]> git.ipfire.org Git - thirdparty/dovecot/core.git/log
thirdparty/dovecot/core.git
8 years agoimap-login: Remove duplicate spaces between ID x-forward parameters
Timo Sirainen [Wed, 14 Jun 2017 08:47:00 +0000 (11:47 +0300)] 
imap-login: Remove duplicate spaces between ID x-forward parameters

8 years agoimap-login: Fix crash in ID if trusted client sends NIL value to internal field
Timo Sirainen [Wed, 14 Jun 2017 07:50:38 +0000 (10:50 +0300)] 
imap-login: Fix crash in ID if trusted client sends NIL value to internal field

Fixes:
Panic: imap-login: file imap-login-client.c: line 215 (cmd_id_handle_keyvalue): assertion failed: (client_id_str == !client_id_reserved_word(key))

8 years agolib-index: Fix cache lookups from uncommitted transactions with in-memory indexes
Timo Sirainen [Tue, 13 Jun 2017 19:14:11 +0000 (22:14 +0300)] 
lib-index: Fix cache lookups from uncommitted transactions with in-memory indexes

This especially fixes mail_get_header_stream() with imapc and
imapc_features=fetch-headers, because it works by first fetching the
specific headers and putting them to the cache transaction, then later on
relying on index_mail_get_header_stream() to get them from the transaction.
Before this fix they wouldn't be found, which triggered another unnecessary
FETCH BODY.PEEK[HEADER].

8 years agolib-mail: istream-header-filter - Fix HEADER_FILTER_ADD_MISSING_EOH with empty input
Timo Sirainen [Tue, 13 Jun 2017 18:54:00 +0000 (21:54 +0300)] 
lib-mail: istream-header-filter - Fix HEADER_FILTER_ADD_MISSING_EOH with empty input

The result for "" should be "\n", not "\n\n". The second "\n" would belong
to the mail body.

This fixes calculating hashes for incremental dsync when mail didn't have
Date or Message-ID headers, resulting in e.g.:

Warning: Deleting mailbox 'INBOX': UID=1 already exists locally for a different mail: Headers hashes don't match (e1c06d85ae7b8b032bef47e42e4c08f9 vs 68b329da9893e34099c7d8ad5cb9c940)

8 years agolib: istream-seekable - Fix reading when parent stream's content is larger than max_b...
Timo Sirainen [Wed, 14 Jun 2017 00:21:19 +0000 (03:21 +0300)] 
lib: istream-seekable - Fix reading when parent stream's content is larger than max_buffer_size

When writing the parent stream to fd and trying to read it back, it would
try to enforce the max_buffer_size and fail with:

istream-seekable: Couldn't read back in-memory input ...: buffer full

But since the data was already in buffer, istream-seekable shouldn't try to
enforce it at this point anymore.

8 years agopush-notification: Switch to main ioloop while calling drivers' commit callbacks
Timo Sirainen [Mon, 12 Jun 2017 20:45:05 +0000 (23:45 +0300)] 
push-notification: Switch to main ioloop while calling drivers' commit callbacks

There aren't any guarantees what ioloop happens to be active at the time
the commit is called. The active ioloop can also be destroyed early on,
which can cause an I/O leak and crashes later on.

8 years agomail-crypt-acl: Fix assert-crash due to missing namespace initialization
Timo Sirainen [Fri, 9 Jun 2017 15:10:14 +0000 (18:10 +0300)] 
mail-crypt-acl: Fix assert-crash due to missing namespace initialization

This was broken by e031d9aaae59a9f79710dc1138b76b69272615a3.

8 years agopop3: send back auth reply sooner
Josef 'Jeff' Sipek [Tue, 23 May 2017 11:12:49 +0000 (14:12 +0300)] 
pop3: send back auth reply sooner

If possible (i.e., pop3_lock_session=no), we can send back the OK
response earlier.

8 years agopop3: initialize namespaces explicitly
Josef 'Jeff' Sipek [Tue, 23 May 2017 10:01:48 +0000 (13:01 +0300)] 
pop3: initialize namespaces explicitly

(instead of relying on mail user initialization to instantiate them for
us)

8 years agopop3: client_create_from_input shouldn't process input
Josef 'Jeff' Sipek [Fri, 19 May 2017 07:58:18 +0000 (10:58 +0300)] 
pop3: client_create_from_input shouldn't process input

This makes the pop3 code resemble the imap code more.  Making the subsequent
changes easier.

