2003-08-15 Not Zed <NotZed@Ximian.com>
** See bug #47634.
* tests/lib/messages.c (test_message_compare): check
write_to_stream returns.
(message_dump_rec): helper to dump message structure.
* camel-mime-part-utils.c
(simple_data_wrapper_construct_from_parser): dont set content
encoding here.
(camel_mime_part_construct_content_from_parser): set it here
instead, on every part. basically same as setting the
mime_type_field always.
* camel-multipart-signed.c (camel_multipart_signed_class_init):
* camel-mime-message.c (camel_mime_message_class_init):
* camel-multipart.c (camel_multipart_class_init): override
decode_to_stream to always do the same as write_to_stream, since
we can never be encoded.
svn path=/trunk/; revision=22252
* tests/message/test4.c: New test suite for the mime parser (which
is where the below 2 fixes were noticed).
* camel-mime-parser.c (folder_boundary_check): Calculate 'len' by
subtracting the boundary start from inend rather than 'atleast'.
(folder_scan_content): Calculate 'inend' differently depending on
the EOF state.
2003-08-08 Jeffrey Stedfast <fejj@ximian.com>
* camel-mime-filter-tohtml.c (html_convert): Rather than checking
*inptr == '\n', check inptr >= inend - this gets rid of an Invalid
Read report from valgrind.
* camel-mime-part.c (write_to_stream): Don't necessarily re-encode
just because the encodings differ. Need to look into making it so
that message/rfc822 and multipart parts ignore the
Content-Transfer-Encoding header and just keep their 'encoding'
bits set to DEFAULT.
2003-08-05 Jeffrey Stedfast <fejj@ximian.com>
* providers/imap/camel-imap-folder.c (get_content): Updated.
* camel-mime-message.c (camel_mime_message_init): Don't override
the mime_type here.
(process_header): Updated to use CamelDataWrapper's mime_type
field.
(find_best_encoding): Same.
(best_encoding): Here too.
* camel-digest-folder.c (camel_digest_folder_new): Updated for
CamelMimePart::content_type change.
* camel-mime-part.c (camel_mime_part_init): Override our parent
class's default mime_type.
(camel_mime_part_finalize): Don't need to unref the content_type
anymore.
(process_header): Updated to use CamelDataWrapper's mime_type
field.
(camel_mime_part_set_filename): Same.
(camel_mime_part_get_filename): Same.
(camel_mime_part_get_content_type): Same.
(set_content_object): Here too.
(write_to_stream): Updated.
(construct_from_parser): Updated.
* camel-mime-part.h: Remove the content_type field.
2003-07-31 Jeffrey Stedfast <fejj@ximian.com>
* tests/lib/messages.c (test_message_compare_content): If the
chunks differ, perform a hexdump on the data being compared so
that we may analyse it easier.
* camel-multipart-signed.c (write_to_stream): Return ssize_t.
* camel-mime-utils.h: Added the CamelMimePartEncodingType enum
here.
* camel-mime-part.h: Removed the CamelMimePartEncodingType enum
from here.
* camel-mime-part.c (write_to_stream): Updated to return
ssize_t. Also minor changes to only re-encode the content stream
if the charset or encoding changed (this way we write it out in
the original raw form if nothing changed).
* camel-mime-part-utils.c
(simple_data_wrapper_construct_from_parser): Drastically
simplify. We no longer scan html content to try and find the
charset, nor do we care about converting the content to UTF-8 and
handling broken windows charsets.
* camel-mime-message.c (find_best_encoding): Use
decode_to_stream() here. Also updated to not assume the content
charset is UTF-8 since it is very likely not the case anymore
since data-wrappers no longer are converted to UTF-8 at parse
time.
* camel-folder-summary.c (summary_build_content_info_message): Use
decode_to_stream instead here too.
* camel-folder-search.c (match_words_1message): Use
decode_to_stream instead of write_to_stream so we can search the
contents.
* camel-data-wrapper.c (camel_data_wrapper_init): Set the default
encoding to DEFAULT.
(write_to_stream): Updated to return ssize_t
(camel_data_wrapper_decode_to_stream): New virtual function to
decode a data wrapper to a stream (results in nearly identical
behaviour to the old write_to_stream method).
(decode_to_stream): Default implementation of above virtual
method. Decodes base64/qp/etc streams.
