Date: Fri, 23 Apr 2004 19:49:25 +0900 (JST) From: NIIMI Satoshi <sa2c@sa2c.net> To: FreeBSD-gnats-submit@FreeBSD.org Cc: nobutaka@FreeBSD.org Subject: ports/65910: editor/emacs ignores C-g when creating network connection Message-ID: <20040423104925.973365C86@berkeley.l.sa2c.net> Resent-Message-ID: <200404231050.i3NAoODg069838@freefall.freebsd.org>
next in thread | raw e-mail | index | archive | help
>Number: 65910
>Category: ports
>Synopsis: editor/emacs ignores C-g when creating network connection
>Confidential: no
>Severity: serious
>Priority: medium
>Responsible: freebsd-ports-bugs
>State: open
>Quarter:
>Keywords:
>Date-Required:
>Class: sw-bug
>Submitter-Id: current-users
>Arrival-Date: Fri Apr 23 03:50:24 PDT 2004
>Closed-Date:
>Last-Modified:
>Originator: NIIMI Satoshi
>Release: FreeBSD 5.2.1-RELEASE-p5 i386
>Organization:
>Environment:
System: FreeBSD berkeley.l.sa2c.net 5.2.1-RELEASE-p5 FreeBSD 5.2.1-RELEASE-p5 #3: Fri Apr 16 08:33:34 JST 2004 root@berkeley.l.sa2c.net:/usr/obj/usr/src/sys/GENERIC i386
>Description:
Emacs checks interrupt key input (C-g) periodically when it blocks long time.
But a patch file for src:process.c in ports tree does not.
This makes open-network-stream() non-interruptable.
>How-To-Repeat:
Emacs on X on FreeBSD 5:
(with-temp-buffer
;; "host" is a resolbable but unreachable host.
(open-network-stream "hoge" (current-buffer) "host" "smtp"))
will block for long time and cannot be interrupted by hitting C-g.
>Fix:
I don't know this is "correct" patch, but this works for me well.
--- patch begins here ---
Index: files/patch-src:process.c
===================================================================
RCS file: /home/ncvs/ports/editors/emacs/files/patch-src:process.c,v
retrieving revision 1.1
diff -u -d -r1.1 patch-src:process.c
--- files/patch-src:process.c 21 Mar 2004 07:38:53 -0000 1.1
+++ files/patch-src:process.c 23 Apr 2004 10:30:34 -0000
@@ -1,32 +1,9 @@
--- src/process.c.orig Mon Mar 17 07:06:56 2003
-+++ src/process.c Tue Aug 12 01:17:12 2003
-@@ -1850,6 +1850,7 @@
- for (lres = res; lres; lres = lres->ai_next)
- {
- s = socket (lres->ai_family, lres->ai_socktype, lres->ai_protocol);
-+
- if (s < 0)
- {
- xerrno = errno;
-@@ -1882,19 +1883,43 @@
- It'd be nice to be able to control the connect timeout
- though. Would non-blocking connect calls be portable? */
- turn_on_atimers (0);
-+
- ret = connect (s, lres->ai_addr, lres->ai_addrlen);
- xerrno = errno;
- turn_on_atimers (1);
-
-- if (ret == 0 || xerrno == EISCONN)
-+ if (ret == 0 || xerrno == EISCONN )
- /* The unwind-protect will be discarded afterwards.
++++ src/process.c Fri Apr 23 19:28:58 2004
+@@ -1891,10 +1891,34 @@
Likewise for immediate_quit. */
break;
- immediate_quit = 0;
-
-- if (xerrno == EINTR)
-- goto loop;
+#if 1
+ if ( (errno == EINTR ) || (errno == EINPROGRESS)){
+ fd_set fds;
@@ -37,6 +14,7 @@
+ FD_ZERO(&fds);
+ FD_SET(s, &fds);
+ resel:
++ QUIT;
+ e = select(s+1, NULL, &fds, NULL, &tv);
+
+ if( e == -1){
@@ -52,19 +30,17 @@
+ }
+#endif
+
+ immediate_quit = 0;
+
+- if (xerrno == EINTR)
+- goto loop;
if (xerrno == EADDRINUSE && retry < 20)
{
/* A delay here is needed on some FreeBSD systems,
-@@ -2000,11 +2025,30 @@
- && errno != EISCONN)
+@@ -2001,10 +2025,30 @@
{
int xerrno = errno;
--
-+
- immediate_quit = 0;
-- if (errno == EINTR)
-- goto loop;
+#if 1
+ if ( (errno == EINTR ) || (errno == EINPROGRESS)){
+ fd_set fds;
@@ -75,6 +51,7 @@
+ FD_CLEAR(&fds);
+ FD_SET(s, &fds);
+ resel:
++ QUIT;
+ e = select(s+1, NULL, fds, NULL, &tv);
+ if( e != 0){
+ if(errno == EINTR)
@@ -86,6 +63,10 @@
+ xerrno = 0;
+ }
+#endif
+ immediate_quit = 0;
+
+- if (errno == EINTR)
+- goto loop;
if (errno == EADDRINUSE && retry < 20)
{
/* A delay here is needed on some FreeBSD systems,
--- patch ends here ---
>Release-Note:
>Audit-Trail:
>Unformatted:
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20040423104925.973365C86>