8 years agopop3: client_create_from_input should return the client struct
Josef 'Jeff' Sipek [Fri, 19 May 2017 07:51:22 +0000 (10:51 +0300)] 
pop3: client_create_from_input should return the client struct

This makes the code more similar to imap's, making subsequent changes
easier.

8 years agopop3: rename login_client_connected() client arg to login_client
Josef 'Jeff' Sipek [Fri, 19 May 2017 07:46:05 +0000 (10:46 +0300)] 
pop3: rename login_client_connected() client arg to login_client

This makes the next commit cleaner.

8 years agopop3: move pop3 session locking out of client_create
Josef 'Jeff' Sipek [Fri, 19 May 2017 07:42:03 +0000 (10:42 +0300)] 
pop3: move pop3 session locking out of client_create

As a result, we can directly return the client structure (instead of
passing it back via a _r arg).

This makes the pop3 client_create look more like the imap version.

8 years agoauth: Don't lose all forward_ fields if the first passdb lookup fails.
Timo Sirainen [Tue, 30 May 2017 13:57:45 +0000 (16:57 +0300)] 
auth: Don't lose all forward_ fields if the first passdb lookup fails.

8 years agolib-http: client: Only drop queued requests when a DNS lookup fails; not also the...
Stephan Bosch [Wed, 24 May 2017 18:19:11 +0000 (20:19 +0200)] 
lib-http: client: Only drop queued requests when a DNS lookup fails; not also the ones that are already in progress.

8 years agolib-http: client: When a request is destroyed prematurely during payload input, consi...
Stephan Bosch [Wed, 24 May 2017 19:59:32 +0000 (21:59 +0200)] 
lib-http: client: When a request is destroyed prematurely during payload input, consider the payload stream destroyed and act accordingly.

The application may hold a reference to the payload stream still, and it may be difficult to prevent that.
This causes lib-http to keep waiting for the payload to be destroyed.
When nothing else is going on, the current ioloop may then become empty, which caused the familiar assert failure.

8 years agolib-http: client: Explicitly destroy the response payload timeout stream when the...
Stephan Bosch [Wed, 24 May 2017 17:08:43 +0000 (19:08 +0200)] 
lib-http: client: Explicitly destroy the response payload timeout stream when the request is destroyed while receiving payload.

This way, the timeout is stopped.
This caused ioloop panics.

8 years agolib-storage: Fix crash in mail_get_header_stream() when its previous stream wasn...
Timo Sirainen [Tue, 30 May 2017 15:25:50 +0000 (18:25 +0300)] 
lib-storage: Fix crash in mail_get_header_stream() when its previous stream wasn't at EOF

At least this could have happened when indexes were disabled and running:

FETCH 1 (envelope body.peek[header.fields (foo)] bodystructure)

Fixes:
Panic: file index-mail-headers.c: line 198 (index_mail_parse_header_init): assertion failed: (!mail->data.header_parser_initialized)

8 years agoquota: assertion for *args != NULL is added in order to avoid warning from static...
Sergey Kitov [Wed, 31 May 2017 11:32:30 +0000 (14:32 +0300)] 
quota: assertion for *args != NULL is added in order to avoid warning from static analyzer and *args[0] changed to (*args)[0]

8 years agoquota: Fix infinite loop when parsing quota parameters
Timo Sirainen [Wed, 24 May 2017 15:27:10 +0000 (18:27 +0300)] 
quota: Fix infinite loop when parsing quota parameters

Broken by previous quota changes.

