On 9/29/2017 6:43 PM, Eric Blake wrote:
On 09/29/2017 06:13 PM, John E. Malmberg wrote:
Hello,
I am confused about the rename managing hardlinks test.
It starts out with conftest.f as an empty file and conftest.fl as a hard
link to conftest.f.
Step 1:
rename ("conftest.f", &q
Hello,
I am confused about the rename managing hardlinks test.
It starts out with conftest.f as an empty file and conftest.fl as a hard
link to conftest.f.
Step 1:
rename ("conftest.f", "conftest.f1")
I would expect after a success, conftest.f would no longer exist, only
conftest.f1.
The TIME_T_IS_SIGNED macro not defined if time_t is unsigned.
This causes the test for working mtime to fail to compile on such
systems such as OpenVMS.
configure:33308: cc -o conftest -g conftest.c >&5
time_t_max = (! TIME_T_IS_SIGNED
..^
%CC-E-UNDECLARED, In this statem
On 7/19/2017 8:31 AM, John E. Malmberg wrote:
Testsuite summary for dummy 0
# TOTAL: 1146
# PASS: 874
# SKIP: 99
# XFAIL: 0
# FAIL
Testsuite summary for dummy 0
# TOTAL: 1146
# PASS: 874
# SKIP: 99
# XFAIL: 0
# FAIL: 173
# XPASS: 0
# ERROR: 0
This is with local changes e
On 7/7/2017 2:43 AM, Bruno Haible wrote:
John E. Malmberg wrote:
I found out what the issue is. The OpenVMS stdio.h file makes the
unlocked IO routines only available as macros. There is no declaration.
Configure detects the macros, and that there is no declaration of them.
The unlocked
On 7/15/2017 10:23 AM, Bruno Haible wrote:
John E. Malmberg wrote:
This patch defines the CMSG_LEN and CMSG_SPACE macros for OpenVMS in
sys_socket.h.
There was no patch attached to your mail.
Sorry, here is the patch.
Regards,
-John
0001-socket-Add-support-for-OpenVMS.patch.gz
This patch defines the CMSG_LEN and CMSG_SPACE macros for OpenVMS in
sys_socket.h.
Regards,
-John
This patch adds the RUSAGE_SELF and RUSAGE_CHILDREN defined constants
missing from the OpenVMS sys/resource.h file.
Regards,
-John
0001-resource.h-OpenVMS-needs-RUSAGE-constants.patch.gz
Description: GNU Zip compressed data
This is a patch to work around that OpenVMS defines NAN in
instead of
Regards,
-John
0001-math-OpenVMS-has-NAN-in-fp.h.patch.gz
Description: GNU Zip compressed data
On 6/28/2017 3:56 PM, Bruno Haible wrote:
John E. Malmberg wrote:
OpenVMS has the function. Configure detects it exists, but the test for
it working is failing.
...
The OpenVMS getdtablesize() is returning the maximum number of open
files that a process can have.
It is not returning the
I have what appears to be a mostly successful build.
That is 1082 object files are created and a libgnu.a is created.
The make is failing in gltests. OpenVMS does not have a concept of
rpaths, so there should not be anything for it to do for rpaths.
Also, I can not run git on VMS.
Are there
On 7/6/2017 5:04 PM, Paul Eggert wrote:
Bruno Haible wrote:
I guess that this would not work:
#include
#include "unlocked-io.h"
would not make the unlocked IO routines visible, because
gets skipped
when it gets included a second time. Right?
If we treat __UNIX_PUTC like we treat _GNU
On 7/3/2017 7:07 PM, Bruno Haible wrote:
> Hello John,
>
>> * Config.h
>> OpenVMS needs __UNIX_PUTC macro defined for putc_unlocked
>> and friends to be visible.
>
> The right place to do this (for a macro that affects multiple
> gnulib modules) is the 'extensions' module.
