At 12:25 31.03.2003 +0200, Simo Sorce wrote:
On Mon, 2003-03-31 at 03:30, Andrew Bartlett wrote:
As per my recent commit, the new (VFS) modules system completely breaks
on tree disconnect!
We need to separate the different cases - the compat and the central
modules, and provide either a flag
At 13:02 31.03.2003 +0200, Simo Sorce wrote:
Eh, the shutdown stuff was just a thing, I was thinking yesterday ...
I agree we should have to way to startup and shutdown the modules, as we
have to way to load it (preload and fork).
to - two
this is what my patch is about :-)
the old modules
On Mon, 2003-03-31 at 03:30, Andrew Bartlett wrote:
As per my recent commit, the new (VFS) modules system completely breaks
on tree disconnect!
We need to separate the different cases - the compat and the central
modules, and provide either a flag or a function pointer to the correct
way
On Mon, 2003-03-31 at 12:25, Simo Sorce wrote:
Eh, the shutdown stuff was just a thing, I was thinking yesterday ...
I agree we should have to way to startup and shutdown the modules, as we
have to way to load it (preload and fork).
to - two
This is mandatory for modules that uses databases
On Mon, Mar 31, 2003 at 11:30:42AM +1000, Andrew Bartlett wrote about 'New modules
system and vfs_done':
As per my recent commit, the new (VFS) modules system completely breaks
on tree disconnect!
We need to separate the different cases - the compat and the central
modules, and provide
As per my recent commit, the new (VFS) modules system completely breaks
on tree disconnect!
We need to separate the different cases - the compat and the central
modules, and provide either a flag or a function pointer to the correct
way to shut down a module.
The code in conn_close is really
At 12:12 27.03.2003 +0100, Jelmer Vernooij wrote:
I noticed that all smb_register_*() fn's return BOOL
and the init_module() fn's returns int
so this is wrong:
I don't have problems with this. We only need the return value in
smb_probe_module() to check for failure, and False is defined to be
Not that this is correct or anything... just my 2cents.
I tend in my own code to return negative integers to indicate warnings or
errors, with different negative values having different meaning (as
needed). That way, I can simply check for 0 on return. If I want to
get specific about what
Hi!
Here's a patch that adds support to the VFS for the new modules
system. The only problem is that it breaks all current modules, but
I don't think I can do anything about that.
Any objections?
Jelmer
--
Jelmer Vernooij [EMAIL PROTECTED] - http://nl.linux.org/~jelmer/
23:26:54 up 8 days
At 23:29 25.03.2003 +0100, Jelmer Vernooij wrote:
I took Jelmer's patch and modifed it,
to support an easy way to have per connection default_vfs_ops
and per connection private data in the vfs modules
so every module registers and got a unique vfs module number for te current
smbd or vfstest
At 00:52 26.03.2003 +0100, Stefan (metze) Metzmacher wrote:
At 23:29 25.03.2003 +0100, Jelmer Vernooij wrote:
I took Jelmer's patch and modifed it,
to support an easy way to have per connection default_vfs_ops
and per connection private data in the vfs modules
so every module registers and got
Hi Jelmer,
I found this in proto.h
/* The following definitions come from modules/xml.c */
BOOL parsePass(xmlDocPtr doc, xmlNsPtr ns, xmlNodePtr cur, SAM_ACCOUNT * u);
BOOL parseUser(xmlDocPtr doc, xmlNsPtr ns, xmlNodePtr cur, SAM_ACCOUNT * u);
NTSTATUS xmlsam_init(PDB_CONTEXT * pdb_context
Hi,
I've finally committed my modules patch to CVS. Here are the docs (now
in dev-doc as well).
Next things I'll work on:
- Adapting VFS to use the new modules system
- Winsdb (after metze resends his patch)
Tomorrow I continue learning for my exams so the VFS patch
probably won't come
Will this break compatability with auth plugins?
-- Luke
--
Luke Howard | PADL Software Pty Ltd | www.padl.com
On Thu, Mar 20, 2003 at 10:45:04PM +1100, Luke Howard wrote about 'Re: The new modules
system':
Will this break compatability with auth plugins?
