Re: [RFA/commit] Handle EOF on terminals opened with ENONBLOCK...

[Joel Brobecker <brobecker@adacore.com>]

[-- Attachment #1: Type: text/plain, Size: 358 bytes --]

> 2009-04-23  Joel Brobecker   <brobecker@adacore.com>
> 
>         * utils.c: Add include of gdb_usleep.h.
>         (defaulted_query): Detect false EOF conditions that happen
>         on terminals opened with the O_NONBLOCK flag when there is
>         nothing to read.

This is what I ended up checking in. Eli, is the comment clearer
for you?

-- 
Joel

[-- Attachment #2: utils.c.diff --]
[-- Type: text/x-diff, Size: 1418 bytes --]

Index: utils.c
===================================================================
RCS file: /cvs/src/src/gdb/utils.c,v
retrieving revision 1.210
diff -u -p -r1.210 utils.c
--- utils.c	24 Apr 2009 22:10:03 -0000	1.210
+++ utils.c	6 May 2009 22:51:37 -0000
@@ -67,6 +67,8 @@
 #include <sys/time.h>
 #include <time.h>
 
+#include "gdb_usleep.h"
+
 #if !HAVE_DECL_MALLOC
 extern PTR malloc ();		/* ARI: PTR */
 #endif
@@ -1477,6 +1479,25 @@ defaulted_query (const char *ctlstr, con
       gdb_flush (gdb_stdout);
 
       answer = fgetc (stdin);
+
+      /* We expect fgetc to block until a character is read.  But
+         this may not be the case if the terminal was opened with
+         the NONBLOCK flag.  In that case, if there is nothing to
+         read on stdin, fgetc returns EOF, but also sets the error
+         condition flag on stdin and errno to EAGAIN.  With a true
+         EOF, stdin's error condition flag is not set.
+
+         A situation where this behavior was observed is a pseudo
+         terminal on AIX.  */
+      while (answer == EOF && ferror (stdin) && errno == EAGAIN)
+        {
+          /* Not a real EOF.  Wait a little while and try again until
+             we read something.  */
+          clearerr (stdin);
+          gdb_usleep (10000);
+          answer = fgetc (stdin);
+        }
+
       clearerr (stdin);		/* in case of C-d */
       if (answer == EOF)	/* C-d */
 	{