8 years agoWhitespace cleanup
Sergey Kitov [Fri, 19 May 2017 08:00:22 +0000 (11:00 +0300)] 
Whitespace cleanup

8 years agoDict backend parameters reading changed to use common parsing function
Sergey Kitov [Fri, 19 May 2017 07:58:52 +0000 (10:58 +0300)] 
Dict backend parameters reading changed to use common parsing function

8 years agofs backend parameter parsing changed to use common function
Sergey Kitov [Fri, 19 May 2017 07:57:02 +0000 (10:57 +0300)] 
fs backend parameter parsing changed to use common function

8 years agoimapc backend parameter parsing changed to use common function
Sergey Kitov [Fri, 19 May 2017 07:52:46 +0000 (10:52 +0300)] 
imapc backend parameter parsing changed to use common function

8 years agomaildir backend parameter parsing changed to use common function
Sergey Kitov [Fri, 19 May 2017 07:49:01 +0000 (10:49 +0300)] 
maildir backend parameter parsing changed to use common function

8 years agoquota root default init changed to use common param parse function
Sergey Kitov [Fri, 19 May 2017 07:46:27 +0000 (10:46 +0300)] 
quota root default init changed to use common param parse function

8 years agocommon quota parameter parsing function implemented
Sergey Kitov [Thu, 18 May 2017 10:30:25 +0000 (13:30 +0300)] 
common quota parameter parsing function implemented

8 years agolib-storage: test-mail-search-args-simplify - test with initialized args
Timo Sirainen [Wed, 7 Jun 2017 16:54:10 +0000 (19:54 +0300)] 
lib-storage: test-mail-search-args-simplify - test with initialized args

8 years agolib-storage: mail_search_args_simplify() - Fix merging already-initialized keywords
Timo Sirainen [Wed, 7 Jun 2017 16:53:00 +0000 (19:53 +0300)] 
lib-storage: mail_search_args_simplify() - Fix merging already-initialized keywords

It was deinitializing the arg that was kept instead of the one being removed.

8 years agolib-storage: Fix mail_search_args_simplify() to not deinit too many args
Timo Sirainen [Wed, 7 Jun 2017 16:37:46 +0000 (19:37 +0300)] 
lib-storage: Fix mail_search_args_simplify() to not deinit too many args

Removing an arg should deinit it, but not its following siblings.

8 years agolib-storage: Add mail_search_arg_one_deinit()
Timo Sirainen [Wed, 7 Jun 2017 16:36:19 +0000 (19:36 +0300)] 
lib-storage: Add mail_search_arg_one_deinit()

8 years agoquota-clone: Fix compile warning
Timo Sirainen [Tue, 13 Jun 2017 23:15:34 +0000 (02:15 +0300)] 
quota-clone: Fix compile warning

8 years agowelcome: Fix sending parameters to welcome script.
Timo Sirainen [Mon, 12 Jun 2017 09:44:37 +0000 (12:44 +0300)] 
welcome: Fix sending parameters to welcome script.

Broken by f38b0dafbffa9d79542e36b4b3e598ed4115e5a3

8 years agopop3-migration: Strip trailing spaces from headers when calculating hash
Timo Sirainen [Mon, 12 Jun 2017 11:24:59 +0000 (14:24 +0300)] 
pop3-migration: Strip trailing spaces from headers when calculating hash

Fixes matching mails with Zimbra.

8 years agoimapc: If mailbox iteration fails, make sure error message isn't lost.
Timo Sirainen [Mon, 12 Jun 2017 11:35:22 +0000 (14:35 +0300)] 
imapc: If mailbox iteration fails, make sure error message isn't lost.

8 years agolib-storage: Fail mailbox list iteration early if it fails to get INBOX flags
Timo Sirainen [Mon, 12 Jun 2017 11:33:46 +0000 (14:33 +0300)] 
lib-storage: Fail mailbox list iteration early if it fails to get INBOX flags

The resulting INBOX reply could be wrong in that case.

Also with imapc if the remote server is down, this causes two connection
errors instead of just one.

8 years agolib-storage: If mailbox list iteration fails, don't add INBOX
Timo Sirainen [Mon, 12 Jun 2017 11:16:13 +0000 (14:16 +0300)] 
lib-storage: If mailbox list iteration fails, don't add INBOX

It possibly should have been returned by the list iteration itself and we
could be returning it wrong here.

Also calling this causes imapc to lost the error message in mailbox_list.

8 years agoimapc: Don't reconnect on DELETE if auth failure was already seen.
Timo Sirainen [Mon, 12 Jun 2017 09:18:28 +0000 (12:18 +0300)] 
imapc: Don't reconnect on DELETE if auth failure was already seen.

