Build fix

more work.

* camel/camel-folder-pt-proxy.c:
more work.

Check pthreads.

1999-09-21  bertrand  <Bertrand.Guiheneuf@aful.org>

* configure.in:
Check pthreads.

* camel/Makefile.am:
camel-folder-pt-proxy.c is only compiled
when pthreads are available.

camel_folder_get_full_name did not work. Fixed.

Preliminary subfolder support for maildirs.

Signal proxying implemenatation. (_signal_marshaller_server_side):

1999-09-21  bertrand  <Bertrand.Guiheneuf@aful.org>

* camel/camel-folder-pt-proxy.c:
Signal proxying implemenatation.
(_signal_marshaller_server_side):
(_signal_marshaller_client_side):
(_init_signals_proxy):
Code not is tested and has to be best
explained as it uses threads conditions and
gtk signal system.
(_thread_notification_catch): notify pending signals
as well as thread availability.

completed

1999-09-20  bertrand  <Bertrand.Guiheneuf@aful.org>

* camel/camel-folder-pt-proxy.c (_op_exec_or_plan_for_exec):
completed

Binch of new funcs:
(_maybe_run_next_op):
Called by the watch notification when
a threaded op is completed
(_thread_notification_catch):
notification watch call back
(_init_notify_system):
set up the notification channel
(notify_availability):
called by threads before completion.

synczZzZ

new func. Try to exec an operation in a thread or queue it if a thread is

1999-09-18  bertrand  <Bertrand.Guiheneuf@aful.org>

* camel/camel-folder-pt-proxy.c (_op_exec_or_plan_for_exec):
new func. Try to exec an operation in a thread
or queue it if a thread is already busy.

* camel/camel-op-queue.c (camel_op_queue_set_service_availability):
(camel_op_queue_get_service_availability):
new funcs.

* camel/camel-op-queue.c (camel_op_new):
(camel_op_free):
new funcs. Uses glib mem chunks.

added notify io_channel. new func. run the next operation.

1999-09-17  bertrand  <Bertrand.Guiheneuf@aful.org>

* camel/camel-folder-pt-proxy.c (_init_with_store):
added notify io_channel.
* (camel_op_queue_run_next_op): new func.
run the next operation.

New object. Operation queue. Meant to be used in non-blocking proxy

1999-09-17  bertrand  <Bertrand.Guiheneuf@aful.org>

* camel/camel-op-queue.h:
* camel/camel-op-queue.c:
New object. Operation queue. Meant to be used in
non-blocking proxy objects.

First proxy (pthread based) implementation starts.

* First proxy (pthread based) implementation starts.

1999-09-14  bertrand  <Bertrand.Guiheneuf@aful.org>

* camel/Makefile.am (libcamel_la_SOURCES):
added camel-folder-pt-proxy.[ch] to the build
process.

* camel/camel-folder-pt-proxy.c (_init_with_store):
started implementation of the pthread-based
folder proxy.

forgotten files

Summary files are here and work !!!

1999-09-08  bertrand  <Bertrand.Guiheneuf@aful.org>

* camel/providers/MH/mh-summary.c
(mh_load_summary):
(mh_save_summary):
(mh_create_summary):
implemented summary (files) for MH folders.

* camel/providers/MH/camel-mh-folder.c
(_get_message_by_uid): implemented.
(camel_mh_folder_class_init):
(_get_message_uid): implemented

impemented.

1999-09-08  bertrand  <Bertrand.Guiheneuf@aful.org>

* camel/providers/MH/camel-mh-folder.c
(_get_message_by_uid): impemented.

* camel/camel-folder.c:
(camel_folder_get_message_uid_by_number):
const'ified uid.
(camel_folder_get_message_by_uid): idem
removed stupid camel_folder_get_message_uid_by_number
method.

Documented UIDs.

1999-09-08  bertrand  <Bertrand.Guiheneuf@aful.org>

* camel/camel-folder.c (camel_folder_has_uid_capability):
(camel_folder_get_message_uid):
(camel_folder_get_message_uid_by_number):
(camel_folder_get_message_by_uid):
(camel_folder_get_uid_list):
Documented UID methods.

More work on UID list in Mh provider.

1999-09-08  bertrand  <Bertrand.Guiheneuf@aful.org>

* tests/ui-tests/store_listing.c (close_all):
close all open folders and stores. (necessary
for UID list saving ).