Currently not as auth and vfs don't use the new modules system yet
(only pdb and rpc). I will have to break auth plugin compatibility
when I'm
Will the migration to the new format be difficult and/or documented?
-- Luke
From: Jelmer Vernooij [EMAIL PROTECTED]
Subject: Re: The new modules system
To: Luke Howard [EMAIL PROTECTED]
Cc: [EMAIL PROTECTED]
Date: Thu, 20 Mar 2003 13:12:39 +0100
On Thu, Mar 20, 2003 at 10:45:04PM +1100, Luke
On Fri, 2003-03-21 at 00:15, Luke Howard wrote:
Will the migration to the new format be difficult and/or documented?
Not difficult, and I trust ctrlsoft will update the example module at
the same time.
Basically we will just change the init function - I don't think it will
take you much to
090156 metze ctrlsoft: I think the modules stuff should not merged into 3_0 untill
we have discused how module loading should be finally done in samba
What I will be merging into 3_0 at first is just some basic helper
functions for the modules. it's currently only the smb_load_module()
function
At 11:43 21.02.2003 +0100, Jelmer Vernooij wrote:
090156 metze ctrlsoft: I think the modules stuff should not merged into
3_0 untill
we have discused how module loading should be finally done in samba
What I will be merging into 3_0 at first is just some basic helper
functions for the modules
On Tue, Nov 26, 2002 at 08:06:07AM +0100, Stefan (metze) Metzmacher wrote about 'Re:
[PATCH] add 'modules path' and handle 'configure --with-configdir' right':
here's the next version of the patch. (it's now attached :-)
I've applied a part of your patch, but I'm waiting with the
lp_modules_dir
Hi Jelmer,
I noticed that most vfs modules use global vars, I think we should avoid
this because the modules are configurable for each share (connection). and
the modules should use a connection context to store the configuration data
to not conflict,
I'm currently writing a vfs_antivir
On Tue, Nov 26, 2002 at 09:27:13AM +0100, Stefan (metze) Metzmacher wrote:
Hi Jelmer,
I noticed that most vfs modules use global vars, I think we should avoid
this because the modules are configurable for each share (connection). and
the modules should use a connection context to store
At 12:08 26.11.2002 +0200, Alexander Bokovoy wrote:
vfs_private was designed for it. Modules which use global vars should be
fixed, indeed, I've tried to do so during introduction of stacked VFS.
But in current CVS HEAD we have a number of misuse of vfs_private, for
example, in vfs_recycle
msg.pgp
Description: PGP message
to sambadatadir
I'm happy with this patch, but NOT with the name 'sambadatadir' DOES
ANYBODY HAVE ABETTER IDEA?
I removed the modules/Makfile.ext_modules.in stuff
I think we should put ALL header files in include/ and then after make
proto we can install all headers to ${prefix}/include/samba
so
At 11:23 23.11.2002 +1100, Tim Potter wrote:
On Fri, Nov 22, 2002 at 01:21:21PM +0100, Jelmer Vernooij wrote:
here's a patch that added the 'modules path' parameter.
lp_modules_path() is prefixed to all lp_modules() witch are not start
with
'/' (not absolute pathes)
the default
the patch doesn't compile fine.
- I forgot to add dyn_MODULESDIR to include/dynconfig.h
- in param/loadparm.c szModulesPath shold be char * not char **
- add a warning to modules/Makefile.ext_modules.in : that the users should
be careful with editing this file
If you want me to send a new
At 12:00 22.11.2002 +0100, Stefan (metze) Metzmacher wrote:
If you want me to send a new patch tell me.
anyway I'll send a new patch with a few more fixes.