* camel-data-wrapper.h: Removed the rawtext bit and added an
encoding member.
svn path=/trunk/; revision=22171
2002-09-04 Not Zed <NotZed@Ximian.com>
* tests/folder/test4.c (main): clear nonfatal stuff.
* tests/folder/test6.c (main): check inbox and 'another folder'
because some servers do different things w/ inbox.
* providers/imap/camel-imap-message-cache.c
(camel_imap_message_cache_get): Dont try to open the directory
instead of an empty "" uid.
* providers/imap/camel-imap-folder.c (imap_get_message): we dont
want to g_return_if_fail, we need to set an exception and always
handle the case.
(get_message_simple): Set an exception if we get a construct
failure.
(imap_refresh_info): if we're refreshing inbox, force a reselect.
this is required for at least cryus. CHECK doesnt work either :(
* tests/lib/folders.c (test_folder_basic): for non-local stores,
the folder will have an extra ref for selection, take this into
account when checking ref leaks.
(test_folder_message_ops): Dont try to delete folder with messages
in it, it works generally with imap. also, change params so we
can test different mailbox types.
(test_folder_message_ops): disconnect remote services before
finishing off. doesn't need to stricly but makes ref count
checking more accurate.
(test_folder_message_ops): removed explicit remote sync, imap does
it itself now ...
* providers/imap/camel-imap-store.c:
(camel_imap_store_finalize): call service_disconnect, so it isn't
called later in the finalise chain, to properly cleanup on exit.
svn path=/trunk/; revision=17962
2002-08-21 Jeffrey Stedfast <fejj@ximian.com>
* camel-session.c (get_service): Don't register the noop timeout
here anymore, we're gonna handle all this in the mailer now.
(camel_session_register_timeout): Removed.
(camel_session_remove_timeout): Removed.
svn path=/trunk/; revision=17830
2002-08-19 Jeffrey Stedfast <fejj@ximian.com>
* tests/lib/camel-test.c (camel_test_init): Updated to pass the
needed arguments to camel_init() and to init gthreads before camel
(since camel expectes it to be initialised already).
* tests/lib/session.c: #include <config.h> needed for
ENABLE_THREADS so that sizeof (CamelSessionClass) gives the
correct size.
* tests/*: Updated to work with camel API changes.
svn path=/trunk/; revision=17805
by sendmail to strip the BCC from the email while still sending to the BCC
recipients.
camel:
2001-06-28 Peter Williams <peterw@ximian.com>
* camel-mime-filter-stripheader.c: New file. Filter that strips a
header from mime output.
* camel-mime-filter-stripheader.h: New file. Header for the above.
* providers/smtp/camel-smtp-transport.c (smtp_data): Use the stripheader
filter to remove the "Bcc" header.
* Makefile.am: Add the stripheader files.
* tests/lib/Makefile.am (INCLUDES): Get this to compile again.
* tests/mime-filter/test-stripheader.c: New file. Test suite for
the CamelMimeFilterStripHeader.
* tests/mime-filter/Makefile.am: New test section: mime filters.
mail:
2001-06-28 Peter Williams <peterw@ximian.com>
* mail-ops.c (mail_send_message): Revert fejj's Bcc header removal;
this unsets the BCC recipients and so doesn't send to the Bcc'd
people at all.
svn path=/trunk/; revision=10576
2001-06-18 Not Zed <NotZed@Ximian.com>
* camel-folder-summary.c
(camel_folder_summary_info_new_from_parser): Only create a missing
uid if we have indexing turned on.
* camel-lock-helper.c (setup_process): Function to setup
process/sanity/security checks. Change to the real uid as soon as
we can.
(lock_path): First try to lock as the real uid, if that fails, try
the root uid.
(unlock_id): Unlock as the uid we created the lock as.
* Makefile.am (INCLUDES): Added -DCAMEL_SBINDIR for lock helper
location.
* providers/local/camel-spool-folder.c (spool_lock): Implemented,
using lock helper locking. Need to work out if the locking
requires a root created lock?
(spool_unlock): Likewise.
2001-06-15 Not Zed <NotZed@Ximian.com>
* camel-lock-helper.c: Setuid Lock helper process. Creates and manages
.locks, keeping them active, removing them, etc. What real perms it
needs is a little system dependent.
2001-06-14 Not Zed <NotZed@Ximian.com>
* providers/local/camel-maildir-store.c (get_folder_info): Implement.