* camel/providers/MH/mh-uid.c (mh_generate_uid_list):
fix: store UID list in CamelMHFolder object.

* camel/providers/MH/camel-mh-folder.c (_open):
read or create UID list.
(_close): save UID list.

oops

correct parameter decl (const)

1999-09-07  bertrand  <Bertrand.Guiheneuf@aful.org>

* camel/md5-utils.c (md5_get_digest_from_file):
correct parameter decl (const)

* camel/md5-utils.h: typo.

* camel/providers/MH/mh-uid.c
More work on UID stuff for MH.
(mh_save_uid_list):
(mh_load_uid_list):
(mh_generate_uid_list):
new funcs. Manage on-disk uid list.

* camel/providers/MH/mh-utils.c (mh_is_a_message_file):
Util routines live here now.

camel/md5-utils.c Documented all funcs.

1999-09-07  bertrand  <Bertrand.Guiheneuf@aful.org>

* camel/md5-utils.c
Documented all funcs.

(md5_get_digest_from_stream):
correct typo.
(md5_get_digest_from_file):
same typo corrected.

raw routines are declared public now. Md5 use has to be versatile.

1999-09-07  bertrand  <Bertrand.Guiheneuf@aful.org>

* camel/md5-utils.h :
raw routines are declared public now.
Md5 use has to be versatile.

* camel/providers/MH/mh-uid.c (mh_uid_get_for_file):
new func. Returns an UID for an MH message.

new function : get file md5 signature. To be used in providers code.

1999-09-06  bertrand  <Bertrand.Guiheneuf@aful.org>

* camel/md5-utils.h:
* camel/md5-utils.c:
(md5_get_digest_from_file):
new function : get file md5 signature.
To be used in providers code.

new methods.

1999-09-06  bertrand  <Bertrand.Guiheneuf@aful.org>

* camel/md5-utils.h:
* camel/md5-utils.c:
(md5_get_digest_from_stream):
new methods.

changed names to follow camel style.

1999-09-06  bertrand  <Bertrand.Guiheneuf@aful.org>

* camel/md5-utils.h:
* camel/md5-utils.c:
changed names to follow camel style.

`

imported md5 coding routine from rpm. Compiles.

1999-09-06  bertrand  <Bertrand.Guiheneuf@aful.org>

* camel/md5-utils.c: imported md5 coding
routine from rpm. Compiles.

Basic UID framework.

1999-09-05  bertrand  <Bertrand.Guiheneuf@aful.org>

* camel/camel-folder.c (camel_folder_has_uid_capability):
(camel_folder_get_message_uid):
(camel_folder_get_message_uid_by_number):
(camel_folder_get_message_by_uid):
(camel_folder_get_uid_list):
Basic UID framework.

New revision. Some fixes.

1999-09-05  bertrand  <Bertrand.Guiheneuf@aful.org>

* devel-docs/misc/ref_and_id_proposition.txt:
New revision. Some fixes.

* camel/camel-folder.h (struct _CamelFolder): added
uid_capability field.

* started to work on UIDs

use folder summary instead of opening all messages.

1999-09-05  bertrand  <Bertrand.Guiheneuf@aful.org>

* tests/ui-tests/store_listing.c (show_folder_messages):
use folder summary instead of opening all messages.

* camel/providers/MH/camel-mh-folder.c (_create_summary):
basic and highly non-efficient summary implementation.
Should be seen as a proof of concept only.
subfolder summary still has to be implemented.

* camel/providers/maildir/camel-maildir-folder.c (_init_with_store):
hasn't summary for the moment.
* camel/providers/maildir/camel-maildir-folder.c
cosmetic changes.

started summary implementation. (_open): correct use of open.

1999-09-04  bertrand  <Bertrand.Guiheneuf@aful.org>

* camel/providers/MH/camel-mh-folder.c (_create_summary):
started summary implementation.
(_open): correct use of open.

* camel/camel-folder.c (camel_folder_get_summary):
get folder associated summary object.

added summary files build

1999-09-04  bertrand  <Bertrand.Guiheneuf@aful.org>

* camel/Makefile.am:
added summary files build

* camel/camel-folder-summary.[ch]:
basic summary framework

sgml doc has camel-recipient now.

1999-09-03  bertrand  <Bertrand.Guiheneuf@aful.org>

* devel-docs/camel/Makefile.am:
sgml doc has camel-recipient now.