8 years agoimapc: Don't reconnect on STATUS if auth failure was already seen.
Timo Sirainen [Mon, 12 Jun 2017 09:16:19 +0000 (12:16 +0300)] 
imapc: Don't reconnect on STATUS if auth failure was already seen.

8 years agoimapc: Remove root_sep_lookup_failed cache
Timo Sirainen [Mon, 12 Jun 2017 08:57:52 +0000 (11:57 +0300)] 
imapc: Remove root_sep_lookup_failed cache

The auth_failed_* should already be enough for this. This removal also
fixes the storage error message to be auth_failed_reason when possible.

8 years agoimapc: Check for auth failures when saving messages
Timo Sirainen [Mon, 12 Jun 2017 09:07:52 +0000 (12:07 +0300)] 
imapc: Check for auth failures when saving messages

First, don't even attempt an APPEND if we've already seen an auth failure.
Second, if APPEND does fail because of auth error, set the correct error to
storage.

8 years agoimapc: Fix APPEND error message to include APPEND, not COPY
Timo Sirainen [Mon, 12 Jun 2017 09:03:39 +0000 (12:03 +0300)] 
imapc: Fix APPEND error message to include APPEND, not COPY

8 years agoimapc: Make sure storage error has the proper auth failure error string
Timo Sirainen [Mon, 12 Jun 2017 08:40:03 +0000 (11:40 +0300)] 
imapc: Make sure storage error has the proper auth failure error string

The first failed command always had the correct error string, but the
following failed commands just returned -1 without updating storage error.
The storage error could have been something completely different by then.

8 years agoimapc: Cleanup - Rename auth_error to auth_failed_reason for consistency
Timo Sirainen [Mon, 12 Jun 2017 08:33:38 +0000 (11:33 +0300)] 
imapc: Cleanup - Rename auth_error to auth_failed_reason for consistency

8 years agoimapc: Change auth_failed boolean to more specific auth_failed_state
Timo Sirainen [Mon, 12 Jun 2017 08:32:23 +0000 (11:32 +0300)] 
imapc: Change auth_failed boolean to more specific auth_failed_state

Since we now rely on auth_failed_state being initialized to zero,
explicitly set IMAPC_COMMAND_STATE_OK to be defined as 0.

8 years agoimapc: Remove auth_failed check from imapc_list_get_storage_name()
Timo Sirainen [Mon, 12 Jun 2017 08:05:31 +0000 (11:05 +0300)] 
imapc: Remove auth_failed check from imapc_list_get_storage_name()

Reverts 9949dc3649ac7a00289f6bf0662c793bee3d4286. It doesn't seem to be
actually needed at least now, and it's breaking:

 * SELECT INBOX -> failure due to remote server being down
 * SELECT INBOX -> NO [CANNOT] Invalid mailbox name: Name is empty

8 years agoimap-quota: If quota lookups fail, return NO for GETQUOTA/GETQUOTAROOT
Timo Sirainen [Thu, 8 Jun 2017 17:34:39 +0000 (20:34 +0300)] 
imap-quota: If quota lookups fail, return NO for GETQUOTA/GETQUOTAROOT

8 years agoquota: Update comments in quota.h
Timo Sirainen [Thu, 8 Jun 2017 17:34:27 +0000 (20:34 +0300)] 
quota: Update comments in quota.h

8 years agoquota: quota_root_iter_next() - Iterate all visible roots
Timo Sirainen [Thu, 8 Jun 2017 17:25:11 +0000 (20:25 +0300)] 
quota: quota_root_iter_next() - Iterate all visible roots

Don't try to skip the roots that don't have an active quota. imap_quota
plugin does the skipping now itself, and quota_clone shouldn't really be
even used with more than one quota root.

8 years agoimap-quota: If quota root doesn't have any quotas, don't send empty QUOTA reply
Timo Sirainen [Thu, 8 Jun 2017 17:24:26 +0000 (20:24 +0300)] 
imap-quota: If quota root doesn't have any quotas, don't send empty QUOTA reply

For example if GETQUOTA is asked for fs quota, but the user doesn't have fs
quota enabled.

8 years agoquota-clone: Don't clone quota resources that aren't enabled.
Timo Sirainen [Thu, 8 Jun 2017 17:29:05 +0000 (20:29 +0300)] 
quota-clone: Don't clone quota resources that aren't enabled.

