>>>>> Ben Collins writes:
Ben> On Fri, Apr 16, 1999 at 06:52:34PM +0200, Andreas Jaeger wrote:
>> > There were a few differences: They have nothing resembling our
>> > sigaction, and sigstack patches (are they still needed?) They have an
>> > additional "libio-vfork" patch, which may fix our problems with
>> > non-cvs kernels (I'm not sure, but I've attached it so you can look at
>> > it). Another difference is that RH is using glibc-990329.tar.gz as a
>> > base and we are using glibc-19990406.tar.gz.
>>
>> The vfork patch is superseded by:
>> 1999-03-29 Paul Eggert <[EMAIL PROTECTED]>
>>
>> * libio/iopopen.c (_IO_new_proc_open): Don't modify proc_file_chain
>> while in child process.
>>
>> which should be in 2.1.1-0.2.
Ben> This does not show in the ChangeLog for 2.1.1-0.2.
The patch went in the same day and therefore it should be in 2.1.1-0.2
(looking at the dates). Nevertheless, I'm appending the patch so that
you can check yourself.
Andreas
Index: libio/iopopen.c
===================================================================
--- libio/iopopen.c 1998/11/23 09:31:38 1.16
+++ libio/iopopen.c 1999/03/29 16:01:58 1.17
@@ -139,6 +139,8 @@
if (child_pid == 0)
{
int child_std_end = mode[0] == 'r' ? 1 : 0;
+ struct _IO_proc_file *p;
+
_IO_close (parent_end);
if (child_end != child_std_end)
{
@@ -148,11 +150,8 @@
/* POSIX.2: "popen() shall ensure that any streams from previous
popen() calls that remain open in the parent process are closed
in the new child process." */
- while (proc_file_chain)
- {
- _IO_close (_IO_fileno ((_IO_FILE *) proc_file_chain));
- proc_file_chain = proc_file_chain->next;
- }
+ for (p = proc_file_chain; p; p = p->next)
+ _IO_close (_IO_fileno ((_IO_FILE *) p));
_IO_execl ("/bin/sh", "sh", "-c", command, (char *) 0);
_IO__exit (127);
--
Andreas Jaeger [EMAIL PROTECTED] [EMAIL PROTECTED]
for pgp-key finger [EMAIL PROTECTED]