(scan_dir): Does the work of scanning for maildir directories.
2001-06-13 Not Zed <NotZed@Ximian.com>
* providers/local/camel-spool-store.c (get_folder_info):
Implemented, just returns a hardcoded INBOX folder.
(free_folder_info): implemented, free's the 1 possible level of
folder info.
* providers/local/camel-spool-folder.c
(camel_spool_folder_construct): Set the real unread message
count on the folder_created thing.
svn path=/trunk/; revision=10261
2001-06-12 Not Zed <NotZed@Ximian.com>
* providers/local/camel-spool-*.[ch]: A new provider, for spool
mailboxes. Mostly a cut and paste of the mbox code, but not a
subclass CamelLocal*. Not tested a lot, doesn't lock yet, use
with extreme caution.
* tests/lib/folders.c (test_folder_message_ops): Added spool arg,
spool folders can't be deleted, renamed, etc.
(test_folder_basic): Same.
* tests/folder/test2.c (main): Added checks for spool type.
* tests/[message|stream|folder|misc|smime]/Makefile.am (LDADD):
Added db3 flags, so make check compiles, doesn't run though.
2001-05-24 Not Zed <NotZed@Ximian.com>
* providers/local/camel-local-provider.c
(camel_provider_module_init): Added spool provider.
svn path=/trunk/; revision=10198
* tests/lib/messages.c (test_message_read_file): Fix an fd leak
* tests/lib/session.c, tests/lib/session.h: a CamelSession
subclass for the test programs.
* tests/lib/Makefile.am: include session.[ch]
* tests/folder/test*.c: Use a CamelTestSession from libcameltest
instead of cut+pasting everywhere.
* tests/misc/url.c (main): Update for a camel_url_new change at
some point.
* tests/*/.cvsignore: Add stuff.
* camel-mime-utils.c (rfc2047_encode_word): Fix a silly ==/!=
mixup.
svn path=/trunk/; revision=10023
Index: ChangeLog
===================================================================
RCS file: /cvs/gnome/evolution/camel/ChangeLog,v
retrieving revision 1.684
diff -r1.684 ChangeLog
0a1,34
> 2001-01-17 Not Zed <NotZed@Ximian.com>
>
> * camel-folder.c (free_summary): Call
> camel_folder_summary_array_free() to do the work.
> (get_summary): Use camel_folder_summary_array() to get the array
> atomically. These fixes allow folder/test8 to work again, and fix
> a sort of race where the summary size can change while we were
> making a copy of it.
>
> * camel-folder-summary.c (camel_folder_summary_array): Get the
> summary array atomically, so it can't contain empty records.
> (camel_folder_summary_array_free): And free it.
>
> * tests/lib/camel-test.c (die): If we are verbose & in threads,
> then goto sleep so we can debug.
>
> * tests/folder/test8.c (worker): Add a missing pull() for
> comnparing content.
>
> * camel-filter-search.c: Fix the symbol table, so match-all is an
> immediate function, as it should be.
>
> * tests/folder/test9.c (main): New test, tests some filtering
> things.
>
> * tests/message/test3.c (main): Dont use a boundary string with
> spaces in it. Folding can corrupt it. Maybe the folding isn't
> working entirely right, but anyway.
>
> * camel-session.c: Debug out the debug.
>
> * camel-filter-driver.c (camel_filter_driver_filter_folder): Plug
> a messageinfo leak.
>
1a36,94
>
> * camel-filter-search.c (header_exists): Changed to support
> multiple args (or'd together).
> (header_contains): Cleaned up to match the search code. Why did
> fejj change it? I'll never know.
> (header_matches):
> (header_starts_with):
> (header_ends_with): Big cleanup of fejj's "i'm the cut & paste
> king" code. Also properly handle or'ing of additional args to
> match what the folder-search code should do.
> (check_match): New function which does the annoying matching
> stuff (for header matches).
> (check_header): Similarly, handles or'ing of the matches together.
> (header_contains):
> (header_matches):
> (header_starts_with):
> (header_ends_with): Call check_header to do the actual work.
> (header_soundex): And here too.
> (match_all): Yeah like match-all isn't passed expression results,
> its passed expression terms. Fix this so match-all works like it
> should, by executing the contained expression.
> (message_body_contains): Copied directly from
> camel-folder-search.c, a more robust/faster/simpler body search
> code.