They would always just be zeros anyway, so this makes the update slightly
more efficient. Although practically this only matters with dirsize and
fs quotas, which people generally don't use with quota_clone.

8 years agolib-imap-client: Fix hang when imapc_client_get_capabilities() is called without...
Timo Sirainen [Thu, 8 Jun 2017 17:39:47 +0000 (20:39 +0300)] 
lib-imap-client: Fix hang when imapc_client_get_capabilities() is called without connection

8 years agolib-mail: message_address_parse() - Fix fill_missing==TRUE handling
Timo Sirainen [Wed, 7 Jun 2017 15:10:10 +0000 (18:10 +0300)] 
lib-mail: message_address_parse() - Fix fill_missing==TRUE handling

Mainly MISSING_DOMAIN wasn't set in all situations. Also added unit tests.

8 years agolib-mail: message_address_parse() - Handle invalid standalone phrases better
Timo Sirainen [Wed, 7 Jun 2017 12:33:42 +0000 (15:33 +0300)] 
lib-mail: message_address_parse() - Handle invalid standalone phrases better

We'll treat "local-part" (without quotes) as a mailbox even without
@domain, but if it continues with anything or if it's a quoted-string,
we'll treat it as a display-name instead.

This is probably better than just converting everything to display-name,
since there are likely to exist headers like "To: localuser"

8 years agolib-mail: Improve test-message-address to test invalid addresses in lists
Timo Sirainen [Wed, 7 Jun 2017 12:22:22 +0000 (15:22 +0300)] 
lib-mail: Improve test-message-address to test invalid addresses in lists

Test in both a group and list of addresses.

8 years agolib-mail: message_address_parse() - don't stop at <> when parsing address lists
Timo Sirainen [Wed, 7 Jun 2017 12:21:10 +0000 (15:21 +0300)] 
lib-mail: message_address_parse() - don't stop at <> when parsing address lists

And similarly don't stop at <@domain>

8 years agolib-mail: message_address_parse() - don't stop at invalid addresses when parsing...
Timo Sirainen [Wed, 7 Jun 2017 12:08:27 +0000 (15:08 +0300)] 
lib-mail: message_address_parse() - don't stop at invalid addresses when parsing group

Also the end-group's message_address shouldn't have invalid_syntax=TRUE,
except when we didn't actually find the ending ';'

8 years agolib-mail: message_address_parse() - Fix reading out-of-bounds read on invalid input
Timo Sirainen [Wed, 7 Jun 2017 11:57:12 +0000 (14:57 +0300)] 
lib-mail: message_address_parse() - Fix reading out-of-bounds read on invalid input

8 years agolib-mail: message_address_parse() - Remove unnecessary NULL checks
Timo Sirainen [Wed, 7 Jun 2017 10:34:39 +0000 (13:34 +0300)] 
lib-mail: message_address_parse() - Remove unnecessary NULL checks

message_address_parse() always initialized rfc822_parser_init() with
last_comment string.

8 years agolib-mail: test-message-address - add more and better organized tests
Timo Sirainen [Wed, 7 Jun 2017 10:19:40 +0000 (13:19 +0300)] 
lib-mail: test-message-address - add more and better organized tests

8 years agolib-mail: message_address_parse() - Add INVALID_ROUTE/SYNTAX_ERROR only if fill_missi...
Timo Sirainen [Wed, 7 Jun 2017 09:58:21 +0000 (12:58 +0300)] 
lib-mail: message_address_parse() - Add INVALID_ROUTE/SYNTAX_ERROR only if fill_missing==TRUE

This allows message_address_write() to write the parsed invalid address in a
way that at least somewhat resembles the original input.

8 years agolib-mail: message_address_write() - Write empy username as ""
Timo Sirainen [Wed, 7 Jun 2017 09:42:28 +0000 (12:42 +0300)] 
lib-mail: message_address_write() - Write empy username as ""

We can't output <@domain> because that conflicts with the <@route:...>
syntax. Also ""@domain seems to be a valid address.

