1
0
mirror of https://github.com/samba-team/samba.git synced 2025-12-24 04:23:53 +03:00
Commit Graph

71 Commits

Author SHA1 Message Date
Stefan Metzmacher
d7970d7032 pidl/NDR: fix handling of multilevel pointers in function elements
The 2nd or higher level of wire pointers needs to be marked as deferred.

metze
2008-02-07 08:33:18 +01:00
Stefan Metzmacher
aa8518521b pidl: get the pointer types correct when an element has multiple pointers
Only the first level gets the pointer type from the
pointer property, the others get them from
the pointer_default() interface property

see http://msdn2.microsoft.com/en-us/library/aa378984(VS.85).aspx
(Here they talk about the rightmost pointer, but testing shows
they mean the leftmost pointer.)

metze
2008-02-07 08:33:18 +01:00
Stefan Metzmacher
145d6c8ea0 pidl: remove 'pointer_default_top()' support
metze
2008-02-04 18:37:09 +01:00
Stefan Metzmacher
0bcc8e53d1 pidl: revert changes it didn't want to push...sorry!
8ebf16c074 Works!!!...pidl/Samba4::NDR::Parser: fix support for embedded "ref" pointers
6fcf2456d0 WORKS!!!...pidl/NDR: fix handling of multilevel pointers in function elements
0569139ca2 LOOKS OK... pidl: get the pointer types correct when an element has multiple pointe
rs
13afc89a87 CHECKED... pidl/Samba4::NDR::Parser: correctly get the name of an array element
29c104944b CHECKED... TODO:MSG pidl/Samba4::NDR::Parser: fix ...
3369015f5d CHECKED... pidl/Samba4::NDR::Parser: move logic for extra get_pointer_of() into a f
unction

metze
2008-02-02 11:13:03 +01:00
Stefan Metzmacher
6fcf2456d0 WORKS!!!...pidl/NDR: fix handling of multilevel pointers in function elements
The 2nd or higher level of wire pointers needs to be marked as deferred.

metze
2008-02-01 23:15:16 +01:00
Stefan Metzmacher
0569139ca2 LOOKS OK... pidl: get the pointer types correct when an element has multiple pointers
Only the first level gets the pointer type from the
pointer property, the others get them from
the pointer_default() interface property

see http://msdn2.microsoft.com/en-us/library/aa378984(VS.85).aspx
(Here they talk about the rightmost pointer, but testing shows
they mean the leftmost pointer.)

metze
2008-02-01 22:36:05 +01:00
Stefan Metzmacher
3c19198143 pidl: cosmetic fix to use the same value everywhere
metze
2008-02-01 18:20:32 +01:00
Stefan Metzmacher
fdf9bcb163 pidl/IDL: don't strip ',' from the properties content
metze
2008-02-01 07:09:38 +01:00
Jelmer Vernooij
b00c1a0724 pidl/python: Fix parsing arguments, fix more pointer issues. 2008-01-14 19:53:06 +01:00
Jelmer Vernooij
893f4102c9 pidl: Add function for determining whether a type has a body. 2008-01-12 23:10:28 +01:00
Jelmer Vernooij
254bf85c2e pidl: Remove declare tests, add more tests for typedef. 2008-01-12 21:21:14 +01:00
Jelmer Vernooij
2fe013f715 pidl: Remove support for the declare keyword in other parts of the pidl code. 2008-01-12 01:22:23 +01:00
Jelmer Vernooij
9e804e0c21 r25185: Check that can_contain_deferred returns true if one of the members of a type can contain deferred data. 2007-10-10 15:06:47 -05:00
Jelmer Vernooij
1afc7dd4d3 r25166: Simplify can_contain_deferred and add tests for it. 2007-10-10 15:06:45 -05:00
Stefan Metzmacher
3e0f22c857 r25114: handle structs and unions without typedefs better
in can_contain_deffered()

jelmer: we need to handle this more genericly, I assume
        we have a lot of other related bugs

metze
2007-10-10 15:06:42 -05:00
Jelmer Vernooij
30c1de30bb r24815: Support cpp_quote(). 2007-10-10 15:03:14 -05:00
Stefan Metzmacher
b7d6ac5215 r24514: IS_* elemements sound be perl integers consistantly...
metze
2007-10-10 15:02:06 -05:00
Stefan Metzmacher
5f1f50cd27 r24493: - it turns out that
foreach my $e (@{$union->{ELEMENTS}}) {
  changes $union->{ELEMENTS} from undef into an empty array.
  this removes the difference between
  struct foo { }; and struct foo;
  So we need to explicit return before.
- we should return the same element for layout for
  structs and unions with no elements.
- fix the testsuite to match

metze
2007-10-10 15:02:01 -05:00
Stefan Metzmacher
47fb2d42df r24488: enum's and bitmap's not have nested types
metze
2007-10-10 15:01:59 -05:00
Stefan Metzmacher
8aacab6f6a r24452: bail out instead of only giving a warning on
incorrect idl files

metze
2007-10-10 15:01:53 -05:00
Stefan Metzmacher
5abbb4673a r22617: assume we have defered elements if the type is just provides via
'declare' this should fix the problem with the winbind IRPC calls

metze
2007-10-10 14:51:52 -05:00
Jelmer Vernooij
74a3681d65 r21682: Remove accidently committed debug command. 2007-10-10 14:49:10 -05:00
Jelmer Vernooij
d65b520f08 r21681: Fix bug in the parsing code that parsed "struct foo;" the same as
"struct foo {};".