> (mime_part_matches): Removed entirely.
> (handle_multipart): Removed entirely.
> (build_match_regex): Copied from camel-folder-search. Builds a
> set of simple strings into a regex pattern that matches any of
> them (for faster & simpler matching). Expanded to accept regex
> patterns itself, so it can merge them together.
> (body_contains): Use build match/match message to match using a
> built regex.
> (body_regex): Likewise, this time we tell it we're building a
> regex though.
> (header_full_regex): Use build_match_regex to take the drudgery
> out of it, and expand it to handle multiple regex's at once.
> (get_full_header): slightly cleaner (well i dunno, the sprintf
> stuff just got to me).
> (header_regex): Cleaned up to use build_match_Regex too, and to
> properly check types.
> (filter_message_search): Just allocate 'fms' on the stack.
>
> * camel-filter-driver.c (camel_filter_driver_finalise):
> (camel_filter_driver_init):
> (camel_filter_driver_class_init):
> (camel_filter_driver_get_type): Changed from gtk object to camel
> object.
> (camel_filter_driver_add_rule): New function to add a rule to be
> processed in sexp form.
> (camel_filter_driver_init): Init the rules list.
> (camel_filter_driver_finalise): Clear the rules/rules list.
> (camel_filter_driver_filter_message): Scan rules list directly
> rather than creating on the fly.
>
> * Makefile.am (libcamelinclude_HEADERS): Added camel-filter-driver.h
> (libcamel_la_SOURCES): Added camel-filter-driver.c, code taken
> from filter-driver, which can drive, uh, filters based on sexp's.
> (libcamelinclude_HEADERS):
> (libcamel_la_SOURCES): Added camel-filter-search.[ch]
svn path=/trunk/; revision=7560
2000-12-24 Not Zed <NotZed@HelixCode.com>
* providers/imap/camel-imap-search.c (imap_body_contains): Lock
the command channel while searching.
(imap_body_contains): If performing a whole uid search, then add
references to our own summary items, dont look it up in the
folder. This way they can't vanish unexpectedly.
* providers/imap/camel-imap-folder.h (CamelImapFolder): Added a
private field.
* providers/imap/camel-imap-private.h: Added lock for imap
searches.
* Merge from camel-mt-branch.
* providers/imap/camel-imap-folder.c (imap_update_summary): Merge
fix, use the folder->summary.
(imap_get_message_flags, imap_set_message_flags,
imap_get_message_user_flag, imap_set_message_user_flag): Removed
again.
(camel_imap_folder_init): Setup private data/lock.
(imap_finalize): Free private data/search lock.
(imap_search_free): Lock the search_lock.
(imap_search_by_expression): Lock the search lock when using the
search object. Also copy/ref hte summary, rather than getting it
directly.
(imap_refresh_info): Free any info lookups. Use folder->summary
not imap_folder->summary. And lock around commands.
svn path=/trunk/; revision=7150
2000-12-06 Not Zed <NotZed@HelixCode.com>
* tests/lib/camel-test.h (check): Change line no format so that
emacs can detect it.
svn path=/trunk/; revision=6848
2000-12-05 Not Zed <NotZed@HelixCode.com>
* camel-seekable-substream.c (stream_flush): stream_flush does
make sense for a substream afterall (if you have a stream_write).
(stream_write): Implement this.
(stream_seek): Change the STREAM_END behaviour to be more sane.
if bounded go from the end of the bound, if unbounded, go from the
end of the parent stream.
* camel-stream-mem.c (stream_read): Dont return error if reading
past the end of data, just return 0.
* camel-stream-fs.c (camel_stream_fs_init): Initialise the stream
to be unbound.
(stream_seek): Fix the logic when seeking from the end of an
unbounded stream.
(camel_stream_fs_new_with_fd): If the fd is invalid (-1), then
return NULL immediately.
(stream_seek): Range check a SEEK_END so it fits within
bound_start.
2000-12-01 Not Zed <NotZed@HelixCode.com>
* tests/lib/folders.c (test_folder_basic): New test to perform
basic store operations on folders (taken from folders/test1).
(test_folder_message_ops): Tkane the guts out of folders/test2.
svn path=/trunk/; revision=6790
2000-11-30 Not Zed <NotZed@HelixCode.com>
* providers/local/camel-mbox-folder.c (mbox_get_message): Remove
assertion that content is there, when it no longer can be.
