keyval: New keyval_parse()
keyval_parse() parses KEY=VALUE,... into a QDict. Works like
qemu_opts_parse(), except:
* Returns a QDict instead of a QemuOpts (d'oh).
* Supports nesting, unlike QemuOpts: a KEY is split into key
fragments at '.' (dotted key convention; the block layer does
something similar on top of QemuOpts). The key fragments are QDict
keys, and the last one's value is updated to VALUE.
* Each key fragment may be up to 127 bytes long. qemu_opts_parse()
limits the entire key to 127 bytes.
* Overlong key fragments are rejected. qemu_opts_parse() silently
truncates them.
* Empty key fragments are rejected. qemu_opts_parse() happily
accepts empty keys.
* It does not store the returned value. qemu_opts_parse() stores it
in the QemuOptsList.
* It does not treat parameter "id" specially. qemu_opts_parse()
ignores all but the first "id", and fails when its value isn't
id_wellformed(), or duplicate (a QemuOpts with the same ID is
already stored). It also screws up when a value contains ",id=".
* Implied value is not supported. qemu_opts_parse() desugars "foo" to
"foo=on", and "nofoo" to "foo=off".
* An implied key's value can't be empty, and can't contain ','.
I intend to grow this into a saner replacement for QemuOpts. It'll
take time, though.
Note: keyval_parse() provides no way to do lists, and its key syntax
is incompatible with the __RFQDN_ prefix convention for downstream
extensions, because it blindly splits at '.', even in __RFQDN_. Both
issues will be addressed later in the series.
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <1488317230-26248-4-git-send-email-armbru@redhat.com>
2017-03-01 00:26:49 +03:00
|
|
|
/*
|
|
|
|
* Parsing KEY=VALUE,... strings
|
|
|
|
*
|
|
|
|
* Copyright (C) 2017 Red Hat Inc.
|
|
|
|
*
|
|
|
|
* Authors:
|
|
|
|
* Markus Armbruster <armbru@redhat.com>,
|
|
|
|
*
|
|
|
|
* This work is licensed under the terms of the GNU GPL, version 2 or later.
|
|
|
|
* See the COPYING file in the top-level directory.
|
|
|
|
*/
|
|
|
|
|
|
|
|
/*
|
|
|
|
* KEY=VALUE,... syntax:
|
|
|
|
*
|
|
|
|
* key-vals = [ key-val { ',' key-val } [ ',' ] ]
|
|
|
|
* key-val = key '=' val
|
|
|
|
* key = key-fragment { '.' key-fragment }
|
|
|
|
* key-fragment = / [^=,.]* /
|
|
|
|
* val = { / [^,]* / | ',,' }
|
|
|
|
*
|
|
|
|
* Semantics defined by reduction to JSON:
|
|
|
|
*
|
2017-03-20 15:55:45 +03:00
|
|
|
* key-vals specifies a JSON object, i.e. a tree whose root is an
|
|
|
|
* object, inner nodes other than the root are objects or arrays,
|
|
|
|
* and leaves are strings.
|
keyval: New keyval_parse()
keyval_parse() parses KEY=VALUE,... into a QDict. Works like
qemu_opts_parse(), except:
* Returns a QDict instead of a QemuOpts (d'oh).
* Supports nesting, unlike QemuOpts: a KEY is split into key
fragments at '.' (dotted key convention; the block layer does
something similar on top of QemuOpts). The key fragments are QDict
keys, and the last one's value is updated to VALUE.
* Each key fragment may be up to 127 bytes long. qemu_opts_parse()
limits the entire key to 127 bytes.
* Overlong key fragments are rejected. qemu_opts_parse() silently
truncates them.
* Empty key fragments are rejected. qemu_opts_parse() happily
accepts empty keys.
* It does not store the returned value. qemu_opts_parse() stores it
in the QemuOptsList.
* It does not treat parameter "id" specially. qemu_opts_parse()
ignores all but the first "id", and fails when its value isn't
id_wellformed(), or duplicate (a QemuOpts with the same ID is
already stored). It also screws up when a value contains ",id=".
* Implied value is not supported. qemu_opts_parse() desugars "foo" to
"foo=on", and "nofoo" to "foo=off".
* An implied key's value can't be empty, and can't contain ','.
I intend to grow this into a saner replacement for QemuOpts. It'll
take time, though.
Note: keyval_parse() provides no way to do lists, and its key syntax
is incompatible with the __RFQDN_ prefix convention for downstream
extensions, because it blindly splits at '.', even in __RFQDN_. Both
issues will be addressed later in the series.