* camel/camel-recipient.c (camel_recipient_foreach_recipient_type):
added in-line documentation.

sync again. Changelog comes tomorow

sync

Added (experimental) maildir provider written by Jukka Zitting

1999-09-02  bertrand  <Bertrand.Guiheneuf@aful.org>

* camel/providers/maildir: Added (experimental) maildir
provider written by Jukka Zitting <hukka@greywolves.org>

* camel/providers/Makefile.am (SUBDIRS): build maildir
provider.

* camel/camel-provider.c (camel_provider_get_for_protocol):
bug fix. patch from  Jukka Zitting <hukka@greywolves.org>

changed decl to fit CRLFunc. (_write_recipients_to_stream): calls

1999-09-02  bertrand  <Bertrand.Guiheneuf@aful.org>

* camel/camel-mime-message.c (_write_one_recipient_to_stream):
changed decl to fit CRLFunc.
(_write_recipients_to_stream):
calls camel_recipient_foreach_recipient_type now.

* camel/camel-recipient.c (camel_recipient_foreach_recipient_type):
new convinience function. Iterate over all recipient types.

* camel/gmime-utils.c (gmime_write_header_table_to_stream):
s/write_header_table_to_stream/gmime_write_header_table_to_stream/
(gmime_write_header_with_glist_to_stream):
s/write_header_with_glist_to_stream/gmime_write_header_with_glist_to_stream/

now use CamelRecipientTable

1999-09-01  bertrand  <Bertrand.Guiheneuf@aful.org>

* camel/camel-mime-message.c (_finalize):
(_add_recipient):
(_remove_recipient):
(_get_recipients): now use CamelRecipientTable

* camel/gmime-content-field.c:
(gmime_content_field_unref): test if object
to free is non void.

Still are some bugs in camel-recipient.c

When using g_free (obj) don't test if obj != NULL g_free () already do

1999-09-01  bertrand  <Bertrand.Guiheneuf@aful.org>

* camel/camel-folder.c (_finalize):
(_set_name):
* camel/camel-mime-message.c (_finalize):
* camel/camel-mime-part.c (_finalize):
(_set_description):
(_set_disposition):
* camel/camel-service.c (_finalize):
* camel/camel-stream-fs.c (_finalize):
* camel/gmime-content-field.c:
(gmime_content_field_construct_from_string):
* camel/url-util.c (g_url_free):

When using g_free (obj) don't test if obj != NULL
g_free () already do that. Thanks to elerium for
the feedback.

19

new func. More work on new independant recipient code.

1999-08-30  bertrand  <Bertrand.Guiheneuf@aful.org>

* camel/camel-recipient.c (camel_recipient_get):
(camel_recipient_remove):
(camel_recipient_add):
new func. More work on new independant recipient code.

1999-08-29  bertrand  <Bertrand.Guiheneuf@aful.org>

* MAINTAINERS: updated my e-mail address.

Address update

text parameter declared const

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.

contruct the content from the buffer before calling CamelMedium

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.

new file, groups mime-part related utils. Meant to be used by providers

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.

now descend from CamelMedium.

1999-08-25  bertrand  <Bertrand.Guiheneuf@aful.org>

* camel/camel-mime-part.c: now descend from CamelMedium.

* tests/test1.c (main): all headers must be strdup'ed
(main): unref created objects

La tout n'est qu'ordre et beaute,
Luxe, calme et voluptee.

these methods are in CamelMedium now.

1999-08-25  bertrand  <Bertrand.Guiheneuf@aful.org>

* camel/camel-medium.c (_set_content_object):
(_get_content_object): these methods are
in CamelMedium now.

Going to revamp MimePart soon so that it inherits from CamelMedium.

new class. Will handle all sort of information media (Mime mail messages,

1999-08-24  bertrand  <Bertrand.Guiheneuf@aful.org>

* camel/camel-medium.c (camel_medium_class_init):
new class. Will handle all sort of information media
(Mime mail messages, Lotus Notes  mail messages,
postit notes, faxes, who knows .... :)
CamelMimePart will inherit from it.

* camel/camel-mime-part.c (_set_disposition):
(_set_description):
description and disposition parameters are now const.

* camel/gmime-content-field.c (gmime_content_field_free): added
assertion code.

* camel/providers/MH/camel-mh-folder.c (_get_message):
uses buffered stream.

* camel/camel-stream-buffered-fs.c:
new stream to accelerate file ops.