* camel-folder-summary.h: Removed pos/bodypos/endpos from
camelmeessagecontentinfo.
(CamelMessageFlags): Added an attachments flag.
* providers/local/camel-local-summary.h: Added load virtual
function.
* tests/lib/folders.c (test_message_info): Accessors.
(test_folder_message): "
* camel-folder-thread.c (get_root_subject): Fix accessors.
(dump_tree_rec): "
* camel-folder-search.c (camel_folder_search_execute_expression):
Accessors for messageinfo.
(search_match_all): "
(search_header_contains): "
(search_header_contains): "
(search_body_contains): "
(camel_folder_search_execute_expression): Use mepool_strdup.
* providers/local/camel-mbox-summary.c (summary_update): Accessors
for messageinfo.
(mbox_summary_sync_full): "
* providers/local/camel-mh-summary.c (remove_summary): Accessors
for messageinfo.
(mh_summary_check): "
(mh_summary_sync_message): "
(mh_summary_sync): "
* providers/local/camel-mh-folder.c (mh_append_message): Use
accessor for uid.
* providers/local/camel-local-summary.c
(local_summary_decode_x_evolution): Use accessor to uid.
(local_summary_encode_x_evolution): Likewise.
(message_info_new): And here.
(camel_local_summary_load): Call virtual load function.
(local_summary_load): Default load function, load summary.
(camel_local_summary_load): Check file exists before trying to
load.
(camel_local_summary_construct): Turn off building content info!
(CAMEL_LOCAL_SUMMARY_VERSION): Bump, since we dont build content
info anymore.
(camel_local_summary_load): After a successful load/check, do a
save too so we dont have to go through it again randomly.
* providers/nntp/camel-nntp-utils.c (get_XOVER_headers): Use
accessors for messageinfo.
* providers/nntp/camel-nntp-folder.c (nntp_folder_get_uids): Use
accessors for uid.
* providers/imap/camel-imap-folder.c (imap_refresh_info): Use
accessor for uid.
(imap_sync): Likewise.
(imap_get_uids): Likewise.
(imap_update_summary): And here.
* providers/vee/camel-vee-folder.c (vfolder_remove_match): Use
accessor for uid.
(vfolder_add_match): Handle estrv stuff.
(vfolder_change_match): Accessor for uid.
(get_real_message): "
(vee_get_uids): "
(vee_folder_build): " + estrv.
(vee_folder_build_folder): "
* providers/local/camel-maildir-folder.c (maildir_append_message):
Use acccessors for uid's.
(maildir_get_message): Here too.
* providers/local/camel-maildir-summary.c
(camel_maildir_summary_init): Setup the string count for us.
(message_info_new): Access the string array directly.
(message_info_free): No need to free string if using array.
(camel_maildir_summary_info_to_name): Use accessor to get to uid.
(remove_summary): And here.
(maildir_summary_check): Likewise.
(maildir_summary_sync): And here.
(maildir_summary_load): Load up a cache of uid->filename mappings
before loading the actual summary file. This saves us having to
waste the diskspace storing the filenames in the summary itself,
and also helps us sync the summary better on load.
(message_info_load): If we have the load_map setup, and the uid
exists, then set the filename cache from it, and update the flags
from the name, incase our summary mismatches it.
* camel-folder-summary.c (camel_folder_summary_init): Setup string
count for compressed info record. An optional compile mode which
stores all strings for a given messageinfo into a packed array,
which should save 36-50 bytes/record.
(camel_folder_summary_info_new): Init the string array.
(message_info_new): Set the string array items, as required.
(message_info_load): And here too.
(message_info_save): Use accessors to get to strings.
(message_info_free): Free strings as one.
(camel_message_info_dup_to): Handle packed array case.
(camel_folder_summary_add): Use accessors. And pack the strv
before storing it.
(summary_assign_uid): New function to assign a unique uid to a
message, if it doesn't have one.
(camel_folder_summary_add): Call assign_uid instead of doing it
ourselves.
(camel_folder_summary_info_new_from_parser): "
(camel_folder_summary_info_new_from_message): "
(camel_folder_summary_encode_string): constify.
(camel_folder_summary_encode_token): "
(summary_build_content_info_message): Fix accessors to messageinfo.