8 years agolib-mail: message_address_write() - Always write <user@domain> as output
Timo Sirainen [Wed, 7 Jun 2017 09:37:09 +0000 (12:37 +0300)] 
lib-mail: message_address_write() - Always write <user@domain> as output

This is the standard format nowadays, so don't convert <user@domain> into
plain user@domain.

8 years agolib-mail: message_address_write() - Don't write empty <>
Timo Sirainen [Wed, 7 Jun 2017 09:35:35 +0000 (12:35 +0300)] 
lib-mail: message_address_write() - Don't write empty <>

8 years agolib-mail: message_address_write() - Don't add '@' if domain is missing
Timo Sirainen [Wed, 7 Jun 2017 09:34:17 +0000 (12:34 +0300)] 
lib-mail: message_address_write() - Don't add '@' if domain is missing

8 years agolib-mail: test-message-address - Test also writing invalid addresses
Timo Sirainen [Wed, 7 Jun 2017 09:33:10 +0000 (12:33 +0300)] 
lib-mail: test-message-address - Test also writing invalid addresses

8 years agolib-mail: test-message-address cleanup
Timo Sirainen [Wed, 7 Jun 2017 09:01:07 +0000 (12:01 +0300)] 
lib-mail: test-message-address cleanup

Merge input and output into the same struct to make it clearer.

8 years ago-Wstrict-bool warning fixes
Timo Sirainen [Tue, 7 Jun 2016 00:41:18 +0000 (03:41 +0300)] 
-Wstrict-bool warning fixes

8 years agolib-mail: Fixed to 91fdb25b5
Timo Sirainen [Mon, 6 Jun 2016 14:10:12 +0000 (17:10 +0300)] 
lib-mail: Fixed to 91fdb25b5

Caused by my manual edits to avoid the -Wstrict-bool warnings.

8 years agolib-mail: Update tests for message address
Pali Rohár [Sun, 5 Jun 2016 13:48:20 +0000 (15:48 +0200)] 
lib-mail: Update tests for message address

8 years agolib-mail: message_address_write: Quote and escape strings if needed
Pali Rohár [Sun, 5 Jun 2016 13:48:19 +0000 (15:48 +0200)] 
lib-mail: message_address_write: Quote and escape strings if needed

ATEXT characters must be properly quoted when are in phrase.

Test case:

  { name = "test\"test", mailbox = "user", domain = "host" }

converts to:

  "test\"test" <user@host>

8 years agolib-mail: parse_mailbox: Set display name instead mailbox when parsing failed
Pali Rohár [Sun, 5 Jun 2016 13:48:18 +0000 (15:48 +0200)] 
lib-mail: parse_mailbox: Set display name instead mailbox when parsing failed

It does not make sense to set mailbox without domain on incorrect input.
Rather set display name which is more likely useable value.

Test case:

  test

is parsed as:

  { name = "test", mailbox = NULL, domain = NULL }

8 years agolib-mail: parse_addr_spec: Email address without local-part is invalid
Pali Rohár [Sun, 5 Jun 2016 13:48:17 +0000 (15:48 +0200)] 
lib-mail: parse_addr_spec: Email address without local-part is invalid

Add explicit invalid_syntax flag also when end of input occure because
address is without domain invalid and in this case it was not correctly
propagated.

8 years agolib-mail: parse_addr_spec: Like in rfc822_skip_comment() check if last_comment is...
Pali Rohár [Sun, 5 Jun 2016 13:48:16 +0000 (15:48 +0200)] 
lib-mail: parse_addr_spec: Like in rfc822_skip_comment() check if last_comment is not NULL

This will fix possible NULL pointer dereference when caller does not set last_comment.

8 years agolib-mail: message_address_write: Fix generating group list with empty name
Pali Rohár [Sun, 5 Jun 2016 13:48:15 +0000 (15:48 +0200)] 
lib-mail: message_address_write: Fix generating group list with empty name

Empty name for group list must be quoted.

Test case:

  { { name = NULL, mailbox = "", domain = NULL }, { name = NULL, mailbox = NULL, domain = NULL } }

converts to:

  "":;

8 years agolib-mail: message_address_write: Fix generating empty group list
Pali Rohár [Sun, 5 Jun 2016 13:48:14 +0000 (15:48 +0200)] 
lib-mail: message_address_write: Fix generating empty group list

Empty group list ends with ": " not with ", ".