Thanks to jwz, I've decided to add a level of abstraction to Camel.
In the future, it should be able to handle other mail systems, but
also non-mail information vehicles.

Enough for today. Roller time!

name parameter is const. This fixes a bug in destroy ()

1999-08-20  bertrand  <Bertrand.Guiheneuf@aful.org>

* camel/camel-stream-fs.c (camel_stream_fs_new_with_name):
name parameter is const. This fixes a bug in destroy ()

More work on RFC 2047 encoder.

return the number of bytes read. How can this have ever worked ?

1999-08-17  bertrand  <Bertrand.Guiheneuf@aful.org>

* camel/camel-stream.c (camel_stream_read):
return the number of bytes read.
How can this have ever worked ?
(camel_stream_flush): don't return anything.

* camel/gmime-utils.c (get_header_table_from_stream):
mem leak fixed.

+ various other mem leaks.

Now has a popup menu on mailbox tree to allow easier tests. Implemented

1999-08-16  bertrand  <Bertrand.Guiheneuf@aful.org>

* tests/ui-tests/store_listing.c:
Now has a popup menu on mailbox tree to allow
easier tests. Implemented the copy stuff.
Works well for the MH provider :)))

* camel/providers/MH/camel-mh-folder.c (_copy_message_to):
Test MH provider fast copy implemented.

new method.

1999-08-15  bertrand  <Bertrand.Guiheneuf@aful.org>

* camel/camel-folder.c (_copy_message_to): new method.

moved the active list readjustment code here. Much saner, providers won't

1999-08-15  bertrand  <Bertrand.Guiheneuf@aful.org>

* camel/camel-folder.c (camel_folder_expunge):
moved the active list readjustment code here.
Much saner, providers won't have to worry about
that.

* camel/providers/MH/camel-mh-folder.c (_expunge):
no more active list readjustment stuff.

* camel/camel-folder.h: the expunge virtual no more
returns a list of expunged messages. Now providers
only have to set the expunge flag on the expunged
messages.

* camel/camel-folder.c (camel_folder_get_message):
moved the caching code here. Finally, I don't want
providers to crash the libs with a bad message cache
implementation.
(_close): do not call the CamelFolder virtual expunge
method directly, use camel_folder_expunge() instead.
(camel_folder_expunge): added the want_list param.
The client can decide if it wants the expunged message
list or not. If yes, it'll have to unref the messages
itself.

Implemented deletion/expunge mechanism.

1999-08-14  bertrand  <Bertrand.Guiheneuf@aful.org>

* tests/ui-tests/store_listing.c (delete_selected_messages):
(expunge_selected_folders):
Implemented deletion/expunge mechanism.

* camel/camel-folder.c (_get_message):
Added some debug info.

* camel/providers/MH/camel-mh-folder.c (_expunge):
implemented and tested.

* camel/camel-mime-message.c (_set_flag):
changed the old braindead implementation.
boolean are inserted in the flag hash
table casted as gpointers.

* camel/camel-mime-message.c: indentation fix

default implementation to be called first by providers methods. It looks

1999-08-13  bertrand  <Bertrand.Guiheneuf@aful.org>

* camel/camel-folder.c (_get_message): default implementation
to be called first by providers methods. It looks in the
folder message list (in memory) to see if the message has
not already been retrieved, and in this case, returns
the same CamelMimeMessage object

Caching messages in folder kworks. Providers will have to
do their job correctly.
Enough for today.

New field (message_list) which will hold a reference on each message

1999-08-13  bertrand  <Bertrand.Guiheneuf@aful.org>

* camel/camel-folder.h (struct _CamelFolder):
New field (message_list) which will hold a
reference on each message obtained by the folder,
which is necessary in order for the caching procedure
to work (Getting the same message from a folder twice
will return the same CamelMimeMessage object).

* camel/camel-folder.c (camel_folder_get_message):
When the store retreives a message put it in its
message list.
(_finalize): free message list.

* ChangeLog: fix typo (parmanent)

new method, returns the list of permanent flags supported by the folder.

1999-08-13  bertrand  <Bertrand.Guiheneuf@aful.org>

* camel/camel-folder.c (_get_parmanent_flag_list):
(camel_folder_get_parmanent_flag_list):
new method, returns the list of permanent
flags supported by the folder.