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <1488317230-26248-4-git-send-email-armbru@redhat.com>
2017-03-01 00:26:49 +03:00
|
|
|
*
|
2017-03-20 15:55:45 +03:00
|
|
|
* Each key-val = key-fragment '.' ... '=' val specifies a path from
|
|
|
|
* root to a leaf (left of '='), and the leaf's value (right of
|
|
|
|
* '=').
|
keyval: New keyval_parse()
keyval_parse() parses KEY=VALUE,... into a QDict. Works like
qemu_opts_parse(), except:
* Returns a QDict instead of a QemuOpts (d'oh).
* Supports nesting, unlike QemuOpts: a KEY is split into key
fragments at '.' (dotted key convention; the block layer does
something similar on top of QemuOpts). The key fragments are QDict
keys, and the last one's value is updated to VALUE.
* Each key fragment may be up to 127 bytes long. qemu_opts_parse()
limits the entire key to 127 bytes.
* Overlong key fragments are rejected. qemu_opts_parse() silently
truncates them.
* Empty key fragments are rejected. qemu_opts_parse() happily
accepts empty keys.
* It does not store the returned value. qemu_opts_parse() stores it
in the QemuOptsList.
* It does not treat parameter "id" specially. qemu_opts_parse()
ignores all but the first "id", and fails when its value isn't
id_wellformed(), or duplicate (a QemuOpts with the same ID is
already stored). It also screws up when a value contains ",id=".
* Implied value is not supported. qemu_opts_parse() desugars "foo" to
"foo=on", and "nofoo" to "foo=off".
* An implied key's value can't be empty, and can't contain ','.
I intend to grow this into a saner replacement for QemuOpts. It'll
take time, though.
Note: keyval_parse() provides no way to do lists, and its key syntax
is incompatible with the __RFQDN_ prefix convention for downstream
extensions, because it blindly splits at '.', even in __RFQDN_. Both
issues will be addressed later in the series.
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <1488317230-26248-4-git-send-email-armbru@redhat.com>
2017-03-01 00:26:49 +03:00
|
|
|
*
|
2017-03-20 15:55:45 +03:00
|
|
|
* A path from the root is defined recursively:
|
|
|
|
* L '.' key-fragment is a child of the node denoted by path L
|
|
|
|
* key-fragment is a child of the tree root
|
|
|
|
* If key-fragment is numeric, the parent is an array and the child
|
|
|
|
* is its key-fragment-th member, counting from zero.
|
|
|
|
* Else, the parent is an object, and the child is its member named
|
|
|
|
* key-fragment.
|
keyval: New keyval_parse()
keyval_parse() parses KEY=VALUE,... into a QDict. Works like
qemu_opts_parse(), except:
* Returns a QDict instead of a QemuOpts (d'oh).
* Supports nesting, unlike QemuOpts: a KEY is split into key
fragments at '.' (dotted key convention; the block layer does
something similar on top of QemuOpts). The key fragments are QDict
keys, and the last one's value is updated to VALUE.
* Each key fragment may be up to 127 bytes long. qemu_opts_parse()
limits the entire key to 127 bytes.
* Overlong key fragments are rejected. qemu_opts_parse() silently
truncates them.
* Empty key fragments are rejected. qemu_opts_parse() happily
accepts empty keys.
* It does not store the returned value. qemu_opts_parse() stores it
in the QemuOptsList.
* It does not treat parameter "id" specially. qemu_opts_parse()
ignores all but the first "id", and fails when its value isn't
id_wellformed(), or duplicate (a QemuOpts with the same ID is
already stored). It also screws up when a value contains ",id=".
* Implied value is not supported. qemu_opts_parse() desugars "foo" to
"foo=on", and "nofoo" to "foo=off".
* An implied key's value can't be empty, and can't contain ','.
I intend to grow this into a saner replacement for QemuOpts. It'll
take time, though.
Note: keyval_parse() provides no way to do lists, and its key syntax
is incompatible with the __RFQDN_ prefix convention for downstream
extensions, because it blindly splits at '.', even in __RFQDN_. Both
issues will be addressed later in the series.
