gmane.network.netatalk.devel

Re: New UAM DHX2

didier — 2008-11-22T15:37:33

Hi, Le samedi 22 novembre 2008 à 15:35 +0100, Frank Lahm a écrit : Oops, my mistake, I've double checked and it works. Didier ------------------------------------------------------------------------- This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK & win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100&url=/

Re: New UAM DHX2

Frank Lahm — 2008-11-22T14:35:55

Hi Didier, 2008/11/22 didier magic.fr>: Leftover. I must have missed them in the cleanup phase. I'll look into it and submit a fix as needed, I have to take a look at it. I don't line that syntax. But the possibility to use libgrypt if installed in nonstandard places would be usefull of course. I can change that. Hm, it did here afair. I've tested all 4 possible combinations. I'll look into it. -Frank ------------------------------------------------------------------------- This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK & win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100&url=/

Re: New UAM DHX2

didier — 2008-11-22T12:29:49

Hi, Le vendredi 21 novembre 2008 à 16:38 +0100, Frank Lahm a écrit : Commit in cvs HEAD but: 1) There's warning msgs: uams_dhx2_passwd.c: Dans la fonction «dhx2_setup» : uams_dhx2_passwd.c:176: attention : unused variable «g_len» uams_dhx2_passwd.c:174: attention : unused variable «ctxerror» uams_dhx2_passwd.c:173: attention : unused variable «ctx» uams_dhx2_passwd.c:170: attention : unused variable «g_uint» uams_dhx2_passwd.c:168: attention : unused variable «i» uams_dhx2_passwd.c:165: attention : unused parameter «ibuf» uams_dhx2_passwd.c: Dans la fonction «logincont1» «ctxerror»«ctxerror»: uams_dhx2_passwd.c:367: attention : ISO C90 forbids mixed declarations and code uams_dhx2_passwd.c:362: attention : unused variable «retID» uams_dhx2_passwd.c:358: attention : unused parameter «obj» uams_dhx2_passwd.c:358: attention : unused parameter «uam_pwd» uams_dhx2_passwd.c: Dans la fonction «logincont2» : uams_dhx2_passwd.c:508: attention : ISO C90 forbids mixed declarations and cod

Re: New UAM DHX2

Frank Lahm — 2008-11-21T15:38:49

Hi again, 2008/11/18 Frank Lahm googlemail.com>: Done. In uam_dhx2_pam, not in passwd. I'd say that should be sufficient, shouldn't it? Also fixed: some memory leaks. Its hard to track these gcry_mpis ! Complete patch attached. -Frank ------------------------------------------------------------------------- This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK & win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100&url=/_______________________________________________ Netatalk-devel mailing list Netatalk-devel< at >lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/netatalk-devel

Re: New UAM DHX2

Frank Lahm — 2008-11-18T18:42:35

Hi everybody, 2008/11/12 Frank Lahm googlemail.com>: Done. Usage: install libgcrypt and libgcrypt-dev, bootstrap, configure. I've only build and run it on Debian Etch so far. To do: change pw support. Please test, audit, report. -Frank ------------------------------------------------------------------------- This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK & win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100&url=/_______________________________________________ Netatalk-devel mailing list Netatalk-devel< at >lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/netatalk-devel

force -5006 error?

Thomas Kaiser — 2008-11-17T18:28:31

Hello all, slightly off-topic but maybe of interest for other AFP/netatalk developers... I'm searching for a way to create a situation where a client accessing a specific file on an AFP share get's reproducible a -5006 error. This error is described as: 5006 afpDenyConflict Permission/Deny mode conflicts with the current mode in which this fork is already open We're struggling with a strange behaviour caused by InDesign. Sometimes ID omits placed images from print output or PDF generation due to a file lock. We've looked at traces and have always seen FPOpenFork reply: file synchronization locks conflict (-5006) in such cases. A reproducible way to lock files in that specific behaviour (either from the client's side or on the server -- maybe with a specific Netatalk tool/ setting I'm not aware of) would be very helpful since we need not to wait until the next such lock happens by accident and causes press spoilage. Anyone? T

Re: New UAM DHX2

Frank Lahm — 2008-11-13T10:00:32

Hi Didier, thanks for the explanations. 2008/11/13 didier magic.fr>: I've had already understood these parts. Thats the part I'm worried about: in DHX2 FPLoginCont will be called twice. So this means I have to demultiplex that inside my logincont function ? ...I'd say yes. Only in that I still have to find a one shot func in libgrypt to generate p ang g, besides that ... Not by name, but the math funcs building DH are there. CAST128 encryption too. Thanks again, -Frank ------------------------------------------------------------------------- This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK & win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100&url=/