* camel/camel-mime-message.c (_get_flag_list):
(camel_mime_message_get_flag_list): new method,
return the list of flag name used by this message.

those two func go here now.

1999-08-13  bertrand  <Bertrand.Guiheneuf@aful.org>

* camel/hash-table-utils.c (g_strcase_equal):
(g_strcase_hash): those two func go here now.

* camel/hash_table_utils.c (hash_table_generic_free):
free a (gpointer, gpointer) hash table pair.

* camel/camel-mime-message.c (camel_mime_message_init): use
case insensitive hash table functions.
(_set_flag):
(camel_mime_message_set_flag):
(_get_flag):
(camel_mime_message_get_flag):
Use const for flag name, they are now
duplicated.

remove stupid debug code. (add_mail_store): use camel_session_get_store

1999-08-12  bertrand  <Bertrand.Guiheneuf@aful.org>

* tests/ui-tests/store_listing.c (show_folder_messages):
remove stupid debug code.
(add_mail_store): use camel_session_get_store instead
of creating MH store directly.
(main): load MH provider.

* camel/camel-provider.c (camel_provider_register_as_module):
register new provider.
(camel_provider_get_for_protocol):
Now, implementation is correct.

* camel/camel-store.c (_finalize):
* camel/camel-store.h (struct _CamelStore):
further disabled url_name field use.
URL will be generated dynamically.

Ben dam don dieu, vla t'y pas que ya tout le bourier qui marche !
(Autoload store/protocol from URL works)

compilation and runtime fixes.

1999-08-12  bertrand  <Bertrand.Guiheneuf@aful.org>

* camel/camel-session.c (camel_session_get_store_for_protocol):
compilation and runtime fixes.

* camel/providers/MH/camel-mh-store.c (_init):
synced with CamelStore.

* camel/camel-store.c (_init):
in CamelStore::init, url_name is now const.
disabled url_name copy.

Last check in did not even compile...

new function: returns a store for an URL.

1999-08-12  bertrand  <Bertrand.Guiheneuf@aful.org>

* camel/camel-session.c (camel_session_get_store):
new function: returns a store for an URL.
(camel_session_get_store_for_protocol):
new functionc: returns a store for a given
store protocol (as IMAP/POP/MH ...)
* camel/string-utils.c (g_strcase_equal):
(g_strcase_hash): case insensitive hash table
funcs.

* camel/camel-session.c (camel_session_init): hash table
keys are case insensitive.

* camel/camel-provider.c (camel_provider_get_for_protocol):
new function, returns the last registered
provider for a protocol.

new file. MH provider registration stuff.

1999-08-12  bertrand  <Bertrand.Guiheneuf@aful.org>

* camel/providers/MH/camel-mh-provider.c:
new file. MH provider registration stuff.

* camel/camel-provider.c (camel_provider_register_as_module):
load a provider from a shared object (plugin).
(camel_provider_register): register a provider
"by hand". Used for statically defined providers.

* tests/test7.c: new test.
tests providers loading framework.

Providers modules loading Works !!! :))))

Message creation still consumes a lot of
memory (maybe 2000k) and I really don't
understand why. As soon as I get gtk working
on SunOS, purify should hopefully give me the answer.

1999-08-11  bertrand  <Bertrand.Guiheneuf@aful.org>

* camel/camel-service.c (_finalize):
* camel/camel-stream-fs.c (_finalize):
(_destroy): close file descriptor.
* camel/camel-stream-mem.c (_finalize):
* camel/camel-store.c (_finalize):
* camel/camel-folder.c (_finalize):

implemented destructors.

1999-08-11  bertrand  <Bertrand.Guiheneuf@aful.org>

* camel/camel-multipart.c (_finalize):
* camel/camel-simple-data-wrapper.c (_finalize):
* camel/camel-mime-part.c (_finalize):
implemented destructors.

* camel/gmime-content-field.c (gmime_content_field_ref):
(gmime_content_field_unref):
New reference mechanism for GMimeContentField objects.

* camel/camel-data-wrapper.c (_finalize):
Started implementing destructors.

* camel/camel-mime-part.c (_construct_from_stream):
* camel/gmime-content-field.c (gmime_content_field_write_to_stream):
* camel/camel-multipart.c (_construct_from_stream):
removed forgotten anarchic traces.

Go away memory leaks!

some test files

sync

more advanced RFC2047 encoder started.

