aboutsummaryrefslogtreecommitdiffstats
path: root/NEWS
diff options
context:
space:
mode:
authorMarcus Brinkmann <[email protected]>2003-05-27 01:31:06 +0000
committerMarcus Brinkmann <[email protected]>2003-05-27 01:31:06 +0000
commit03bcb7f4c14e771392ccf04d1d5cd38c5baf47b2 (patch)
tree03f00ce78b2feec8187145b102edb6cdd8c30c7f /NEWS
parentUpdate Miguel's email address. (diff)
downloadgpgme-03bcb7f4c14e771392ccf04d1d5cd38c5baf47b2.tar.gz
gpgme-03bcb7f4c14e771392ccf04d1d5cd38c5baf47b2.zip
doc/
2003-05-27 Marcus Brinkmann <[email protected]> * gpgme.texi (Passphrase Callback): Document new prototype. gpgme/ 2003-05-26 Marcus Brinkmann <[email protected]> * engine.h (EngineCommandHandler): Change last argument to int fd. * gpgme.h (gpgme_passphrase_cb_t): Rewritten to take parts of the description and fd. (gpgme_edit_cb_t): Change last argument to int fd. * ops.h (_gpgme_passphrase_command_handler_internal): New prototype. * passphrase.c: Include <assert.h>. (op_data_t): Rename userid_hint to uid_hint, remove last_pw_handle. (release_op_data): Check values before calling free. (_gpgme_passphrase_status_handler): Likewise. (_gpgme_passphrase_command_handler_internal): New function. (_gpgme_passphrase_command_handler): Rewritten. * edit.c (edit_status_handler): Pass -1 as fd argument. (command_handler): Update prototype. New variable processed. Use it to store return value of _gpgme_passphrase_command_handler_internal which is now used instead _gpgme_passphrase_command_handler. Use it also to check if we should call the user's edit function. Pass fd to user's edit function. * rungpg.c (struct gpg_object_s): Change type of cmd.cb_data to void *. (gpg_release): Check value before calling free. Do not release cmd.cb_data. (command_cb): Function removed. (command_handler): New function. Thus we don't use a data object for command handler stuff anymore, but handle it directly. This allows proper error reporting (cancel of passphrase requests, for example). Also all callbacks work via direct writes to the file descriptor (so that passphrases are not kept in insecure memory). (gpg_set_command_handler): Rewritten to use even more ugly hacks. (read_status): Check cmd.keyword before calling free. Install command_handler as the I/O callback handler with GPG as private data. tests/ 2003-05-27 Marcus Brinkmann <[email protected]> * (t-decrypt-verify.c, t-decrypt.c, t-edit.c, t-encrypt-sign.c, t-encrypt-sym.c, t-sign.c, t-signers.c): Include <unistd.h>. (passphrase_cb): Rewritten. * t-edit.c (edit_fnc): Rewritten.
Diffstat (limited to 'NEWS')
-rw-r--r--NEWS14
1 files changed, 14 insertions, 0 deletions
diff --git a/NEWS b/NEWS
index a6523add..00f3e9e5 100644
--- a/NEWS
+++ b/NEWS
@@ -46,6 +46,18 @@ Noteworthy changes in version 0.4.1 (unreleased)
signatures, the progress callback is invoked for both the detached
signature and the plaintext message, though.
+ * gpgme_passphrase_cb_t has been changed to not return a complete
+ description, but the UID hint, passphrase info and a flag
+ indicating if this is a repeated attempt individually, so the user
+ can compose his own description from this information. Furthermore
+ it does not expect the user to return a string, but write the
+ passphrase followed by a newline character directly to a file
+ descriptor.
+
+ * gpgme_edit_cb_t has been changed to take a file descriptor argument.
+ The user is expected to write the response to the file descriptor,
+ followed by a newline.
+
* gpgme_op_verify and gpgme_op_decrypt_verify don't return a status
summary anymore. Use gpgme_get_sig_status to retrieve the individual
stati.
@@ -226,6 +238,8 @@ gpgme_trust_item_t NEW
gpgme_status_code_t NEW
gpgme_io_cb_t CHANGED: Return type from void to GpgmeError.
gpgme_event_io_t CHANGED: New event type (all numbers changed).
+gpgme_passphrase_cb_t CHANGED: Desc decomposed, write directly to FD.
+gpgme_edit_cb_t CHANGED: Write directly to FD.
gpgme_key_get_string_attr CHANGED: Don't handle GPGME_ATTR_IS_SECRET.
gpgme_op_verify CHANGED: Drop R_STAT argument.
gpgme_op_decrypt_verify CHANGED: Drop R_STAT argument.