_flags to
        a bitfield.
        (list_permanent_flags): Renamed to get_permanent_flags, and
        returns a bitfield.
        (camel_folder_expunge): Changed expunge to a void type.  The
        messages would no longer be useful after they have been removed
        ...

        * camel-mime-message.c (set_flag): Removed.
        (camel_mime_message_set_flag): Removed.
        (get_flag): Removed.
        (camel_mime_message_get_flag): Removed.
        (add_flag_to_list): Removed.
        (get_flag_list): Removed.
        (camel_mime_message_get_flag_list): Removed.
        (camel_mime_message_get_flags): New interface to get system flags.
        (camel_mime_message_set_flags):  " to set ".
        (camel_mime_message_get_user_flag): To get a user flag.
        (camel_mime_message_set_user_flag): To set a user flag.
        (finalize): Hmm, the old one free'd the key and data, not good
        when the data is a boolean ...

oops. s/CAMEL_TRANSPORT_TYPE/CAMEL_PROVIDER_TRANSPORT/

Tweak the definition of CamelProvider. Among other things, a provider may

* camel-provider.h: Tweak the definition of CamelProvider. Among
other things, a provider may now be both a store and a transport.

* camel-provider.c: Remove a lot of code we had no intention of
using. This now only contains two functions: camel_provider_init
to read the installed .urls files, and camel_provider_load to
load and register a new provider.

* camel-session.c: Remove more unused code and simplify some of
the remaining code. The list of available provider modules is now
stored in the session, and it handles calling camel_provider_load
to load them as needed. Provider registration is now done by
calling back from the module init routine, which allows a single
module to register providers for multiple URL types.

* providers/*: Update provider structures and init routines for
the new stuff. Add a .urls file to each provider specifying what
urls it handles, and install that with the library.

* providers/nntp/camel-nntp-provider.c: Add hints towards
supporting both news: and nntp: URLs, and using nntp as both a
store and a transport.

const poison

* camel-internet-address.c (camel_internet_address_get): const
poison

camel_mime_parser_tell() returns an offset from where it started parsing,

* 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.

remove an unused #include

fix a bug that would cause corruption with very long headers.

* camel-mime-parser.c (folder_scan_header): fix a bug that would
cause corruption with very long headers.

Made Evolution OAF-compatible.

Dont try and check a NULL header.

2000-04-27  NotZed  <NotZed@HelixCode.com>

        * camel-mime-utils.c (check_header): Dont try and check a NULL
        header.

        * camel-recipient.[ch]: Dead.  Its not pining.

        * camel-mime-message.h: Dont include recipients.h anymore.

        * camel-mime-message.c (camel_mime_message_add_recipient): Accept
        name/address separately, and store in an CamelInternetAddress.
        (add_recipient): Removed.
        (remove_recipient): Removed.
        (remove_recipient_address): Renamed from remove_receipient, works
        via address.
        (camel_mime_message_remove_recipient_name): New function to remove
        by name.
        (get_recipients): Removed.
        (camel_mime_message_get_recipients): Return a camel-internet-address.
        (write_to_stream): No longer write receipients directly.
        (write_recipients_to_stream): Removed.
        (write_one_recipient_to_stream): Removed.
        (camel_mime_message_init): Setup recipients hashtable, rather than
        usign the recipients stuff.
        (set_recipient_list_from_string): Killed, a violent and lengthy
        death.
        (process_header): Simplified recipient handling code a lot.
        (received_date_str, sent_date_str, reply_to_str, subject_str,
        from_str): Removed some oddly-defined global statics.
        (camel_mime_message_class_init): Dont initialise above variables
        anymore.
        (init_header_name_table): Removed, use a table to init this, and
        do it in class init (2 lines of code ...).

        * camel-news-address.c: Class to represent news addresses -
        currently empty, and not built.

        * camel-internet-address.h: Class to represent internet (email)
        addresses.

        * camel-address.h: Abstract class to represent (lists of)
        addresses.

Class to represent news addresses - currently empty, and not built.

2000-04-27  NotZed  <NotZed@HelixCode.com>

* camel-news-address.c: Class to represent news addresses -
currently empty, and not built.

* camel-internet-address.h: Class to represent internet (email)
addresses.

* camel-address.h: Abstract class to represent (lists of)
addresses.

Revert previous change. I was confused.

* camel-mime-part.c (write_to_stream): Revert previous change. I
was confused.

expose these routines.

* camel-url.[ch] (camel_url_encode, camel_url_decode): expose
these routines.

Diagram of the camel-stream and derived classes used within
camel.

Only write a newline between the headers and the content object if the

* camel-mime-part.c (write_to_stream): Only write a newline
between the headers and the content object if the content object
is not a CamelMedium. (If the content is a medium, it may have its
own headers, which then need to go before the blank line.)

Remove. We weren't using the fields that made this different from

* 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

sink the content object after referencing it.

* camel-medium.c (set_content_object): sink the content object
after referencing it.

* camel-mime-part.c: fix various little things in the handling
of CamelMedium methods. Change camel_mime_part_set_text to the
more generic camel_mime_part_set_content.

sync to current reality

        * camel.h: sync to current reality

        * camel-folder-utils.[ch]: removed

fix format specifier for time zone. Fix typo in month names array.

* camel-mime-utils.c (header_format_date): fix format specifier
for time zone. Fix typo in month names array.

Big cleanup of camel-stream-*, got rid of 3 classes, improved the interfaces,
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.

only return TRUE if the parent is at eos AND the buffer has been exhausted

* camel-stream-buffer.c (_eos): only return TRUE if the parent is
at eos AND the buffer has been exhausted

fix some incorrect macro usage that resulted in bogus casts

* camel-mime-message.c: fix some incorrect macro usage that
resulted in bogus casts