Andrew Tridgell [Fri, 8 Jul 2005 10:56:20 +0000 (10:56 +0000)]
r8242: support bitmap constants from ejs calls
Andrew Tridgell [Fri, 8 Jul 2005 10:41:27 +0000 (10:41 +0000)]
r8241: - take advantage of pointer comparison
- added testing of samr_EnumDomainGroups()
Andrew Tridgell [Fri, 8 Jul 2005 10:39:16 +0000 (10:39 +0000)]
r8240: support comparing pointers in ejs. This allows for
if (io.output.samr == NULL) {
print("no result\n");
}
Andrew Tridgell [Fri, 8 Jul 2005 10:30:50 +0000 (10:30 +0000)]
r8239: - added testing of LookupDomain, OpenDomain and EnumDomainUsers
- used SEC_FLAG_MAXIMUM_ALLOWED for access mask. This gets the samr.js
test working with w2k3 as the server
Andrew Tridgell [Fri, 8 Jul 2005 10:29:18 +0000 (10:29 +0000)]
r8238: - fixed handling of NULL pointers from ejs
- added automatic creation of all constants in IDL as ejs variables
Andrew Tridgell [Fri, 8 Jul 2005 09:38:41 +0000 (09:38 +0000)]
r8237: expanded the samr.js test to do a samr_Connect(), samr_EnumDomains() and samr_Close()
Andrew Tridgell [Fri, 8 Jul 2005 09:37:54 +0000 (09:37 +0000)]
r8236: fixed support for arrays of structures
Stefan Metzmacher [Fri, 8 Jul 2005 09:00:15 +0000 (09:00 +0000)]
r8235: fix the build
metze
Andrew Tridgell [Fri, 8 Jul 2005 08:19:59 +0000 (08:19 +0000)]
r8234: started on testing samr calls from ejs. So far it only does samr_Connect()
Andrew Tridgell [Fri, 8 Jul 2005 08:18:52 +0000 (08:18 +0000)]
r8233: - added support for more base types in pidl ejs
- added auto generation of a header with prototypes for public ejs functions
- make public functions non-static
- fixed allocation of fixed sized arrays
- added 'noejs' flag indicating that a typedef will be handled manually by ejs
- added manual functions for sid and GUID, so they show up as nice
strings in ejs scripts
This allows ejs to bring in samr, security, lsa and misc IDL functions
Stefan Metzmacher [Fri, 8 Jul 2005 08:09:02 +0000 (08:09 +0000)]
r8232: remove samr_String and netr_String as they are the same as lsa_String
metze
Andrew Tridgell [Fri, 8 Jul 2005 07:26:45 +0000 (07:26 +0000)]
r8230: prevent authentication dying on a NULL domain
Volker Lendecke [Fri, 8 Jul 2005 06:29:06 +0000 (06:29 +0000)]
r8229: A level2 oplock does not conflict a no-oplock open, but is broken on first
write. Test that.
Volker
Stefan Metzmacher [Fri, 8 Jul 2005 05:27:47 +0000 (05:27 +0000)]
r8228: fix compiler warnings
metze
Stefan Metzmacher [Fri, 8 Jul 2005 05:22:31 +0000 (05:22 +0000)]
r8227: add STR_LARGE_SIZE flag, to support strings where the size is length+1,
metze
Stefan Metzmacher [Fri, 8 Jul 2005 05:19:28 +0000 (05:19 +0000)]
r8226: w2k3 adds a '.' after the forest and domain dns names for the primary domain
and NULL for the trusted domains forest dns name
metze
Stefan Metzmacher [Fri, 8 Jul 2005 05:16:43 +0000 (05:16 +0000)]
r8225: make nETBIOSName case insensitive, so that lower case netbios domain REALM's
are working in the hdb-ldb module
metze
Stefan Metzmacher [Fri, 8 Jul 2005 05:14:46 +0000 (05:14 +0000)]
r8224: - add objectGUID ldif_handler
- fix some compiler warnings
metze
Stefan Metzmacher [Fri, 8 Jul 2005 05:10:02 +0000 (05:10 +0000)]
r8223: fix the values of nt_version and type we reply in the server code,
also name the struct like the nt_version number
metze
Stefan Metzmacher [Fri, 8 Jul 2005 05:04:57 +0000 (05:04 +0000)]
r8222: 0 entries are no error, unless it's a base search
metze
Andrew Tridgell [Fri, 8 Jul 2005 04:55:07 +0000 (04:55 +0000)]
r8220: added auto-generation of ENUM constants in ejs wrapper. So we can now use the enum name
instead of a integer in ejs scripts making rpc calls
Andrew Tridgell [Fri, 8 Jul 2005 04:16:36 +0000 (04:16 +0000)]
r8218: added testing of echo_TestSurrounding() and
echo_TestDoublePointer(). All echo.idl functions are now tested and
working from ejs.