(CAMEL_FOLDER_SUMMARY_VERSION): Bumped, for removal of
contentinfo->pos data.
(camel_folder_summary_info_new_from_parser): Calculate the size
based on the parser position, not the removed contentinfo stuff.
(camel_folder_summary_info_new_from_message): Remove size stuff.
(camel_folder_summary_offset_content): Removed, no longer means anything.
(content_info_new):
(content_info_load):
(content_info_save):
(summary_build_content_info): Remove stuff for contentinfo->pos*.
(summary_build_content_info): Take a msginfo argument, set
attachments flag if we find any attachments.
(summary_build_content_info_message): set attachments flag if we
find any attachments.
(camel_folder_summary_info_new_from_parser): Always scan the
content info, even if we dont save it.
(camel_folder_summary_info_new_from_message): And here too.
(summary_build_content_info): Only create the contentinfo stuff if
we have it turned on, otherwise just parse and discard.
(summary_build_content_info_message): Likewise.
svn path=/trunk/; revision=6731
2000-11-29 Not Zed <NotZed@HelixCode.com>
* tests/message/test2.c (main): Add a missed unref.
* camel-stream-mem.c (camel_stream_mem_set_buffer): We must set
ourselves as the owner of the byte-array.
Weird, someone has modified this file (its been reindented), but i
can't see any changelogs ...
* tests/lib/messages.c (content_finalise): Fix memleak in tester,
free byte array when our content object is deleted.
* camel-folder-search.c (camel_folder_search_finalize): Yeah
great, so the sexp is a gtk object, not a camel object. Isn't
that going to be fun to fix?
* camel-session.c (camel_session_finalise): Free the storage path.
* providers/local/camel-local-store.c (camel_local_store_init): If
store->folders is setup, free it first, then overwrite. Hmm,
this seems a bit crappy to me.
* camel-store.c (camel_store_init): Dont setup store->folders if
its already setup.
* camel-exception.c (camel_exception_setv): Removed a memleak. no
need to strdup after a strdup_printf!!!
* camel-address.c (camel_address_finalize): Free the address
ptrarray, once finished.
* providers/local/camel-local-folder.c (local_finalize): Make sure
we dont leave the folder locked on close.
(local_finalize): Free summary/search.
* providers/local/camel-mh-summary.c (mh_summary_next_uid_string):
Small memleak, always free name after using it.
* camel-mime-part.c (set_content_object): Free txt after setting
the header.
* providers/local/camel-maildir-summary.c (maildir_summary_check):
Fix a memleak, close the dir after scanning new.
(message_info_free): Added so we can free the filename cached in
the messageinfo.
(camel_maildir_summary_finalise): Free the hostname.
* tests/folder/test[12].c (main): Clear out camel-test before
starting.
* providers/local/camel-mbox-summary.c (mbox_summary_sync_quick):
Because encode_x_evolution folds the line (sigh, because
encode_param does, unlike every other function in
camel-mime-utils), unfold the encoded result before comparing.
(mbox_summary_sync_quick): Another small memleak, free xevnew once
finished with it.
* camel-mime-utils.c (header_decode_quoted_string): Removed a
redundant check for c=0.
(header_unfold): New function to un-fold headers.
* providers/local/camel-local-summary.c
(local_summary_encode_x_evolution): some problems with encoding
tags, using the wrong output strings.
(local_summary_encode_x_evolution): We dont need to append a ;
either, param_list_format_append() will do it for us.
`
svn path=/trunk/; revision=6711
2000-11-28 Not Zed <NotZed@HelixCode.com>
* providers/local/camel-maildir-summary.c
(camel_maildir_summary_init): Set the info size's properly, oops!
* tests/lib/folders.[ch]: Folder testing helpers.
* tests/folder/test2.c: Test basic message ops on folders.
* tests/folder/test1.c (main): Test basic folder ops on (local)
stores.
* providers/local/camel-local-provider.c
(camel_provider_module_init): Removed some debug.
* providers/local/camel-maildir-folder.c
(camel_maildir_folder_class_init): fix parent class.
* providers/local/camel-mh-folder.c (camel_mh_folder_class_init):
Fix parent class (damn cut & paste).
* providers/local/camel-maildir-store.c (get_folder): Call parent
impl.
(camel_maildir_store_class_init): Fix parent class setup.
(delete_folder): Check the folder exists before trying to delete
it.