Test case:

  { { name = NULL, mailbox = "group", domain = NULL }, { name = NULL, mailbox = NULL, domain = NULL } }

converts to:

  group:;

8 years agolib: ioloop-kqueue - Improve kevent() panic log message
Timo Sirainen [Fri, 9 Jun 2017 20:32:28 +0000 (23:32 +0300)] 
lib: ioloop-kqueue - Improve kevent() panic log message

8 years agolib-storage: Fix sort index writing to not leave sort_id=0 gaps
Timo Sirainen [Fri, 9 Jun 2017 09:42:51 +0000 (12:42 +0300)] 
lib-storage: Fix sort index writing to not leave sort_id=0 gaps

This fixes errors like:
Error: INBOX: Broken sort-f indexes, resetting

8 years agolib-storage: Add details to "Broken sort-* indexes" error
Timo Sirainen [Fri, 9 Jun 2017 09:25:36 +0000 (12:25 +0300)] 
lib-storage: Add details to "Broken sort-* indexes" error

8 years agolib-storage: Fix setting the correct cache record corrupted
Timo Sirainen [Fri, 9 Jun 2017 11:31:15 +0000 (14:31 +0300)] 
lib-storage: Fix setting the correct cache record corrupted

It was mixing UIDs and sequences, so a wrong mail could have been set
corrupted or it could have crashed with:

Panic: file mail-index-transaction-update.c: line 1018 (mail_index_update_ext): assertion failed: (seq > 0 && (seq <= mail_index_view_get_messages_count(t->view) || seq <= t->last_new_seq))

8 years agoquota: Fix negative quota_warnings with count backend
Timo Sirainen [Fri, 9 Jun 2017 09:07:00 +0000 (12:07 +0300)] 
quota: Fix negative quota_warnings with count backend

For example this never triggered:

  quota_warning = -messages=100%% quota-warning %u -100

The change to quota_alloc() to update the count_used doesn't seem to
actually fix anything right now, but it makes the code more correct.

8 years agolib: Use unix streams with connection when using unix socket
Aki Tuomi [Wed, 31 May 2017 11:39:55 +0000 (14:39 +0300)] 
lib: Use unix streams with connection when using unix socket

This makes it easier to send file descriptors over connection

8 years agoquota: Fix imapc backend not to try to find out mail's size.
Timo Sirainen [Tue, 30 May 2017 18:57:06 +0000 (21:57 +0300)] 
quota: Fix imapc backend not to try to find out mail's size.

This removes unnecessary FETCH RFC822.SIZE command sent to server when
expunging mails.

8 years agolib-imap-client: Fix test-imapc-client unit test to work on OSX
Timo Sirainen [Tue, 30 May 2017 07:37:24 +0000 (10:37 +0300)] 
lib-imap-client: Fix test-imapc-client unit test to work on OSX

connect() to port 0 fails in it, so for the "connect failure" test first
open a random free listener port and then close it. Hopefully nothing else
reopens it in the mean time.

8 years agoimapc: Don't send UID FETCH 1:* after SELECT if mailbox is empty
Timo Sirainen [Tue, 30 May 2017 21:07:46 +0000 (00:07 +0300)] 
imapc: Don't send UID FETCH 1:* after SELECT if mailbox is empty

8 years agoimapc: SEARCH - Don't add parenthesis around OR
Timo Sirainen [Tue, 30 May 2017 21:00:32 +0000 (00:00 +0300)] 
imapc: SEARCH - Don't add parenthesis around OR

There's no need to write "(OR a b)" when "OR a b" works just the same.

8 years agolib-storage: mail_search_args_to_imap() - don't add parenthesis around OR
Timo Sirainen [Tue, 30 May 2017 20:53:44 +0000 (23:53 +0300)] 
lib-storage: mail_search_args_to_imap() - don't add parenthesis around OR

There's no need to write "(OR a b)" when "OR a b" works just the same.