Reported by one of the OpenChange folks, thanks!
2007-10-10 14:49:10 -05:00
Jelmer Vernooij
3ac68e858d r21584: Support for tagged types has landed!
It's now possible to use "struct foo" without a typedef in IDL files.

echo_info4 is the first type that's been converted.
2007-10-10 14:49:00 -05:00
Jelmer Vernooij
6a4033464b r21574: Fix handling of DECLARE. 2007-10-10 14:48:58 -05:00
Jelmer Vernooij
bbbfbfa870 r21573: Remove more code that assumed all types are typedefs. 2007-10-10 14:48:58 -05:00
Jelmer Vernooij
4d28396f09 r21572: More work towards supporting tagged types. 2007-10-10 14:48:58 -05:00
Jelmer Vernooij
ec3c9ebfd0 r21484: Fix Needed* for nested datastructures. 2007-10-10 14:48:41 -05:00
Jelmer Vernooij
9b0416b5d0 r21457: Cope with anonymous nested types in the NDR layer. This doesn't handled
named nested types yet, as these have to be registered.
2007-10-10 14:48:38 -05:00
Jelmer Vernooij
a716aa70f0 r21430: Support tagged types without typedef. This means:
struct foo {
 ...
};

in IDL will now work. This is the first step towards nested types and
using typedefs for partial types (such as "typedef int *bar;"), a requirement
for complex uses of represent_as().
2007-10-10 14:48:32 -05:00
Jelmer Vernooij
34517c69e6 r21428: Handle representation types in Needed(). 2007-10-10 14:48:32 -05:00
Jelmer Vernooij
37c0da541e r21222: Merge a couple of pidl fixes:
* Pidl will now warn when trying to use pointers as integers in expressions.
* "subcontext()" is now marked as deprecated. The alternatives,
  transmit_as() / represent_as() should be available soon.
* More tests.
* Remove some unused code in smbtorture.
2007-10-10 14:44:48 -05:00
Jelmer Vernooij
ded25eca70 r20631: Add some tests for the ndr parser. 2007-10-10 14:37:18 -05:00
Jelmer Vernooij
959adfd0a6 r20511: Combine warnings/errors/fatal functions and move them to Parse::Pidl. 2007-10-10 14:36:04 -05:00
Jelmer Vernooij
006ab1d4a4 r19856: Use sptr as basis for full ptr implementation. Will add checks for duplicates later. 2007-10-10 14:28:26 -05:00
Stefan Metzmacher
f5aa620b1b r19854: readd support for 'sptr' pointers, to work against windows servers
until jelmer commits his 'ptr' support

metze
2007-10-10 14:28:25 -05:00
Jelmer Vernooij
3f0da2388d r19845: Add warning for pointer_default_top() 2007-10-10 14:28:24 -05:00
Jelmer Vernooij
a1b6308b21 r19842: Complain about unknown pointer types.
Fallback to "unique" when "ptr" is specified instead of failing.
2007-10-10 14:28:23 -05:00
Jelmer Vernooij
21efd2bdb8 r19834: Prevent some inaccurate warnings. 2007-10-10 14:28:22 -05:00
Jelmer Vernooij
78153200ac r19830: Warn about non-ref top-level pointers. 2007-10-10 14:28:21 -05:00
Jelmer Vernooij
324395afc7 r19752: Remove support for the `depends' attribute (use "import") instead. 2007-10-10 14:28:19 -05:00
Jelmer Vernooij
1ccab71cb8 r19585: Add support for some more standard IDL instructions:
- `include' (replaces helper())
 - `import' (replaces depends())

Add support for parsing importlib() - importlib() is now ignored (with a
warning), but no longer causes syntax errors.

helper() and depends() are now marked deprecated and will cause warnings.
2007-10-10 14:24:57 -05:00
Jelmer Vernooij
4f1f541837 r19544: Fix case of 'warning'. 2007-10-10 14:24:47 -05:00
Jelmer Vernooij
d1364ef0cd r18639: Get rid of the keepref support 2007-10-10 14:18:59 -05:00
Jelmer Vernooij
86b4624226 r18559: [string] always applies to the last pointer 2007-10-10 14:18:46 -05:00
Jelmer Vernooij
caac6cf817 r18554: Fix warnings about [out] arguments. 2007-10-10 14:18:45 -05:00
Jelmer Vernooij
f67b4d58ac r18523: Allow [out] on arrays as well as pointers, use in unixinfo. 2007-10-10 14:18:42 -05:00
Jelmer Vernooij
bdc6dd3750 r15776: Don't generate ref pointers in Samba4-generated code. There is no point
in having pointers for outgoing data when you can already modify the top-level
element.

This can be overridden (temporarily) by specifying the new "keepref"
attribute. Once we've removed keepref from all IDL files, I'll remove this
attribute as well.
2007-10-10 14:08:18 -05:00
Jelmer Vernooij
3e89ef0875 r15652: Fix aliases mechanism (fixes #3710) 2007-10-10 14:08:03 -05:00
Jelmer Vernooij
c4ab021ee8 r15593: Warn about [out] arguments that are not pointers. These can all be
fixed by adding [ref] pointers.

This will cause a lot of warnings to be outputted by pidl for now. I will
fix these gradually over the next few days.

We need to avoid [out] arguments that are not pointers because they are
not understood by other IDL compilers and don't work with some of
our output modules (Samba3, Samba3NDR and ethereal)
2007-10-10 14:06:02 -05:00