1999-08-08  Robert Brady  <rwb197@ecs.soton.ac.uk>

* camel/gmime-rfc2047.c: more advanced RFC2047 encoder started.

fix. In MH, message number is not related to message position in folder.

1999-08-08  bertrand  <Bertrand.Guiheneuf@aful.org>

* camel/providers/MH/camel-mh-folder.c (_get_message):
fix. In MH, message number is not related to
message position in folder.

util func. (_get_message_count): implemented.

1999-08-08  bertrand  <Bertrand.Guiheneuf@aful.org>

* camel/providers/MH/camel-mh-folder.c (_is_a_message_file):
util func.
(_get_message_count): implemented.

updated some autogen doc stuff. Still don't understand warnings :(

1999-08-06  bertrand  <Bertrand.Guiheneuf@aful.org>

* devel-docs/camel/:
updated some autogen doc stuff.
Still don't understand warnings :(

* camel/camel-data-wrapper.c:
* camel/providers/MH/camel-mh-store.c:
* camel/url-util.c:
* camel/gmime-content-field.c:
* camel/camel-store.c:
various inline doc corrections.

* camel/camel-folder.c (get_message_count):
new method. Returns the number of message
in the folder.

encoder test.

1999-08-06  Robert Brady  <rwb197@ecs.soton.ac.uk>

* tests/test6.c: encoder test.

* camel/gmime-rfc2047.c: Fixed decoder bug : sequence
?= is not always the terminator for an encoded-string.

stat was not testing the good file. Fixed.

1999-08-06  bertrand  <Bertrand.Guiheneuf@aful.org>

* camel/providers/MH/camel-mh-folder.c (_list_subfolders):
stat was not testing the good file. Fixed.

test for RFC2047 decoder.

1998-08-06  Robert Brady  <rwb197@ecs.soton.ac.uk>

* tests/test5.c: test for RFC2047 decoder.

* camel/gmime-rfc2047.c: Improved RFC2047 decoder.

add debug information (_list_subfolders): test if first char in folder

1999-08-06  bertrand  <Bertrand.Guiheneuf@aful.org>

* camel/providers/MH/camel-mh-folder.c (_exists): add debug information
(_list_subfolders): test if first char in folder name is not '.'
before adding it to the folder list.

* camel/camel-store.c (_init):
disable session check temporarily
(_get_separator):
(_get_folder): new static func.
Fixed several oddities in class definition.

* camel/providers/MH/camel-mh-store.c (camel_mh_store_get_type):
parent type is CAMEL_STORE_TYPE not CAMEL_FOLDER_TYPE

* camel/camel-store.c:
prent class is CamelServiceClass not GtkObjectClass

* camel/url-util.c :
cosmetic changes + use of const when possible.
(find_host):
fix a bug: when there is no host and no port don't skip the '/'
all static find_* func are now named _func_*
(g_url_free): destructor func.
cache field has been disabled. Constructing the url string
won't be too slow and will occur rarely enough that we
do not need to add complexity to this code.

* camel/providers/MH/camel-mh-store.c:
parent class is CamelStorClass not GtkObjectClass

test mh provider.

1999-08-05  bertrand  <Bertrand.Guiheneuf@aful.org>

* tests/test4.c:
test mh provider.

* camel/providers/MH/camel-mh-folder.c (_get_message):
implemented

new method. (camel_folder_get_message): corresponding public call

1999-08-05  bertrand  <Bertrand.Guiheneuf@aful.org>

* camel/camel-folder.c (_get_message):
new method.
(camel_folder_get_message):
corresponding public call

Some notes.

1999-08-05  bertrand  <Bertrand.Guiheneuf@aful.org>

* camel/README.HACKING:
Some notes.

* camel/CODING.STYLE:
short note about coding style.

* camel/README.COPYRIGHT:
Note about copyright policy.

* camel/providers/MH/camel-mh-folder.c (_list_subfolders):
minor typo fixes.

indentation and cosmetic changes.

1999-08-04  bertrand  <Bertrand.Guiheneuf@aful.org>

* camel/gmime-rfc2047.c:
* camel/gmime-rfc2047.h:
indentation and cosmetic changes.

Fixed copyright issue. ;)

implemented.

1999-08-04  bertrand  <Bertrand.Guiheneuf@aful.org>

* camel/providers/MH/camel-mh-folder.c (_list_subfolders):
implemented.

finshed implementation (_delete_messages): implemented.