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <1488317230-26248-4-git-send-email-armbru@redhat.com>
2017-03-01 00:26:49 +03:00
|
|
|
*
|
2017-03-20 15:55:45 +03:00
|
|
|
* This constrains inner nodes to be either array or object. The
|
|
|
|
* constraints must be satisfiable. Counter-example: a.b=1,a=2 is
|
|
|
|
* not, because root.a must be an object to satisfy a.b=1 and a
|
|
|
|
* string to satisfy a=2.
|
|
|
|
*
|
|
|
|
* Array subscripts can occur in any order, but the set of
|
|
|
|
* subscripts must not have gaps. For instance, a.1=v is not okay,
|
|
|
|
* because root.a[0] is missing.
|
|
|
|
*
|
|
|
|
* If multiple key-val denote the same leaf, the last one determines
|
|
|
|
* the value.
|
|
|
|
*
|
|
|
|
* Key-fragments must be valid QAPI names or consist only of decimal
|
|
|
|
* digits.
|
2017-03-01 00:27:05 +03:00
|
|
|
*
|
keyval: New keyval_parse()
keyval_parse() parses KEY=VALUE,... into a QDict. Works like
qemu_opts_parse(), except:
* Returns a QDict instead of a QemuOpts (d'oh).
* Supports nesting, unlike QemuOpts: a KEY is split into key
fragments at '.' (dotted key convention; the block layer does
something similar on top of QemuOpts). The key fragments are QDict
keys, and the last one's value is updated to VALUE.
* Each key fragment may be up to 127 bytes long. qemu_opts_parse()
limits the entire key to 127 bytes.
* Overlong key fragments are rejected. qemu_opts_parse() silently
truncates them.
* Empty key fragments are rejected. qemu_opts_parse() happily
accepts empty keys.
* It does not store the returned value. qemu_opts_parse() stores it
in the QemuOptsList.
* It does not treat parameter "id" specially. qemu_opts_parse()
ignores all but the first "id", and fails when its value isn't
id_wellformed(), or duplicate (a QemuOpts with the same ID is
already stored). It also screws up when a value contains ",id=".
* Implied value is not supported. qemu_opts_parse() desugars "foo" to
"foo=on", and "nofoo" to "foo=off".
* An implied key's value can't be empty, and can't contain ','.
I intend to grow this into a saner replacement for QemuOpts. It'll
take time, though.
Note: keyval_parse() provides no way to do lists, and its key syntax
is incompatible with the __RFQDN_ prefix convention for downstream
extensions, because it blindly splits at '.', even in __RFQDN_. Both
issues will be addressed later in the series.
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <1488317230-26248-4-git-send-email-armbru@redhat.com>
2017-03-01 00:26:49 +03:00
|
|
|
* The length of any key-fragment must be between 1 and 127.
|
|
|
|
*
|
keyval: Support lists
Additionally permit non-negative integers as key components. A
dictionary's keys must either be all integers or none. If all keys
are integers, convert the dictionary to a list. The set of keys must
be [0,N].
Examples:
* list.1=goner,list.0=null,list.1=eins,list.2=zwei
is equivalent to JSON [ "null", "eins", "zwei" ]
* a.b.c=1,a.b.0=2
is inconsistent: a.b.c clashes with a.b.0
* list.0=null,list.2=eins,list.2=zwei
has a hole: list.1 is missing
Similar design flaw as for objects: there is no way to denote an empty
list. While interpreting "key absent" as empty list seems natural
(removing a list member from the input string works when there are
multiple ones, so why not when there's just one), it doesn't work:
"key absent" already means "optional list absent", which isn't the
same as "empty list present".
Update the keyval object visitor to use this a.0 syntax in error
messages rather than the usual a[0].
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <1488317230-26248-25-git-send-email-armbru@redhat.com>
[Off-by-one fix squashed in, as per Kevin's review]
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
2017-03-01 00:27:10 +03:00
|
|
|
* Design flaw: there is no way to denote an empty array or non-root
|
|
|
|
* object. While interpreting "key absent" as empty seems natural
|
keyval: New keyval_parse()
keyval_parse() parses KEY=VALUE,... into a QDict. Works like
qemu_opts_parse(), except:
* Returns a QDict instead of a QemuOpts (d'oh).
* Supports nesting, unlike QemuOpts: a KEY is split into key
fragments at '.' (dotted key convention; the block layer does
something similar on top of QemuOpts). The key fragments are QDict
keys, and the last one's value is updated to VALUE.
* Each key fragment may be up to 127 bytes long. qemu_opts_parse()
limits the entire key to 127 bytes.
* Overlong key fragments are rejected. qemu_opts_parse() silently
truncates them.
* Empty key fragments are rejected. qemu_opts_parse() happily
accepts empty keys.
* It does not store the returned value. qemu_opts_parse() stores it
in the QemuOptsList.
* It does not treat parameter "id" specially. qemu_opts_parse()
ignores all but the first "id", and fails when its value isn't
id_wellformed(), or duplicate (a QemuOpts with the same ID is
already stored). It also screws up when a value contains ",id=".