Andrew Tridgell [Fri, 8 Jul 2005 04:08:17 +0000 (04:08 +0000)]
r8217: added testing of echo_TestSleep() and echo_TestEnum() in echo js code
Andrew Tridgell [Fri, 8 Jul 2005 04:07:26 +0000 (04:07 +0000)]
r8216: - handle union pull in ejs pidl generation
- added debugging calls for missing structure/union elements
Andrew Tridgell [Fri, 8 Jul 2005 03:33:32 +0000 (03:33 +0000)]
r8215: switched the pull side of the ejs generator over to the recursive LEVELS based approach.
This allows for much more complex structure mappings to be generated.
Andrew Tridgell [Fri, 8 Jul 2005 03:11:17 +0000 (03:11 +0000)]
r8214: added testing of echo_TestCall2(), which tests the union push code
Andrew Tridgell [Fri, 8 Jul 2005 03:10:30 +0000 (03:10 +0000)]
r8213: I've started to understand the LEVELS stuff in pidl much better now,
and have re-coded the push side of the ejs generator to use it
properly. It ends up being very neat and small, and should handle much
more complex structures (like arrays of pointers to unions etc).
Also added push side support for unions. This should get more of the
echo pipe working via ejs.
Stefan Metzmacher [Thu, 7 Jul 2005 19:49:35 +0000 (19:49 +0000)]
r8212: fix pushing of nbt_string's:
- we now use an ndr_token_list, for the nbt string label pointer offsets
this avoids to scan the whole buffer
- we need to check for already send string on a per component basis
not only for the fullname
e.g.
w2k3 response this in the CLDAP netlogon replies
forest: w2k3.vmnet1.vm.base
dns_name: sub1.
pdc_dns_name: w2k3-104.
and this will be interpreted like
forest: w2k3.vmnet1.vm.base
dns_name: sub1.w2k3.vmnet1.vm.base
pdc_dns_name: w2k3-104.w2k3.vmnet1.vm.base
metze
Stefan Metzmacher [Thu, 7 Jul 2005 19:40:25 +0000 (19:40 +0000)]
r8211: fix some cldap replies
metze
Stefan Metzmacher [Thu, 7 Jul 2005 19:13:32 +0000 (19:13 +0000)]
r8210: - make the ndr_token_* function public
- allow comparison function to be passed for ndr_token_retrive_cmp_fn(),
this is for matching the keys, if NULL is passed, the old behavior
tok->key == key is used
metze
Stefan Metzmacher [Thu, 7 Jul 2005 11:16:25 +0000 (11:16 +0000)]
r8203: fix the build temporary,
we also get perl warnings uninitialized value used in ejs.pm line 26
metze
Andrew Tridgell [Thu, 7 Jul 2005 08:32:35 +0000 (08:32 +0000)]
r8200: - added stub functions for union pull/push
- kill the js interpreter with an exception on internal errors
Andrew Tridgell [Thu, 7 Jul 2005 08:21:45 +0000 (08:21 +0000)]
r8199: - we don't need to pre-declare 'var status;' everywhere
- added testing of echo_TestCall(), which uses strings
Andrew Tridgell [Thu, 7 Jul 2005 08:20:57 +0000 (08:20 +0000)]
r8198: - handled push/pull of simple strings in ejs
- improved the error handling, so the ejs wrappers don't just ignore a type they
don't handle, instead an exception is issued saying what isn't handled
Andrew Tridgell [Thu, 7 Jul 2005 07:47:30 +0000 (07:47 +0000)]
r8197: added testing of echo_SinkData() and echo_SourceData()
Andrew Tridgell [Thu, 7 Jul 2005 07:43:45 +0000 (07:43 +0000)]
r8196: - added testing of the EchoData interface in the echo test script
- add asserts on all calls for correct results
Andrew Tridgell [Thu, 7 Jul 2005 07:42:47 +0000 (07:42 +0000)]
r8195: - fixed handling of simple arrays. To keep the logic simple, I moved to making all push
functions taking a constant pointer to the type rather than having a different calling
convention for scalars and pointers
- fixed the setting of the 'length' element in arrays
Andrew Tridgell [Thu, 7 Jul 2005 06:27:23 +0000 (06:27 +0000)]
r8194: delete the old hand-written ejs code for echo_AddOne. This is now
autogenerated by pidl
Andrew Tridgell [Thu, 7 Jul 2005 06:26:22 +0000 (06:26 +0000)]
r8193: fixed the echo.js example code to work with the new syntax for rpc
calls from js. I chose the new syntax to match the C calling syntax,
so if you are familiar with using the Samba4 rpc libraries from C,
then using them from js should be easy
Andrew Tridgell [Thu, 7 Jul 2005 06:25:04 +0000 (06:25 +0000)]
r8192: updated the glue code for the generated ejs functions from pidl
only handles a small subset of all IDL files so far
Andrew Tridgell [Thu, 7 Jul 2005 06:23:36 +0000 (06:23 +0000)]
r8191: updated the ejs code generator in pidl to generate enough code for
simple rpc calls to work. Still very rough, but its a start.