1999-08-04  bertrand  <Bertrand.Guiheneuf@aful.org>

* camel/providers/MH/camel-mh-folder.c (_delete):
finshed implementation
(_delete_messages): implemented.

added an implemention of RFC2047 (support for character sets other than

1999-08-04  Robert Brady  <rwb197@ecs.soton.ac.uk>

* camel/gmime-rfc2047.[ch]: added an implemention of RFC2047
  (support for character sets other than US-ASCII in MIME
  headers). Not actually called from anywhere yet.

my file I never commited

Started implementing mh store/folder as a test provider. Next will be pop3.

1999-08-03  bertrand  <Bertrand.Guiheneuf@aful.org>

* camel/providers/MH/camel-mh-folder.c (_create):
implemented.
(_delete): started implementation.

* camel/camel-folder.c (_get_folder): default implementation
calls camel_store_get_folder ().

* camel/providers/MH/camel-mh-folder.c (_init_with_store):
implemented.

* camel/camel-folder.h (struct _CamelFolder):
remove useless exist_on_store field.

* camel/camel-folder.c (_exists):
do not use exist_on_store field.

new public functions (_set_name): set full_path.

* camel/camel-folder.c (camel_folder_set_name):
(camel_folder_get_name):
new public functions
(_set_name): set full_path.

(_set_full_name):
(camel_folder_set_full_name):
commented out this functions def.
It would make things very difficult to handle, and would not
be very useful.

* camel/providers/MH/camel-mh-store.h:
* camel/providers/MH/camel-mh-store.c (camel_mh_store_set_toplevel_dir):
(camel_mh_store_get_toplevel_dir):
* camel/providers/MH/camel-mh-folder.c (_set_name):
* camel/providers/MH/camel-mh-folder.h:
use (gchar *) instead of (GString *) everywhere.
use const when necessary.

* camel/camel-folder.h (struct _CamelFolder):
removed unused message_list field.

There is a probleme here. We can not allow mime part content-type field

* camel/camel-mime-part.c (_set_content_object):
There is a probleme here. We can not allow mime part
content-type field and content_object mime-type to be
different. I thus chosed to set mime part object
content field to be freed (if necessary) and set
to be a pointer to content_object mime type
field.
(_construct_from_stream): set content_object mime type
to be the same as mime_part's one. This is necessary
because we use _set_content_type.

This two things are a bit hackish ansd may need
to be redesigned.

Multipart Mime message parsing works with plain text parts. Woohooo :))))
Making it work with other types is now just a matter of writing
the various data wrappers. And display them will just be a matter of writing
the good bonobo components.

1999-08-03  bertrand  <Bertrand.Guiheneuf@aful.org>

* camel/camel-simple-data-wrapper.c (_construct_from_stream):
more debugging output +  nb_bytes_read is now a signed int
to avoid bug when eos is encountered.

* camel/camel-mime-part.c (_construct_from_stream):
sync to data_wrapper_repository function name changes.
Use default "text/plain" type when conten-type field
is not found. (following RFC 2046 spec).

* camel/data-wrapper-repository.c (data_wrapper_repository_set_data_wrapper_type):
(data_wrapper_repository_get_data_wrapper_type):
change function name prefix (s/data_wrapper/data_wrapper_repository/)

* camel/camel-multipart.c (_read_part):
add `\n` at eol but not before boundary.

* camel/gmime-utils.c (get_header_table_from_stream):
correct implementation of end of stream detection.

Update logs and sync home/work versions.

* camel/camel-multipart.c (_read_part):
use a stream to store the part instead of GString.

* camel/camel-mime-part.c (camel_mime_part_set_text):
set data wrapper content type to "text/plain".

* camel/camel-stream-mem.c:
* camel/camel-stream-mem.h:
new memory buffer based stream.

* camel/camel-stream-fs.c (_seek):
implementation for file system based stream.

* camel/camel-stream.c (camel_stream_seek):
new method.

* camel/camel-stream-fs.c (camel_stream_fs_class_init):
pass CamelStreamFsClass instead of CamelStreamClass.

Anther sync

sync before going home

new memory buffer based stream.

* camel/camel-stream-mem.c:
* camel/camel-stream-mem.h:
new memory buffer based stream.

* camel/camel-stream-fs.c (_seek):
implementation for file system based stream.

* camel/camel-stream.c (camel_stream_seek):
new method.