This makes it v
On 7/3/2017 7:26 PM, Bruno Haible wrote:
John E. Malmberg wrote:
For Bash and Coreutils, I have a close to a pipe replacement that uses a
slightly different API in that you have to know if the pipe is parent to
child or child to parent when you create the pipes. This handles pipes
up to the
On 7/3/2017 7:58 PM, Bruno Haible wrote:
John E. Malmberg wrote:
The OpenVMS Mailboxes work for data up to about 65K and then have buffer
full problems.
...
For Bash and Coreutils, I have a close to a pipe replacement that uses a
slightly different API in that you have to know if the pipe is
On 7/3/2017 7:07 PM, Bruno Haible wrote:
Hello John,
* Config.h
OpenVMS needs __UNIX_PUTC macro defined for putc_unlocked
and friends to be visible.
The right place to do this (for a macro that affects multiple
gnulib modules) is the 'extensions' module.
* lib/passfd.c
This nee
On 7/3/2017 6:30 PM, Bruno Haible wrote:
Hi John,
This Patch gets the spawn-pipe.c and spawni.c modules to compile on OpenVMS.
When I get to the state of testing these, I may need to make some more
changes.
On this one as well, I'd like to ask you to come back when you have
worked out the is
On 7/3/2017 6:26 PM, Bruno Haible wrote:
Hello John,
This patch gets the forkpty.c, grantpt.c, and openpty.c modules building
on OpenVMS.
I will need to figure out later how to actually implement these for
OpenVMS pseudo terminals.
I wouldn't like to apply intermediate patches like this one.
Hello,
This is some of the things that I still need to do to get gnulib to the
point of compiling on OpenVMS.
* Config.h
OpenVMS needs __UNIX_PUTC macro defined for putc_unlocked
and friends to be visible.
* lib/passfd.c
This needs "_X_OPEN_SOURCE_EXTENDED" for the XPG4 V2 features
i
On 7/3/2017 4:07 PM, Paul Eggert wrote:
On 07/03/2017 03:30 PM, John E. Malmberg wrote:
OpenVMS does not have a fork() call and savewd.c is one of the few
cases where the OpenVMS vfork() call can be directly substituted.
This is true only if the application limits itself to a fairly
Hello,
This Patch gets the spawn-pipe.c and spawni.c modules to compile on OpenVMS.
When I get to the state of testing these, I may need to make some more
changes.
It should be possible to implement the spawni() routine on OpenVMS at a
future time.
The OpenVMS pipe() implementation is not
Hello,
OpenVMS does not have a fork() call and savewd.c is one of the few cases
where the OpenVMS vfork() call can be directly substituted.
Regards,
-John
0001-savewd.c-savewd-OpenVMS-uses-vfork.patch.gz
Description: GNU Zip compressed data
Hello,
This patch gets the forkpty.c, grantpt.c, and openpty.c modules building
on OpenVMS.
I will need to figure out later how to actually implement these for
OpenVMS pseudo terminals.
Regards,
-John
0001-pty-No-pty-support-for-OpenVMS.patch.gz
Description: GNU Zip compressed data
On 7/1/2017 7:47 PM, Bruno Haible wrote:
Hi John,
On OpenVMS, the struct _iobuf is needed to access members of the struct
FILE.
This is a patch for the affected stdioext modules to compile on OpenVMS.
The approach looks generally right, but I'd request a couple of changes
before this can go
On OpenVMS, the struct _iobuf is needed to access members of the struct
FILE.
This is a patch for the affected stdioext modules to compile on OpenVMS.
Regards,
-John
0001-stdioext-Port-to-OpenVMS.patch.gz
Description: GNU Zip compressed data
On 6/27/2017 10:13 AM, Bruno Haible wrote:
John E. Malmberg wrote:
The getdtablesize.c module will not compile on OpenVMS and is really not
needed.
...
Is there someway for configure to be changed to not set the macros
causing it to be built when the platform is OpenVMS?
gnulib builds the
Hello,
The getdtablesize.c module will not compile on OpenVMS and is really not
needed.