Volker Lendecke [Wed, 6 Jul 2005 14:56:45 +0000 (14:56 +0000)]
r8185: Delete on close on directories:
Creating a file in a directory with delete-on-close set returns
DELETE_PENDING, and trying to set the flag on a non-empty directory returns
DIRECTORY_NOT_EMPTY.
Volker
Volker Lendecke [Wed, 6 Jul 2005 13:24:38 +0000 (13:24 +0000)]
r8182: Little more testing delete-on-close: Check flag with qfileinfo.
Volker
Andrew Bartlett [Wed, 6 Jul 2005 12:33:52 +0000 (12:33 +0000)]
r8181: Allow host/foo.realm/realm@REALM requests, assuming that the realm
forms both differ only in case. We may need a better solution than
this later.
Andrew Bartlett
Volker Lendecke [Wed, 6 Jul 2005 09:50:31 +0000 (09:50 +0000)]
r8179: Delete-on-close is really a shared DB. Setting it on one connection(!) and
resetting it on another resets it for both.
Volker
Volker Lendecke [Wed, 6 Jul 2005 08:13:11 +0000 (08:13 +0000)]
r8177: More explorations.
The share mode db is actually checked on qpathinfo even before the
delete-on-close is executed.
Volker
Volker Lendecke [Wed, 6 Jul 2005 07:45:22 +0000 (07:45 +0000)]
r8176: Exploring the share mode database...
A delete-on-close deleted file is still around while open on another fd. But
only for findfirst, not for qpathinfo :-)
Volker
Jeremy Allison [Wed, 6 Jul 2005 03:13:17 +0000 (03:13 +0000)]
r8174: Check DOS error codes in torture chkpath test.
Jeremy.
Tim Potter [Wed, 6 Jul 2005 01:53:34 +0000 (01:53 +0000)]
r8171: According to Samba 3 and Ethereal, the winreg_OpenUnkown stuff is
actually a uint16 * without the [string] attribute, a la the the
system_name argument to samr_Connect().
Initialising the pointer to NULL is sufficient and we still pass the
RPC-WINREG test against win2k3.
Stefan Metzmacher [Tue, 5 Jul 2005 12:42:54 +0000 (12:42 +0000)]
r8168: after testing I saw that w2k3 uses unique pointers on the mgmt pipe
that also matches the spec at http://www.opengroup.org/onlinepubs/
9629399/apdyq.htm
metze
Stefan Metzmacher [Tue, 5 Jul 2005 12:09:33 +0000 (12:09 +0000)]
r8167: - use the same algorithm than w2k3 for 'unique' pointer values
- add a new 'sptr' (simple-full) pointer type to simulate what we need to support pipes
like epmapper and mgmt that uses 'ptr' full pointer in their spec
- I runned make test and test_w2k3.sh with this, all fine
does we have any other pipe using 'unique' pointer where we need 'ptr'?
btw: jelmer, what does 'ignore' pointers do? they are allowed in pidl but not implemented!
metze
Stefan Metzmacher [Tue, 5 Jul 2005 10:57:39 +0000 (10:57 +0000)]
r8164: - match the ordering w2k3 uses for the PAC_BUFFER:
LOGON_INFO
LOGON_NAME
SRV_CHECKSUM
KDC_CHECKSUM
- w2k3 also don't use the groupmembership array with rids
it uses the othersids array
metze
Stefan Metzmacher [Tue, 5 Jul 2005 10:52:26 +0000 (10:52 +0000)]
r8163: if sidcount is zero it happened that we return NT_STATUS_NO_MEMORY...
metze
Andrew Bartlett [Tue, 5 Jul 2005 10:06:15 +0000 (10:06 +0000)]
r8162: Revert my pad8 hack.