(delete_folder): Try and make the delete operation atomic/rollback
failures. e.g. if one directory isn't empty, then create the
other empty ones back. Also clear the tmp directory fully first.
* providers/local/camel-mbox-store.c (get_folder): Call parent
impl.
(camel_mbox_store_class_init): parent class is camel_local_store,
not camel_folder, oops.
(delete_folder): Return an error if it doesn't exist, rather than
covering it up.
* providers/local/camel-mh-store.c (get_folder): Call parent impl.
(camel_mh_store_class_init): fix parent class setup.
(delete_folder): Error if it doesn't exist now.
* camel-folder.c (camel_folder_move_message_to):
(camel_folder_copy_message_to): Added warnings as these functions
are going to be removed later.
* camel-store.c (camel_store_get_root_folder): Fix for an early
api change. We want CAMEL_STORE_FOLDER_CREATE, not TRUE, since
its a flag.
(camel_store_get_default_folder): And here too.
* providers/local/camel-local-store.c (xrename): Handle renaming
folders differently to renaming files.
(get_default_folder_name): local stores dont have a default
folder, so make it so. Or at least, it doesn't seem to make sense
to have one.
(get_root_folder_name): Same for root.
(get_folder): Added parent implementation, that makes sure the
service path exists, if we are creating a new folder (but doesn't
create the folder).
2000-11-27 Not Zed <NotZed@HelixCode.com>
* providers/local/camel-local-store.c (xrename): Fixed races. Use
link/unlink, rather than rename, to properly detect overwriting
another file. And allow some files to be missing.
* providers/Makefile.am: Removed mh, mbox, added local, to the default.
svn path=/trunk/; revision=6693
* tests/data/getaddr.pl: little util to scan mailboxes for any and
every address they contain.
* tests/message/test2.c (main): Added a bunch of stuff to test
decoding/reencoding/etc of internationalised addresses.
* tests/message/lib/address-data.h: Copy of some unicode/other
testing data. **Beware** of editing this file in emacs, it'll
probably try and convert all the characters to something
unusable.
* tests/lib/camel-test.c (camel_test_break): Add a debugger hook
point.
* camel-mime-utils.c (quoted_encode): Check for space and convert
to _ separately.
(header_decode_mailbox): Fixed the 'check comments for realname'
code, problem was the domain getting code was skipping all
whitespace/comments before we could get a look-in. This is
approximate but fairly robust.
(header_decode_text): Dont use the c-type isspace func here, we
want a specific whitespace only.
(header_decode_text): If we have decoded words next to each other,
do not insert whitespaces between them, which is what rfc2047 requires.
(header_decode_text): Make c unsigned too.
svn path=/trunk/; revision=6658
2000-11-24 Not Zed <NotZed@HelixCode.com>
* tests/README: Added at least some explanation of all this stuff.
* tests/lib/camel-test.h (check_msg): Added a non-gcc version of
the fail command, we dont get the expression that failed, but no
matter. Should be (more) portable now.
(check, check_msg): Put the file/lineno in the default message.
svn path=/trunk/; revision=6657
2000-11-24 Not Zed <NotZed@HelixCode.com>
* Makefile.am (SUBDIRS): Add tests.
* camel-mime-filter-basic.c (filter): Well, I'll add the extra
bytes here too, lathough not strictly needed, might save a
re-malloc when we get to complete().
* camel-mime-filter-charset.c (filter): Make sure we have room if
we only convert very short data.
(complete): and here too.
* tests/Makefile.am: Initial test harness & tests. Requires gcc
for this.
* camel-internet-address.c (d): Turn off debug.
* camel-charset-map.c (camel_charset_step): Oops, & masks for set
intersection, not | them. Dunno how this got even close to
working.
2000-11-23 Not Zed <NotZed@HelixCode.com>
* camel-mime-filter-basic.c (filter): For base64 encoding, the
output size for 0, 1, or 2 bytes of input can exceed input*2, so
make sure we account for that as well.
(complete): And here.
(complete): Similarly for qp encoding, if we have a trailing
space, we need some extra bytes (not needed for 'filter()', as any
such bytes are stored in state/save).
* camel-mime-utils.c (quoted_decode_step): Removed fixme not required.
(quoted_encode_close): Dont append a trailing afterall. Otherwise
a pass through the encode/decode will grow the message each time.
svn path=/trunk/; revision=6656