Re: New UAM DHX2

Bolke de Bruin — 2008-11-13T07:04:36

Hi, It might NOT help you out as much, but some time ago I started writing some code for DHX2. I attached it. If I remember correctly it is a mix currently between dhc and dhx2 and it definitely does not work at the moment, but maybe it could give you some ideas. regards, Bolke. Op 13 nov 2008, om 05:24 heeft didier het volgende geschreven: ------------------------------------------------------------------------- This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK & win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100&url=/_______________________________________________ Netatalk-devel mailing list Netatalk-devel< at >lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/netatalk-devel

Re: New UAM DHX2

didier — 2008-11-13T04:24:08

Hi, Le mercredi 12 novembre 2008 à 22:32 +0100, Frank Lahm a écrit : All uams modules functions are called from etc/afp/auth.c with helper functions in etc/afpd/uam.c Roughly in your code you define a structure: UAM_MODULE_EXPORT struct uam_export MODULE_NAME = { UAM_MODULE_SERVER, UAM_MODULE_VERSION, uam_setup, uam_cleanup }; with MODULE_NAME your module file name (without .so). Some uams define this structure twice because for legacy reason they may have two names. Ex: UAM_MODULE_EXPORT struct uam_export uams_clrtxt = { ... uam_setup function is called by afpd/uam.c when loading the module and uam_cleanup on exit (both can be called many times if config files are reloaded). in uam_setup you called: uam_register(UAM_SERVER_LOGIN_EXT, path, "UAM string", login, logincont, logout, login_ext) "UAM string" as defined in the spec, ex: DHCAST128 if your module can change password add: uam_register(UAM_SERVER_CHANGEPW, path, "UAM string", changepw) and if it can be used for print

New UAM DHX2

Frank Lahm — 2008-11-12T21:32:30

I've begun implementing DHX2 with libgcrpyt. Can anybody give me some pointers on where and how the necessary chaining is organized ? I've just begun to look at the relevant places and although tags make digesting the code within emacs quite nice I'd appreciate any shortcuts! -Frank ------------------------------------------------------------------------- This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK & win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100&url=/

Re: Fix for .cvsignore in man/[man1|man8]

didier — 2008-11-12T07:17:42

Hi, Le mercredi 05 novembre 2008 à 10:51 +0100, Frank Lahm a écrit : Checked in, thanks Didier ------------------------------------------------------------------------- This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK & win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100&url=/

Re: case insensitive option

didier — 2008-11-12T07:17:06

Hi, Le jeudi 09 octobre 2008 à 00:40 +0900, HAT a écrit : Sorry for the delay but my netatalk subscription was closed, I disabled it because in some cases it's broken, I will fix and put it back. Didier. ------------------------------------------------------------------------- This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK & win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100&url=/

Fix for .cvsignore in man/[man1|man8]

Frank Lahm — 2008-11-05T09:51:22

Both man/[man1|man8]/. cvsignore contain *.1 in order to ignore generated mans. But some mans aren't generated. Fix: man/man1/.cvsignore should contain: --- Makefile Makefile.in apple_cp.1 apple_mv.1 apple_rm.1 asip-status.pl.1 uniconv.1 --- man/man8/.cvsignore should contain: --- Makefile Makefile.in afpd.8 atalkd.8 cnid_dbd.8 cnid_metad.8 papd.8 papstatus.8 psf.8 --- -Frank ------------------------------------------------------------------------- This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK & win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100&url=/

Re: OpenSSL and licensing

Andrew Morgan — 2008-10-21T16:48:48

So, why don't you find all these old email addresses and contact them for us? No need to fork the project or anything. Andy ------------------------------------------------------------------------- This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK & win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100&url=/

OpenSSL and licensing

John Barberio — 2008-10-21T08:18:06

Hi, I'm rather dismayed that altho this has been known to be a problem since 2004, nothing has been done about it. I'm further dismayed that the most recent concerns were dismissed as only 'debian and ubuntu users whining'. Now, considering the fast increasing popularity of Ubuntu, I'd have thought Ubuntu users were a pretty important group now. Now, the objection to adding the required OpenSSL exception to the licence used, is that it would require consent from all the copyright holders. This is not true. As this is an attempt to correct an honest mistake in the licence, It only requires a good faith attempt to contact all the copyright holders, which can be done here, and a trawl of email addresses in old archives. Send a message saying the exception will be added, and that since the use of OpenSSL was a fundamental part of the project, it had been assumed that the exception was implicitly accepted. In the *hugely* unlikely case that someone objects, then don't make the change