* Implied value is not supported. qemu_opts_parse() desugars "foo" to
"foo=on", and "nofoo" to "foo=off".
* An implied key's value can't be empty, and can't contain ','.
I intend to grow this into a saner replacement for QemuOpts. It'll
take time, though.
Note: keyval_parse() provides no way to do lists, and its key syntax
is incompatible with the __RFQDN_ prefix convention for downstream
extensions, because it blindly splits at '.', even in __RFQDN_. Both
issues will be addressed later in the series.
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <1488317230-26248-4-git-send-email-armbru@redhat.com>
2017-03-01 00:26:49 +03:00
|
|
|
* (removing a key-val from the input string removes the member when
|
|
|
|
* there are more, so why not when it's the last), it doesn't work:
|
keyval: Support lists
Additionally permit non-negative integers as key components. A
dictionary's keys must either be all integers or none. If all keys
are integers, convert the dictionary to a list. The set of keys must
be [0,N].
Examples:
* list.1=goner,list.0=null,list.1=eins,list.2=zwei
is equivalent to JSON [ "null", "eins", "zwei" ]
* a.b.c=1,a.b.0=2
is inconsistent: a.b.c clashes with a.b.0
* list.0=null,list.2=eins,list.2=zwei
has a hole: list.1 is missing
Similar design flaw as for objects: there is no way to denote an empty
list. While interpreting "key absent" as empty list seems natural
(removing a list member from the input string works when there are
multiple ones, so why not when there's just one), it doesn't work:
"key absent" already means "optional list absent", which isn't the
same as "empty list present".
Update the keyval object visitor to use this a.0 syntax in error
messages rather than the usual a[0].
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <1488317230-26248-25-git-send-email-armbru@redhat.com>
[Off-by-one fix squashed in, as per Kevin's review]
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
2017-03-01 00:27:10 +03:00
|
|
|
* "key absent" already means "optional object/array absent", which
|
|
|
|
* isn't the same as "empty object/array present".
|
keyval: New keyval_parse()
keyval_parse() parses KEY=VALUE,... into a QDict. Works like
qemu_opts_parse(), except:
* Returns a QDict instead of a QemuOpts (d'oh).
* Supports nesting, unlike QemuOpts: a KEY is split into key
fragments at '.' (dotted key convention; the block layer does
something similar on top of QemuOpts). The key fragments are QDict
keys, and the last one's value is updated to VALUE.
* Each key fragment may be up to 127 bytes long. qemu_opts_parse()
limits the entire key to 127 bytes.
* Overlong key fragments are rejected. qemu_opts_parse() silently
truncates them.
* Empty key fragments are rejected. qemu_opts_parse() happily
accepts empty keys.
* It does not store the returned value. qemu_opts_parse() stores it
in the QemuOptsList.
* It does not treat parameter "id" specially. qemu_opts_parse()
ignores all but the first "id", and fails when its value isn't
id_wellformed(), or duplicate (a QemuOpts with the same ID is
already stored). It also screws up when a value contains ",id=".
* Implied value is not supported. qemu_opts_parse() desugars "foo" to
"foo=on", and "nofoo" to "foo=off".
* An implied key's value can't be empty, and can't contain ','.
I intend to grow this into a saner replacement for QemuOpts. It'll
take time, though.
Note: keyval_parse() provides no way to do lists, and its key syntax
is incompatible with the __RFQDN_ prefix convention for downstream
extensions, because it blindly splits at '.', even in __RFQDN_. Both
issues will be addressed later in the series.
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <1488317230-26248-4-git-send-email-armbru@redhat.com>
2017-03-01 00:26:49 +03:00
|
|
|
*
|
2017-03-20 15:55:47 +03:00
|
|
|
* Design flaw: scalar values can only be strings; there is no way to
|
|
|
|
* denote numbers, true, false or null. The special QObject input
|
|
|
|
* visitor returned by qobject_input_visitor_new_keyval() mostly hides
|
|
|
|
* this by automatically converting strings to the type the visitor
|
qapi: Reject alternates that can't work with keyval_parse()
Alternates are sum types like unions, but use the JSON type on the
wire / QType in QObject instead of an explicit tag. That's why we
require alternate members to have distinct QTypes.
The recently introduced keyval_parse() (commit d454dbe) can only
produce string scalars. The qobject_input_visitor_new_keyval() input
visitor mostly hides the difference, so code using a QObject input
visitor doesn't have to care whether its input was parsed from JSON or
KEY=VALUE,... The difference leaks for alternates, as noted in commit
0ee9ae7: a non-string, non-enum scalar alternate value can't currently
be expressed.