The way that the an OpenVMS getrlimit() would be most likely implemented
to return the RLIMIT_NOFILE is to call the system supplied getdtablesize().
Is there someway for configure to be changed to not se
On 6/16/2017 9:47 PM, John E. Malmberg wrote:
On 6/16/2017 12:21 PM, Bruno Haible wrote:
Hello John,
VMS does not currently provide a sched.h and has that definition in
pthread.h instead.
Please try the attached patch.
Thank you, I will try to get to that this weekend.
The configure
On 6/16/2017 12:21 PM, Bruno Haible wrote:
Hello John,
VMS does not currently provide a sched.h and has that definition in
pthread.h instead.
Please try the attached patch.
Thank you, I will try to get to that this weekend.
Please also tell us identification of your environment:
- Is i
Hello All,
The make step on VMS next stopped at glthread/cond.c with:
cc -DHAVE_CONFIG_H -DEXEEXT=\"\" -DEXEEXT=\"\" -DNO_XMALLOC
-DEXEEXT=\"\" -I. -I.. -DGNULIB_STRICT_CHECKING=1 -g -c -o
glthread/cond.o glthread/cond.c
struct sched_param
^
%CC-E-REDEFSTRUCT, In this declaration, the str
On 6/15/2017 2:26 PM, Paul Eggert wrote:
On 06/15/2017 07:05 AM, John E. Malmberg wrote:
It should only be include files VMS has issues with.
If that's the problem, then it shouldn't be much trouble to have a
separate script that rewrites the file names and source code to follow
It should only be include files VMS has issues with.
Regards,
-John
--
Sent from my Android device with K-9 Mail. Please excuse my brevity.
I seem to have coerced configure into running on NFS served files
created by on VMS using GNV updated kits.
rm -rf ${GNU_MIRRORS}/gnulib/gnulib
./gnulib-tool --create-testdir --with-tests
--dir=${GNU_MIRRORS}/gnulib/gnulib
# Work around some issue with NFS timestamps.
pushd ${GNU_MIRRORS}/gnu
I seem to have found an issue with how timestamps from an NFS served
volume are showing up on VMS.
I see the correct timestamps on the Linux system, but the wrong
timestamps on the VMS system which is causing make to try to run
autoconf or automake, which I do not yet have available.
So far
On 6/8/2017 5:28 PM, Paul Eggert wrote:
On 06/08/2017 08:40 AM, Bruno Haible wrote:
Feel free to submit a patch to doc/build-automation.texi that references
DEPENDENCIES ...
I gave a shot at patching the documentation. The new version is at:
https://www.gnu.org/software/gnulib/manual/gnulib.h
On 6/9/2017 12:08 PM, Paul Eggert wrote:
On 06/09/2017 05:56 AM, John E. Malmberg wrote:
Fedora 25 Upgrade. VMS will not mount the NFS export on any version.
Does the VMS client mount over UDP only? If so, that is probably the
problem. Recent nfs-utils (as used in Fedora 25) disabled UDP
On 6/8/2017 8:46 AM, John Malmberg wrote:
On 6/8/2017 8:34 AM, Eric Blake wrote:
On 06/08/2017 08:21 AM, John Malmberg wrote:
I issued the command below and it failed on a Fedora/22 system with a
You _do_ realize that Fedora 22 is unsupported upstream, and is probably
lacking in security pat
What is the status of http://autobuild.josefsson.org/gnulib/ ?
It is not reachable by me at the moment.
Regards,
-John
On 6/6/2017 11:01 AM, Bruno Haible wrote:
Hi,
John Malmberg wrote:
1. The VMS CRTL getcwd() will not pass configure tests and will be
marked for replacement from gnulib.
2. The replacement from gnulib causes canonicalize-lgpl.c to fail
because the replacement only has 2 parameters, not three.
On 6/4/2017 4:32 PM, Paul Eggert wrote:
John E. Malmberg wrote:
VMS 8.2 and later provide a 64 bit ino_t type when _USE_STD_STAT is
defined on the command line.