Andrew Bartlett
Andrew Bartlett [Tue, 5 Jul 2005 10:05:40 +0000 (10:05 +0000)]
r8161: Update Samba4 for the new Heimdal update.
Andrew Bartlett
Stefan Metzmacher [Tue, 5 Jul 2005 07:10:52 +0000 (07:10 +0000)]
r8158: - use the timestring for the serial number of the bind zone file
- add --krbtgtpass and --machinepass options, with them you can easy set them to default
values for testing so that you don't need to setup a new keytab file when you rerun provision.pl
metze
Stefan Metzmacher [Tue, 5 Jul 2005 06:17:47 +0000 (06:17 +0000)]
r8157: add the algorithm for unique pointers that w2k3 uses.
this is ifdef'ed out currently because we use 'unique' pointers in the epmapper pipe,
where we should use 'ptr' full pointers.
metze
Stefan Metzmacher [Tue, 5 Jul 2005 06:13:12 +0000 (06:13 +0000)]
r8156: I found out that the unknown[2] field of the unknown[4] array is a length too,
it's always 16 bytes smaller than the size in the PAC_BUFFER
we now dump the blob's on LOCAL-PAC with -d 10
metze
Stefan Metzmacher [Mon, 4 Jul 2005 17:04:36 +0000 (17:04 +0000)]
r8154: - fix some mem_leals
- check if the buffer length of the original and created buffer are equal
metze
Stefan Metzmacher [Mon, 4 Jul 2005 16:16:22 +0000 (16:16 +0000)]
r8150: these should be static
metze
Stefan Metzmacher [Mon, 4 Jul 2005 15:42:08 +0000 (15:42 +0000)]
r8148: - make the PAC generation code a bit more readable and add some outof memory checks
- move to handmodified pull/push code for PAC_BUFFER
to get the _ndr_size field and the subcontext size right
- after looking closely to the sample w2k3 PAC in our torture test (and some more in my archive)
I found out that the first uint32 before the netr_SamInfo3 was also a pointer,
(and we passed a NULL pointer there before, so I think that was the reason why the windows clients doesn't want our PAC)
w2k3 uses this for unique pointers:
ptr = ndr->ptr_count * 4;
ptr |= 0x00020000;
ndr->ptr_count;
- do one more pull/push round with the sample PAC
metze
Stefan Metzmacher [Mon, 4 Jul 2005 15:06:05 +0000 (15:06 +0000)]
r8146: fix compiler warning
metze
Stefan Metzmacher [Mon, 4 Jul 2005 10:38:35 +0000 (10:38 +0000)]
r8136: remove unused var
metze
Stefan Metzmacher [Mon, 4 Jul 2005 10:37:05 +0000 (10:37 +0000)]
r8135: fix the linking on my SuSE 7.3 box
metze
Stefan Metzmacher [Mon, 4 Jul 2005 10:26:25 +0000 (10:26 +0000)]
r8134: remove unused var
metze
Rafal Szczesniak [Mon, 4 Jul 2005 07:32:48 +0000 (07:32 +0000)]
r8128: Janitor work...
rafal
Andrew Tridgell [Mon, 4 Jul 2005 07:26:27 +0000 (07:26 +0000)]
r8127: fixed code in function error
Andrew Tridgell [Mon, 4 Jul 2005 07:10:21 +0000 (07:10 +0000)]
r8126: - moved to 16 byte alignment for talloc. This is in response to a bug
report from robert collins.
