* gpgme.texi (Run Control): Update this section.
(Waiting For Completion): Likewise for this subsection.
(Cancelling an Operation): Likewise for this subsection.
(Using External Event Loops): New subsection with several
subsubsections.
* gpgme.c (gpgme_set_io_cbs): Deal with CTX being NULL.
* gpgme.c (_gpgme_op_event_cb_user): New function.
* op-support.c (_gpgme_op_reset): Support a new mode of operation
for private or user event loop. Use new user event callback
wrapper.
* trustlist.c (gpgme_op_trustlist_start): Use this new mode.
* keylist.c (gpgme_op_keylist_start): Likewise.
* rungpg.c (_gpgme_gpg_io_event): New function.
* rungpg.h (_gpgme_gpg_io_event): New prototype.
* engine-gpgsm.c (_gpgme_gpg_io_event): New function.
* engine-gpgsm.h (_gpgme_gpgsm_io_event): New prototype.
* engine.c (_gpgme_engine_io_event): New function.
* engine.h (_gpgme_engine_io_event): New prototype.
* keylist.c (finish_key): Call _gpgme_engine_io_event, and move
the real work for the default IO callback routines to ...
(_gpgme_op_keylist_event_cb): ... here. New function.
* trustlist.c (trustlist_colon_handler): Signal
GPGME_EVENT_NEXT_TRUSTITEM. Move queue manipulation to ...
(_gpgme_op_trustlist_event_cb): ... here. New function.
* gpgme.c (_gpgme_op_event_cb): Call _gpgme_op_keylist_event_cb
and _gpgme_op_trustlist_event_cb when appropriate.
* ops.h (_gpgme_op_keylist_event_cb): New prototype.
(_gpgme_op_trustlist_event_cb): Likewise.
* op-support.c (_gpgme_op_reset): Add comment why we don't use the
user provided event handler directly.
* gpgme.h (GpgmeRegisterIOCb): Return GpgmeError value, and TAG in
a pointer argument.
* wait.c (_gpgme_add_io_cb): Likewise.
* wait.h (_gpgme_add_io_cb): Likewise for prototype.
* rungpg.c (_gpgme_gpg_add_io_cb): Call IO_CBS->add with new
argument. Fix up error handling.
* engine-gpgsm.c (_gpgme_gpgsm_add_io_cb): Call IO_CBS->add with
new argument, fix up error handling.
when to change it.
* gpgme.spec.in: New. Contributed by Wojciech Polak.
* Makefile.am (dist-hook): New.
* AUTHORS: Added Wojciech and bug reporting addresses.
have been removed after my last fix. This avoids a memory leak
when a fingerprint was not found. Removed the double loop
increment in the code to release the arrays.
(make_fingerprint): Removed superfluous check on retrun value of
xmalloc.
(safe_free): Removed. Changed all callers to use a regular free
and at appropriate palces set the free pointer to NULL. That
safe_free stuff seems to have been copied verbatim from some
Mutt example code I posted.
(storeNewCharPtr): Use xmalloc instead of an unchecked
malloc. Removed superfluous string termination.
(parseAddress): Use xmalloc instead of an unchecked malloc.
(nextAddress): Ditto.
* gpgmeplug.c: Moved a few helper functions more to the top.
Fixed comment syntax. Merged a copyright notice somewhere in the
middle of the file with the one at the top.
have been removed after my last fix. This avoids a memory leak
when a fingerprint was not found. Removed the double loop
increment in the code to release the arrays.
(safe_malloc): Removed this macro and replaced it at all places
without return values checks by xmalloc.
(xstrdup): New. Replaces funny named macro with different
semantics. Changed all callers to the new semantic.
(findCertificates): Don't free duplicate free the FPR array
values. Removed the unneeded initialization. Replaces the
gcc-ish use dynamic allocation of automatic variables by repalce
maxCerts with a macro MAXCERTS. Made some comments Real C (tm).
(startListCertificates): Removed uneeded cast from xmalloc.
this also fixes the dividing factor.
(signatureCertificateDaysLeftToExpiry)
(preceiverCertificateDaysLeftToExpiry): using the new macro
(caCertificateDaysLeftToExpiry)
(rootCertificateDaysLeftToExpiry): using new macro in deactivated code.
* ops.h (_gpgme_wait_on_condition): Remove HANG argument from
prototype and change return type to GpgmeError.
(_gpgme_wait_one): New prototype.
* wait.c (gpgme_wait): Replace with the meat from
_gpgme_wait_on_condition here, and remove the support for
conditions.
(_gpgme_wait_on_condition): Remove HANG argument from prototype
and change return type to GpgmeError. Replace with meat from
_gpgme_wait_one and add support for conditions.
(_gpgme_wait_one): Just call _gpgme_wait_on_condition without
condition.
* keylist.c (gpgme_op_keylist_ext_start): Always use our own FD
table (eg use synchronous mode).
(gpgme_op_keylist_next): Remove HANG argument from
_gpgme_wait_on_condition. Check its return value.
* trustlist.c (gpgme_op_trustlist_start): Always use our own FD
table (eg use synchronous mode).
(gpgme_op_trustlist_next): Remove HANG argument from
_gpgme_wait_on_condition. Check its return value.
(main): Print operation info.
* engine-gpgsm.c (map_assuan_error): Map No_Data_Available to EOF.
* import.c (append_xml_impinfo): Kludge to print fingerprint
instead of keyid for use with gpgsm.
(import_status_handler): Set a flag to know whether any import
occured.
(gpgme_op_import): Reurn -1 if no certificate ewas imported.
* gpgme.texi (Importing Keys): Document the return value -1 of
gpgme_op_import.
* engine-gpgsm.c (_gpgme_gpgsm_op_export): Only export the keys
listed in RECP.
* export.c (gpgme_op_export): If no data was returned, return
GPGME_No_Recipients.
* engine-gpgsm.c (gpgsm_assuan_simple_command): Return ERR.
(parse_status): New function.
(gpgsm_status_handler): Use parse_status.
(gpgsm_assuan_simple_command): Accept new arguments STATUS_FNC and
STATUS_FNC_VALUE and process status messages.
(gpgsm_set_recipients): Pass new arugments to gpgsm_assuan_simple_command.
(gpgsm_set_fd): Likewise.
(_gpgme_gpgsm_op_keylist): Likewise.
(_gpgme_gpgsm_op_keylist_ext): Likewise.
(_gpgme_gpgsm_op_sign): Likewise.
BTW, why did you change the sequence of the parts. According to the
specs there is no standard but our implementation has to make sure
that it uses the same sequence everywhere. Adding the extra parts is
useless because they are not defined in rfc2253 and thus can't occur.