Thanks for the heads-up. I see that 8.2 came out in 2005; I guess news
travels slowly from the VMS world to the Gnulib world.
The
VMS 8.2 and later provide a 64 bit ino_t type when _USE_STD_STAT is
defined on the command line.
This patch makes the fallback to the old VMS behavior for only VMS 8.1
and earlier.
This patch also adds an assert to require that VMS programmers have the
_USE_STD_STAT defined for VMS 8.2+.
R
VMS has files that the information on how to interpret the structure is
stored in the file header, not the file it self.
Currently gnu tar is not recording this information, and I am trying to
determine a portable way to record it.
The best hack that I can come up with is to mangle the name o
Announcing GNU Tar 1.19 for OpenVMS Alpha 8.3.
This version is designed to work with GNV gzip and bzip2, but can also
be used as a standalone utility.
I ran as many of the tests as I could from the test suite for gnu tar,
and was able to create and read an 8 GB archive. This is documented in
Jim Meyering wrote:
"John E. Malmberg" <[EMAIL PROTECTED]> wrote:
If you're going to contribute much more, you'll have to fill out
copyright assignment paperwork. Also (assuming you continue) please
provide ChangeLog entries as well as "git format-patch"
Bruno Haible wrote:
John E. Malmberg wrote:
On VMS, think of "// can be any
device mounted on the system, or logical name.
A logical name behaves like a symbolic link either in the current
directory, or in the root directory as far as the file system is
concerned, but it is stor
Bruno Haible wrote:
Jim Meyering wrote:
How about the patch below instead?
I've moved the prototype down, so that we see
the offensive, "$"-containing symbol name only once ;-)
It is less understandable than John Malmberg's patch, because
- normally you should have a declaration of a functio
Just like on rtapelib, the vms vfork() does not really fork, it sets up
a longjump, and the exec* command starts up the child.
VMS C library maps pipes to a device known as a mailbox, and unlike
pipes, mailboxes are small, limited to 65535 bytes. So for the output
pipe, in external module, a
This is what VMS needs to be able to issue the rsh command.
VMS does not really fork(), what happens is that the fork() calls a
routine named vfork() via a macro which sets a longjump to appear to
return twice.
First it vfork() returns with a status of 0 to pretend to be a child,
but it real
Update VMS special code to work with gnulib headers.
-John
[EMAIL PROTECTED]
Personal Opinion Only
--- /src_root/tar-1.19/lib/canonicalize-lgpl.c Fri Sep 28 08:11:36 2007
+++ /stage_vms_src/tar-1.19/lib/canonicalize-lgpl.c Mon Dec 17 18:37:41 2007
@@ -68,9 +68,15 @@
# include "pathmax.h"
#
John E. Malmberg wrote:
Eric Blake wrote:
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
According to John E. Malmberg on 3/8/2008 1:32 PM:
| I believe that this attached patch gets intprops.h TYPE_MAXIMUM to
| handle unsigned types.
TYPE_MAXIMUM already handles unsigned types. It seems like
Eric Blake wrote:
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
According to John E. Malmberg on 3/8/2008 1:32 PM:
| I believe that this attached patch gets intprops.h TYPE_MAXIMUM to
| handle unsigned types.
TYPE_MAXIMUM already handles unsigned types. It seems like you have a
buggy compiler
I believe that this attached patch gets intprops.h TYPE_MAXIMUM to
handle unsigned types.
A mask is needed because of an extension that most C compilers have that
allow ((unsigned int) -1) to be tested as being less than 0 through
operand promotion.
This still leaves me a problem with getda
The next test for me to get working is options.at.
The issue is that the VMS UID is 32 bits with the GID encoded in it,
unless special mapping is in place.