- updated talloc guide to reflect the fact that over the last few
months talloc overhead compared to malloc has dropped, probably due
to a bunch of small changes. It now costs about 4% more than malloc
on my box
Andrew Tridgell [Mon, 4 Jul 2005 05:55:32 +0000 (05:55 +0000)]
r8125: fixed an error code mapping based on the updated torture tests
Andrew Tridgell [Mon, 4 Jul 2005 05:54:41 +0000 (05:54 +0000)]
r8124: added a set of file sharing tests that pass against w2k3
Andrew Tridgell [Mon, 4 Jul 2005 05:54:09 +0000 (05:54 +0000)]
r8123: fixed the RAW-NOTIFY and RAW-QFSINFO tests against w2k3
Andrew Tridgell [Mon, 4 Jul 2005 05:24:39 +0000 (05:24 +0000)]
r8122: more fixes from testing dos error code handling against w2k3
Andrew Tridgell [Mon, 4 Jul 2005 05:16:57 +0000 (05:16 +0000)]
r8121: yuck. w2k3 seems to choose ERRDOS:ERRbaduid or
NT_STATUS_INVALID_HANDLE on a per call basis for a bad vuid. That
means it is doing checking for a valid vuid in each backend function,
rather than globally. I don't want to emulate that as it is way too
error prone, and could easily lead to a security hole, so instead
accept either error code in our test suite.
Andrew Tridgell [Mon, 4 Jul 2005 05:08:27 +0000 (05:08 +0000)]
r8120: added in the newly found DOS locking error codes into the pvfs backend
Andrew Tridgell [Mon, 4 Jul 2005 05:07:18 +0000 (05:07 +0000)]
r8119: fixed two error code returns in the smb server now that we have
torture code that can tell the difference between dos and ntstatus
codes without mapping
Andrew Tridgell [Mon, 4 Jul 2005 05:06:16 +0000 (05:06 +0000)]
r8118: remove a debugging hack that should not have been in the last commit
Andrew Tridgell [Mon, 4 Jul 2005 05:05:28 +0000 (05:05 +0000)]
r8117: fixed a bunch more dos error code handing.
The biggest change was fixing the RAW-CONTEXT test. It was forcing
capabilities to zero in an attempt to not negotiated extended
security, but as a side effect it was forcing negotiation of dos error
codes. This confused the hell out of the test code!
Also fixed a bunch of places incorrectly using NT_STATUS_V() instead
of NT_STATUS_EQUAL() and several places that had the wrong dos status
codes
Andrew Tridgell [Mon, 4 Jul 2005 05:02:29 +0000 (05:02 +0000)]
r8116: demonstrate a little trick that can be used to track down where an
error is coming from. In this case I needed to know where a
NT_STATUS_ACCESS_DENIED was being returned, which is a very common
error, but I needed to know which place in the code was giving it
Andrew Tridgell [Mon, 4 Jul 2005 05:01:22 +0000 (05:01 +0000)]
r8115: added support for 2 more dos error codes found during testing
Andrew Tridgell [Mon, 4 Jul 2005 04:20:34 +0000 (04:20 +0000)]
r8114: fixed the build after tpots ejs commit ....
Andrew Tridgell [Mon, 4 Jul 2005 03:42:28 +0000 (03:42 +0000)]
r8113: this should fix the build on systems without heimdal
Andrew Bartlett [Mon, 4 Jul 2005 03:34:56 +0000 (03:34 +0000)]
r8112: Remove extra headers, and add #ifdef to allow the 'not yet using
Heimdal' case.
Andrew Bartlett
Andrew Tridgell [Mon, 4 Jul 2005 02:54:32 +0000 (02:54 +0000)]
r8111: fixed the client library to work against w2k3 with nt status codes
disabled. The main change is to turn off spnego, which cannot work at
all without nt status codes (w2k3 gives a ERRHRD:ERRgeneral error when
you try)
I also modified NT_STATUS_EQUAL() to allow for nt->dos code equality,
but only when nt status codes are disabled in smb.conf. That keeps all
the existing torture code working, while still allowing us to
correctly catch the cases where forced dos error codes are needed
The dos->ntstatus mapping table has been removed completely, as it
doesn't really make sense, is impossible to get right, and with the
new dos status handling isn't needed. When matching a nt status code
to a dos status code it makes far more sense to map from the nt code
to the dos code and compare, rather than the reverse, as the nt->dos
mapping is what windows has to do internally, so there really is a
valid mapping table.
Andrew Bartlett [Mon, 4 Jul 2005 02:36:16 +0000 (02:36 +0000)]
r8110: More PAC work. I still can't get WinXP to accept the PAC, but we are
much closer.