Re: case insensitive option

HAT — 2008-10-08T15:40:07

In general, home directory is case-sensitive. Didier, why did you remove case-insesitive option? It is necessary to leave the setting to the admin.

case insensitive option

HAT — 2008-09-05T16:51:39

Recent CVS HEAD version was removed case insensitive option. When running it together with the samba, the problem occurs. I think that casefold option should be extended. # casefold options [syntax: casefold:option]: # tolower -> lowercases names in both directions # toupper -> uppercases names in both directions # xlatelower -> client sees lowercase, server sees uppercase # xlateupper -> client sees uppercase, server sees lowercase # sens -> standard HFS+ compatible # insens -> case-insensitive HFSX compatible

Re: netatalk: cannot login as root

Thomas Kaiser — 2008-09-04T06:29:55

And the list wouldn't get flooded with questions from people that do not even want to understand what netatalk is. Which is not that bad IMO :-) Perfect. Netatalk isn't the tool of choice to modify the file system of lifestyle gadgets like iPhones. It's an AFP server implementation (amongst other things) for *file servers* that stores files in a way incompatible with your iPhone's file system. So if you're thinking about bug reports inform the person(s) that made it possible to install Netatalk on an iPhone instead ('Cydia'?). Not be able to login as root is definitely a feature and not a bug. If you're interested in details please compare with http://article.gmane.org/gmane.network.netatalk.devel/32 and become familiar with the -admingroup feature in afpd.conf: http://netatalk.sourceforge.net/2.0/htmldocs/afpd.conf.5.html And please have always in mind that Netatalk is a file server implementation for Unix servers and Apple Clients and not a way to share your iPhone's root filesystem (exp

Re: netatalk: cannot login as root (fwd)

Andrew Morgan — 2008-09-03T19:32:55

---------- Forwarded message ---------- Date: Wed, 03 Sep 2008 15:26:23 -0400 From: metaclam gmail.com> To: Andrew Morgan orst.edu> Subject: Re: netatalk: cannot login as root Andy --- thanks a bunch but seriously -- too much trouble for something which, though I'm appreciative of, I probably will rarely use. If your group makes it so hard to post a bug report or email someone responsible, lots of bugs will go unreported. I tried for at least 15-20 minutes including this message and frankly that's too much. I could spend a good part of my day reporting bugs if it took that long each time. Please feel free to pass this message on or trash it. Thanks. On 08/31/08 1:25 AM, "Andrew Morgan" orst.edu> wrote: ------------------------------------------------------------------------- This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK & win great prizes Grand prize is a trip for two to an Open

Re: Permission semantics: inherited privs withUNIX privs and strict POSIX privs

Frank Lahm — 2008-08-26T14:28:51

2008/8/26 didier magic.fr>: I've none left from 10.5.[2|3] and -- uhoh -- now with 10.5.4 Apple's pulling the plug: http://lists.apple.com/archives/xsan-users/2008/Jul/msg00004.html: the ability to set a share point to use inheritance is gone from the ServerAdmin GUI in 10.5.4. It's still there in sharing(1), but first testing gives me the impression that inherited privs has been removed completely: when I `touch /Volumes/Inherited_10.5.4Server_Volume/777_dir/file` from a 10.4.11 client (777_dir is, you guess, 777) file is 644. Hm. I see you _raise_ permission for file g and U. I guess I've never tested raising permission in my tests with 10.5.[2|3] afpd and with HELIOS Ethershare 4.5. ;-( Indeed it looks like (with HELIOS Ethershare) as you can raise permission above perms of the parent folder. But you can't lower. As far as I can tell from your trace this is the case there too: - in #236 you create file "g" ----> how did you creat it? From Finder.app or from CLI? - in #242 we see that the

Re: Permission semantics: inherited privs withUNIX privs and strict POSIX privs

Frank Lahm — 2008-08-25T12:14:47

2008/8/22 Frank Lahm googlemail.com>: Done. Answer part one: we can't, we're missing the corresponding #define for (1<<8) in volume.h: /* volume attributes */ ...last one: #define VOLPBIT_ATTR_NONETUID (1 << 7) Part two: Apples afpd 10.5 doesn't set "kDefaultPrivsFromParent" for volumes with inherited semantics. AFAIR 10.4.x doesn't set it too. Frank ------------------------------------------------------------------------- This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK & win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100&url=/