Loop on our reads while errno is EINTR or EAGAIN. Also make sure to

2002-06-27  Jeffrey Stedfast  <fejj@ximian.com>

	* camel-gpg-context.c (gpg_ctx_op_step): Loop on our reads while
	errno is EINTR or EAGAIN. Also make sure to completely read stdout
	and stderr by keeping better state.
	(gpg_ctx_parse_status): In the case of a NODATA response from gpg,
	try to use the diagnostics that gpg may have written to its
	stderr.
	(gpg_verify): Check that the gpg process has not exited.
	(gpg_ctx_op_wait): Make sure we haven't already exited (as by
	gpg_ctx_op_is_exited()) and if we have, retrieve the exit status
	from the gpg context.

diff --git a/camel/ChangeLog b/camel/ChangeLog
index 468596b..07095f7 100644 (file)
--- a/camel/ChangeLog
+++ b/camel/ChangeLog
@@ -6,6 +6,10 @@
        (gpg_ctx_parse_status): In the case of a NODATA response from gpg,
        try to use the diagnostics that gpg may have written to its
        stderr.
+       (gpg_verify): Check that the gpg process has not exited.
+       (gpg_ctx_op_wait): Make sure we haven't already exited (as by
+       gpg_ctx_op_is_exited()) and if we have, retrieve the exit status
+       from the gpg context.
 
 2002-06-26  Jeffrey Stedfast  <fejj@ximian.com>
 

diff --git a/camel/camel-gpg-context.c b/camel/camel-gpg-context.c
index 7d5f167..c847b9c 100644 (file)
--- a/camel/camel-gpg-context.c
+++ b/camel/camel-gpg-context.c
@@ -249,6 +249,9 @@ struct _GpgCtx {
        
        GByteArray *diagnostics;
        
+       int exit_status;
+       
+       unsigned int exited:1;
        unsigned int complete:1;
        unsigned int seen_eof1:1;
        unsigned int seen_eof2:1;
@@ -262,7 +265,7 @@ struct _GpgCtx {
        unsigned int validsig:1;
        unsigned int trust:3;
        
-       unsigned int padding:19;
+       unsigned int padding:18;
 };
 
 static struct _GpgCtx *
@@ -279,6 +282,8 @@ gpg_ctx_new (CamelSession *session, const char *path)
        gpg->seen_eof1 = FALSE;
        gpg->seen_eof2 = FALSE;
        gpg->pid = (pid_t) -1;
+       gpg->exit_status = 0;
+       gpg->exited = FALSE;
        
        gpg->path = g_strdup (path);
        gpg->userid = NULL;
@@ -1100,16 +1105,38 @@ gpg_ctx_op_complete (struct _GpgCtx *gpg)
        return gpg->complete && gpg->seen_eof1 && gpg->seen_eof2;
 }
 
+static gboolean
+gpg_ctx_op_is_exited (struct _GpgCtx *gpg)
+{
+       pid_t retval;
+       int status;
+       
+       if (gpg->exited)
+               return TRUE;
+       
+       retval = waitpid (gpg->pid, &status, WNOHANG);
+       if (retval == gpg->pid) {
+               gpg->exit_status = status;
+               gpg->exited = TRUE;
+               return TRUE;
+       }
+       
+       return FALSE;
+}
+
 static void
 gpg_ctx_op_cancel (struct _GpgCtx *gpg)
 {
        pid_t retval;
        int status;
        
+       if (gpg->exited)
+               return;
+       
        kill (gpg->pid, SIGTERM);
        sleep (1);
        retval = waitpid (gpg->pid, &status, WNOHANG);
-       if (retval == 0) {
+       if (retval == (pid_t) 0) {
                /* no more mr nice guy... */
                kill (gpg->pid, SIGKILL);
                sleep (1);
@@ -1124,25 +1151,30 @@ gpg_ctx_op_wait (struct _GpgCtx *gpg)
        pid_t retval;
        int status;
        
-       sigemptyset (&mask);
-       sigaddset (&mask, SIGALRM);
-       sigprocmask (SIG_BLOCK, &mask, &omask);
-       alarm (1);
-       retval = waitpid (gpg->pid, &status, 0);
-       alarm (0);
-       sigprocmask (SIG_SETMASK, &omask, NULL);
-       
-       if (retval == (pid_t) -1 && errno == EINTR) {
-               /* The child is hanging: send a friendly reminder. */
-               kill (gpg->pid, SIGTERM);
-               sleep (1);
-               retval = waitpid (gpg->pid, &status, WNOHANG);
-               if (retval == (pid_t) 0) {
-                       /* Still hanging; use brute force. */
-                       kill (gpg->pid, SIGKILL);
+       if (!gpg->exited) {
+               sigemptyset (&mask);
+               sigaddset (&mask, SIGALRM);
+               sigprocmask (SIG_BLOCK, &mask, &omask);
+               alarm (1);
+               retval = waitpid (gpg->pid, &status, 0);
+               alarm (0);
+               sigprocmask (SIG_SETMASK, &omask, NULL);
+               
+               if (retval == (pid_t) -1 && errno == EINTR) {
+                       /* The child is hanging: send a friendly reminder. */
+                       kill (gpg->pid, SIGTERM);
                        sleep (1);
                        retval = waitpid (gpg->pid, &status, WNOHANG);
+                       if (retval == (pid_t) 0) {
+                               /* Still hanging; use brute force. */
+                               kill (gpg->pid, SIGKILL);
+                               sleep (1);
+                               retval = waitpid (gpg->pid, &status, WNOHANG);
+                       }
                }
+       } else {
+               status = gpg->exit_status;
+               retval = gpg->pid;
        }
        
        if (retval != (pid_t) -1 && WIFEXITED (status))
@@ -1282,7 +1314,7 @@ gpg_verify (CamelCipherContext *context, CamelCipherHash hash,
                goto exception;
        }
        
-       while (!gpg_ctx_op_complete (gpg)) {
+       while (!gpg_ctx_op_complete (gpg) && !gpg_ctx_op_is_exited (gpg)) {
                if (camel_operation_cancel_check (NULL)) {
                        camel_exception_set (ex, CAMEL_EXCEPTION_USER_CANCEL,
                                             _("Cancelled."));