This changes PIDL to allow a subcontext to have a pad8 flag, saying to
pad behind to an 8 byte boundary. This is the only way I can explain
the 4 trainling zeros in the signature struct.
Far more importantly, the PAC code is now under self-test, both in
creating/parsing our own PAC, but also a PAC from my win2k3 server.
This required changing auth_anonymous, because I wanted to reuse the
anonymous 'server_info' generation code.
I'm still having trouble with PIDL, particulary as surrounds value(),
but I'll follow up on the list.
Andrew Bartlett
Andrew Bartlett [Mon, 4 Jul 2005 02:27:18 +0000 (02:27 +0000)]
r8109: Try to print out more helpful debug messages on DCERPC server-side
gensec failure to start.
Andrew Bartlett
Andrew Bartlett [Mon, 4 Jul 2005 02:25:33 +0000 (02:25 +0000)]
r8108: Fix indentation, and remove a discard_const_p() that we don't need any more.
Perhaps like minimal_includes.pl, we should have a way to find these
automaticly. Hmm...
Andrew Bartlett
Andrew Tridgell [Mon, 4 Jul 2005 01:57:53 +0000 (01:57 +0000)]
r8107: now that we properly separate DOS and NT status codes all the places
that relied on the mapping need to be fixed. The first thing is to get
all the torture tests working against w2k3 again with nt status codes
enabled. The 2nd step will be to make them pass with nt status
disabled.
This starts on the first task, fixing the assumption that
NT_STATUS_INVALID_LOCK_SEQUENCE is a valid substitute for
ERRDOS:ERRbadaccess
Andrew Tridgell [Mon, 4 Jul 2005 01:45:52 +0000 (01:45 +0000)]
r8106: the use of a static string for dos error codes was causing problems in
the torture code. To fix this, get rid of dos_errstr() and instead
move the strings into the nt_errstr() table, using cpp to generate the
strings
Tim Potter [Mon, 4 Jul 2005 01:31:34 +0000 (01:31 +0000)]
r8105: Add ejsrpc push/pull functions for uint16 and uint8.
Andrew Tridgell [Mon, 4 Jul 2005 01:23:38 +0000 (01:23 +0000)]
r8104: - added support for our client library to not negotiate nt status codes, controlled
with 'nt status support' option.
- make nt_errstr() display nice strings for dos status codes encoded
using NT_STATUS_DOS()
- no longer map between dos and nt status codes in the client library,
instead return using NT_STATUS_DOS()
- fixed the RAW-CONTEXT test to look for
NT_STATUS_DOS(ERRSRV, ERRbaduid) instead of NT_STATUS_INVALID_HANDLE
Tim Potter [Mon, 4 Jul 2005 01:20:02 +0000 (01:20 +0000)]
r8103: Add an argument separator between the hardcoded args and $PIDL_EXTRA_ARGS
so the first extra arg isn't interpreted as a parameter to the last
hardcoded arg.
Tim Potter [Mon, 4 Jul 2005 00:30:06 +0000 (00:30 +0000)]
r8101: Fix some indentation.
Volker Lendecke [Sun, 3 Jul 2005 17:35:19 +0000 (17:35 +0000)]
r8099: clean up after BASE-TCON
Rafal Szczesniak [Sun, 3 Jul 2005 17:17:32 +0000 (17:17 +0000)]
r8098: Add my copyright and remove unecessary header dependency.
rafal
Rafal Szczesniak [Sun, 3 Jul 2005 14:47:19 +0000 (14:47 +0000)]
r8096: Remove function that has became libnet_Lookup fuction.
rafal
Rafal Szczesniak [Sun, 3 Jul 2005 14:03:41 +0000 (14:03 +0000)]
r8095: Fix compiler warning.
rafal
Rafal Szczesniak [Sun, 3 Jul 2005 13:58:47 +0000 (13:58 +0000)]
r8094: Fix compiler warnings.
rafal
Simo Sorce [Sun, 3 Jul 2005 08:49:38 +0000 (08:49 +0000)]
r8091: fix invalid attribute names s/_/-/g
this will fix the build
Simo Sorce [Sat, 2 Jul 2005 18:43:22 +0000 (18:43 +0000)]
r8084: do not leak memory on errors
Simo Sorce [Sat, 2 Jul 2005 18:34:13 +0000 (18:34 +0000)]
r8083: check attribute type is valid (only ascii alphanum chars and '-' char)
fail if not