* camel/camel-stream-fs.c (camel_stream_fs_class_init):
pass CamelStreamFsClass instead of CamelStreamClass.

do not return NULL when line is empty.

* camel/gmime-utils.c (gmime_read_line_from_stream):
do not return NULL when line is empty.

* camel/camel-multipart.c (_read_part): return true when end
of multipart is found, not the contrary

Adding missing files

merge a lot of work. Rough ChangeLog will follow

sync before going home

Do not use GStrings any more. Added assertion code. Cosmetic reformating

1999-07-15  bertrand  <Bertrand.Guiheneuf@aful.org>

* camel/url-util.c:
* camel/url-util.h:
Do not use GStrings any more.
Added assertion code.
Cosmetic reformating

* AUTHORS:
* ChangeLog:
Changed my email address.

Implemented base64 encoder based on CamelStreams. Should the

1999-07-13  Miguel de Icaza  <miguel@gnu.org>

* camel/gmime-base64.c (gmime_encode_base64): Implemented base64
encoder based on CamelStreams.  Should the encoder/decoder be a
Stream itself?

* camel/gmime-utils.c: include config.h here.
* camel/url-util.c: ditto.
* camel/gstring-util.c: ditto.
* camel/gmime-content-field.c: ditto.
* camel/camel-stream.c: ditto.
* camel/camel-stream-fs.c: ditto.
* camel/camel-store.c: ditto.
* camel/camel-simple-data-wrapper.c: ditto.
* camel/camel-session.c: ditto.
* camel/camel-service.c: ditto.
* camel/camel-mime-part.c: ditto.
* camel/camel-mime-message.c: ditto.
* camel/camel-log.c: ditto.
* camel/camel-data-wrapper.c: ditto
* camel/camel-folder.c: ditto.

* camel/camel-stream.c (camel_stream_write): Moved api
documentation to the places that they document.
(camel_stream_class_init): Virtual classes do not need to have a
default implementation.  So null them all.
(camel_stream_write): Return value from write.
(camel_stream_available): implement.
(camel_stream_write_strings): documented.

* devel-docs/query/virtual-folder-in-depth.sgml: Small
reformatting

1999-06-28  bertrand  <Bertrand.Guiheneuf@inria.fr>

* tests/test2.c (main): now use
CamelDataWrapper::contruct_form_stream to test

now use CamelDataWrapper::contruct_form_stream to test message parsing

1999-06-28  bertrand  <Bertrand.Guiheneuf@inria.fr>

* tests/test2.c (main): now use
CamelDataWrapper::contruct_form_stream to test
message parsing

* camel/camel-data-wrapper.c:
* camel/camel-data-wrapper.h:
construct_from_stream no longer has maximimum size arg.

* camel/camel-mime-part.c (_construct_from_stream): new.
Construct the mime_part from a stream.

new field (content_type) and associated methods. (camel_mime_part_init):

1999-06-28  bertrand  <Bertrand.Guiheneuf@inria.fr>

* camel/camel-mime-part.c:
new field (content_type) and associated methods.
(camel_mime_part_init): initialize content_type field.
(_parse_header_pair): now set content_type MimePart field
instead of using DataWrapper Mime typing facility.

1999-06-28  bertrand  <Bertrand.Guiheneuf@inria.fr>

* camel/camel-data-wrapper.h:
s/content_type/mime_type/

1

new class. Simple implementation of a data wrapper: simply keeps the

1999-06-24  bertrand  <Bertrand.Guiheneuf@inria.fr>

* camel/camel-simple-data-wrapper.[ch]:
new class. Simple implementation of a data wrapper:
simply keeps the stream result in a byte array.

* camel/camel-mime-part.c (_parse_header_pair): added a warning.
Have to think about the correct way to store content type stuff.

Various patches from Ulrich Drepper <drepper at cygnus.com>

1999-06-24  bertrand  <Bertrand.Guiheneuf@inria.fr>

* camel/camel-mime-message.c (_write_one_recipient_to_stream):
includes gmime-utils header.
patch from Ulrich Drepper <drepper at cygnus.com>
set separator string in write_header_with_glist_to_stream()

* camel/camel-log.c (camel_log):
patch from Ulrich Drepper <drepper at cygnus.com>
Do not use stderr in initialization of logfile descriptor.

* camel/camel-stream-fs.c (camel_stream_fs_new_with_name):
patch from Ulrich Drepper <drepper at cygnus.com>
initialize mode field in open().