8 years agolib-mail: Fix compiler warnings
Timo Sirainen [Fri, 9 Jun 2017 10:17:38 +0000 (13:17 +0300)] 
lib-mail: Fix compiler warnings

8 years agoglobal: Use i_realloc_type() wherever possible
Timo Sirainen [Wed, 7 Jun 2017 08:35:18 +0000 (11:35 +0300)] 
global: Use i_realloc_type() wherever possible

8 years agolib: Add test-mempool unit test.
Timo Sirainen [Wed, 7 Jun 2017 08:24:38 +0000 (11:24 +0300)] 
lib: Add test-mempool unit test.

It verifies that the p_new() and p_realloc_type() handle overflows
correctly. test-malloc-overflow already does this for the low-level
MALLOC_*() macros, but here we're cheking that the actually used
memory allocation macros also work.

8 years agolib: Add i_realloc_type() for i_realloc() that checks for overflows
Timo Sirainen [Wed, 7 Jun 2017 08:24:03 +0000 (11:24 +0300)] 
lib: Add i_realloc_type() for i_realloc() that checks for overflows

8 years agolmtp: Removed (Dovecot) from Received header.
Timo Sirainen [Thu, 12 May 2016 09:40:29 +0000 (05:40 -0400)] 
lmtp: Removed (Dovecot) from Received header.

Some people want to hide it, and I don't really see much benefit in
including it anyway. So lets just hide it from everyone.

8 years agoimapc: Fix potential crash if initial sync fails.
Timo Sirainen [Thu, 8 Jun 2017 08:35:58 +0000 (11:35 +0300)] 
imapc: Fix potential crash if initial sync fails.

Happened at least if mailbox had 0 mails and a failing mailbox_sync() was
called twice.

Fixes:
Panic: file imapc-sync.c: line 477 (imapc_sync_index): assertion failed: (mbox->sync_fetch_first_uid == 1)

8 years agolib: test for guid_128_generate() time handling
Josef 'Jeff' Sipek [Mon, 5 Jun 2017 11:21:23 +0000 (14:21 +0300)] 
lib: test for guid_128_generate() time handling

Make sure that guids always increase regardless of what is happening to
the time.

8 years agolib: guid_128_generate() shouldn't generate invalid timestamps
Josef 'Jeff' Sipek [Mon, 5 Jun 2017 09:55:56 +0000 (12:55 +0300)] 
lib: guid_128_generate() shouldn't generate invalid timestamps

The nsec validity check must take into account the soon-to-be-done increment.

8 years agolib: guid_128_generate() needs to correctly convert usecs to nsecs
Josef 'Jeff' Sipek [Mon, 5 Jun 2017 09:53:23 +0000 (12:53 +0300)] 
lib: guid_128_generate() needs to correctly convert usecs to nsecs

The timestamp in the guid starts off with real time, but "runs ahead" if a lot
of guids are generated.  If not many guids are being generated, the timestamp
is fast-forwarded to the current ioloop_timeval.  The buggy comparison prevented
fast forwarding when tv_sec matched, but tv_nsec value was > 0 (which
was true most of the time).

8 years agolib-index: Fix reset_id handling in mail_cache_get_missing_reason() cache
Timo Sirainen [Tue, 6 Jun 2017 14:18:27 +0000 (17:18 +0300)] 
lib-index: Fix reset_id handling in mail_cache_get_missing_reason() cache

Broken by bd897f82a5d9b00ba256b462d2056c2dc7df257c

8 years agolib-storage: Fix assert-crash in SORT caused by earlier changes
Timo Sirainen [Tue, 6 Jun 2017 06:39:34 +0000 (09:39 +0300)] 
lib-storage: Fix assert-crash in SORT caused by earlier changes

Broken by 0a6a527f0c42b5478d80ac53ab357885676fd516

Fixes:
Panic: file index-sort-string.c: line 668 (index_sort_add_ids_range): assertion failed: (str != NULL)

8 years agolib-index: Add cache to mail_cache_get_missing_reason()
Timo Sirainen [Mon, 5 Jun 2017 18:53:18 +0000 (21:53 +0300)] 
lib-index: Add cache to mail_cache_get_missing_reason()

This avoids excessive CPU usage when it's called in a loop for many mails.

8 years agolib-index: mail_cache_get_missing_reason() - Minor code cleanup
Timo Sirainen [Mon, 5 Jun 2017 18:44:45 +0000 (21:44 +0300)] 
lib-index: mail_cache_get_missing_reason() - Minor code cleanup