But we have to discuss on IRC...
metze
-
Stefan metze
On Fri, Nov 22, 2002 at 11:36:48AM +0100, Stefan (metze) Metzmacher wrote:
Hi Jelmer,
here's a patch that added the 'modules path' parameter.
lp_modules_path() is prefixed to all lp_modules() witch are not start with
'/' (not absolute pathes)
the default for lp_modules_path
At 13:21 22.11.2002 +0100, Jelmer Vernooij wrote:
On Fri, Nov 22, 2002 at 11:36:48AM +0100, Stefan (metze) Metzmacher wrote:
Hi Jelmer,
here's a patch that added the 'modules path' parameter.
lp_modules_path() is prefixed to all lp_modules() witch are not start with
'/' (not absolute pathes
On Fri, Nov 22, 2002 at 01:21:21PM +0100, Jelmer Vernooij wrote:
here's a patch that added the 'modules path' parameter.
lp_modules_path() is prefixed to all lp_modules() witch are not start with
'/' (not absolute pathes)
the default for lp_modules_path() is selectable
yep, see samba/examples/VFS/
there are few modules here,
Simo.
On Wed, 2002-10-09 at 19:38, Kris Van Hees wrote:
Has anyone implemented a VFS module already? It does not seem that there are
any as part of the CVS HEAD version, and I could not find a reference to any
on the web site
On Wed, Oct 09, 2002 at 01:38:21PM -0400, Kris Van Hees wrote:
Has anyone implemented a VFS module already? It does not seem that there are
any as part of the CVS HEAD version, and I could not find a reference to any
on the web site.
HEAD:/samba/examples/VFS/
--
/ Alexander Bokovoy
---
He
On Wed, Oct 09, 2002 at 01:38:21PM -0400, Kris Van Hees wrote about 'VFS modules?':
Has anyone implemented a VFS module already? It does not seem that there are
any as part of the CVS HEAD version, and I could not find a reference to any
on the web site.
see the examples/VFS directory in CVS
VFS - modlues
I want to check the fd on my first write call to see if it just got opened
by write call or is a susequent write call.
The purpose is to preform a backup only if the user actually write data to
thefile.
I have been study the vfs.h, smb.h, skel.c to figure this out
fsp-prev-fd
Juergen Hasch wrote:
Am Montag, 20. Mai 2002 01:18 schrieb Juergen Hasch:
Am Montag, 20. Mai 2002 00:56 schrieb Juergen Hasch:
Hi,
attached is a patch and a few new files to make VFS modules work under
Solaris and AIX.
I have tested it with Solaris 8 using gcc and Sun Workshop cc
Am Montag, 20. Mai 2002 01:18 schrieb Juergen Hasch:
Am Montag, 20. Mai 2002 00:56 schrieb Juergen Hasch:
Hi,
attached is a patch and a few new files to make VFS modules work under
Solaris and AIX.
I have tested it with Solaris 8 using gcc and Sun Workshop cc. For AIX I
tested
Am Montag, 20. Mai 2002 00:56 schrieb Juergen Hasch:
Hi,
attached is a patch and a few new files to make VFS modules work under
Solaris and AIX.
I have tested it with Solaris 8 using gcc and Sun Workshop cc. For AIX I
tested it with gcc and xlc_r. Linux still works, too
While trying to compile the VFS modules for Samba 2.2.4 under
FreeBSD 4.5-STABLE, I encountered several problems:
1. Missing include directory, unable to find popt.h if using bundled popt library.
2. Undefined Symbols when smbd loads a VFS module.
- defined DYNEXP for the *bsd* case
3
Hi,
I've attched a patch for the examples/VFS modules of 2.2 CVS to build
all modules successful.
--
Lars Müller
SuSE Linux AG, Deutschherrnstraße 15-19, 90429 Nürnberg, Germany
Find daily samba CVS 2.2 RPMs for SuSE Linux 7.2, 7.3 and 8.0 at
ftp://ftp.SuSE.com/pub/people/lmuelle
Lars Mueller schrieb:
Hi,
I've attched a patch for the examples/VFS modules of 2.2 CVS to build
all modules successful.
Hm, if -D_LARGEFILE64_SOURCE is set probably -D_FILE_OFFSET_BITS=64
should be set, too.
I wrote a qd Makefile.in for examples/VFS, so that Makefile is
generated
41 matches
Mail list logo