So to encode the UID in uid_to_chars(), the value needed is (uid >> 16)
unless the special mapping is in place, in which case the UID wou
Bruno Haible wrote:
John E. Malmberg wrote:
For situations like this, gnulib usually "overloads" the system provided
function, so that it has the POSIX specified prototype and behaves like
POSIX says. This allows programmers to write code for POSIX systems, without
#ifdefs, an
Bruno Haible wrote:
John E. Malmberg wrote:
In lib/sys_stat.in :
The VMS DECC compiler needs to test the undefined macros with !defined.
What do you mean by this?
There is a bug where in some cases it does not work, particularly if
more than one undefined macro is present on an #if
Bruno Haible wrote:
John E. Malmberg wrote:
All versions of the above C compilers define the __DECC and __VMS macros.
...
The __ALLOCA is a compiler built in, so it is specific to the compiler
and not to the version of the operating system.
...
I am doing the build with the compiler that
Ralf Wildenhues wrote:
Hello John,
* John E. Malmberg wrote on Mon, Mar 03, 2008 at 07:01:46AM CET:
In tests/testsuite :
VMS can not deal with having a file/directory named testname and
testname.dir in the same directory.
This is because all directories are named .dir, and to be more
Bruno Haible wrote:
John E. Malmberg wrote:
In src/names.c
The declarations for getgrnam and getpwnam interfere with the ones in
the system supplied headers.
--- /src_root/tar-1.19/src/names.c Sun Aug 12 02:46:25 2007
+++ src/names.c Mon Dec 17 19:53:01 2007
@@ -27,6 +27,7
Bruno Haible wrote:
John E. Malmberg wrote:
In lib/alloca.in.h :
VMS needs a definition for the built-in __ALLOCA when using the DECC
compiler.
I committed your change with modifications:
- This definition was at the wrong place. You put it in the middle of a
declaration.
- The style in
I have TAR passing the pipe test, so I though it would be a good point
to pass the source changes I needed so far.
TODO: I still have to add the vfork/exec/pipe code from the VMS bash
port so that the compression options work.
In lib/alloca.in.h :
VMS needs a definition for the built-in __AL
Hello again,
The lib/getcwd replacement algorithm does not work on VMS.
I am still trying to figure out exactly why it is failing.
On VMS, the directory specification /foo/bar/baz can actually be
multiple overlayed directory trees, such as /sys_foo/bar/baz and
/usr_foo/bar/baz.
And on my te
James Youngman wrote:
On Tue, Feb 19, 2008 at 3:46 AM, John E. Malmberg <[EMAIL PROTECTED]> wrote:
Hello,
I am trying to get gnu tar to build on VMS.
I have run into a problem: GNU TAR now requires that the open() call
support O_DIRECTORY.
You may find some useful resources at
Bruno Haible wrote:
John E. Malmberg wrote:
The open() would have to open or dup a file descriptor to the /dev/null
device to reserve a unique file descriptor, and then something needs to
map that file descriptor back to the pointer to the *DIR structure.
Do you need a pointer to the DIR
Eric Blake wrote:
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
According to John E. Malmberg on 2/18/2008 8:46 PM:
| I have run into a problem: GNU TAR now requires that the open() call
| support O_DIRECTORY.
Tar is using O_DIRECTORY to close a security hole, and not to directly
read the
Hello,
I am trying to get gnu tar to build on VMS.
I have run into a problem: GNU TAR now requires that the open() call
support O_DIRECTORY.
On the VMS platform, this support is not available. Using open() on a
directory will always fail. Only opendir() can be used, and readdir()
must be
Paul Eggert wrote:
"John E. Malmberg" <[EMAIL PROTECTED]> writes:
The quick patch that I can do is to move where the #undef of the
_GL_JUST_INCLUDE_SYSTEM_DIRECTORY_H is.
Yes, thanks, that sounds like a reasonable fix. I installed the patch
below: does it work for you?
I
Hello,
I am attempting to build tar-1.19 on VMS.
I am getting an error assert from inttypes.h:
"This file assumes that 'int' has exactly 32 bits. Please report your
platform and compiler to .
From the analysis of the macro values used to generate the error
assert, I have discovered that IN
68 matches
Mail list logo