In part, this is just our insufficiently sophisticated implementation.
Consider alternate type 'GuestFileWhence'. It has an integer member
and a 'QGASeek' member. The latter is an enumeration with values
'set', 'cur', 'end'. The meaning of b=set, b=cur, b=end, b=0, b=1 and
so forth is perfectly obvious. However, our current implementation
falls apart at run time for b=0, b=1, and so forth. Fixable, but not
today; add a test case and a TODO comment.
Now consider an alternate type with a string and an integer member.
What's the meaning of a=42? Is it the string "42" or the integer 42?
Whichever meaning you pick makes the other inexpressible. This isn't
just an implementation problem, it's fundamental. Our current
implementation will pick string.
So far, we haven't needed such alternates. To make sure we stop and
think before we add one that cannot sanely work with keyval_parse(),
let's require alternate members to have sufficiently distinct
representation in KEY=VALUE,... syntax:
* A string member clashes with any other scalar member
* An enumeration member clashes with bool members when it has value
'on' or 'off'.
* An enumeration member clashes with numeric members when it has a
value that starts with '-', '+', or a decimal digit. This is a
rather lazy approximation of the actual number syntax accepted by
the visitor.
Note that enumeration values starting with '-' and '+' are rejected
elsewhere already, but better safe than sorry.
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <1495471335-23707-5-git-send-email-armbru@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
2017-05-22 19:42:15 +03:00
|
|
|
* expects. Breaks down for type 'any', where the visitor's
|
|
|
|
* expectation isn't clear. Code visiting 'any' needs to do the
|
|
|
|
* conversion itself, but only when using this keyval visitor.
|
|
|
|
* Awkward. Note that we carefully restrict alternate types to avoid
|
|
|
|
* similar ambiguity.
|
2017-03-20 15:55:47 +03:00
|
|
|
*
|
keyval: New keyval_parse()
keyval_parse() parses KEY=VALUE,... into a QDict. Works like
qemu_opts_parse(), except:
* Returns a QDict instead of a QemuOpts (d'oh).
* Supports nesting, unlike QemuOpts: a KEY is split into key
fragments at '.' (dotted key convention; the block layer does
something similar on top of QemuOpts). The key fragments are QDict
keys, and the last one's value is updated to VALUE.
* Each key fragment may be up to 127 bytes long. qemu_opts_parse()
limits the entire key to 127 bytes.
* Overlong key fragments are rejected. qemu_opts_parse() silently
truncates them.
* Empty key fragments are rejected. qemu_opts_parse() happily
accepts empty keys.
* It does not store the returned value. qemu_opts_parse() stores it
in the QemuOptsList.
* It does not treat parameter "id" specially. qemu_opts_parse()
ignores all but the first "id", and fails when its value isn't
id_wellformed(), or duplicate (a QemuOpts with the same ID is
already stored). It also screws up when a value contains ",id=".
* Implied value is not supported. qemu_opts_parse() desugars "foo" to
"foo=on", and "nofoo" to "foo=off".
* An implied key's value can't be empty, and can't contain ','.
I intend to grow this into a saner replacement for QemuOpts. It'll
take time, though.
Note: keyval_parse() provides no way to do lists, and its key syntax
is incompatible with the __RFQDN_ prefix convention for downstream
extensions, because it blindly splits at '.', even in __RFQDN_. Both
issues will be addressed later in the series.
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <1488317230-26248-4-git-send-email-armbru@redhat.com>
2017-03-01 00:26:49 +03:00
|
|
|
* Additional syntax for use with an implied key:
|
|
|
|
*
|
|
|
|
* key-vals-ik = val-no-key [ ',' key-vals ]
|
|
|
|
* val-no-key = / [^=,]* /
|
|
|
|
*
|
|
|
|
* where no-key is syntactic sugar for implied-key=val-no-key.
|
|
|
|
*/
|
|
|
|
|
|
|
|
#include "qemu/osdep.h"
|
|
|
|
#include "qapi/error.h"
|
2018-02-01 14:18:39 +03:00
|
|
|
#include "qapi/qmp/qdict.h"
|
2018-02-01 14:18:38 +03:00
|
|
|
#include "qapi/qmp/qlist.h"
|
keyval: New keyval_parse()
keyval_parse() parses KEY=VALUE,... into a QDict. Works like
qemu_opts_parse(), except:
* Returns a QDict instead of a QemuOpts (d'oh).
* Supports nesting, unlike QemuOpts: a KEY is split into key
fragments at '.' (dotted key convention; the block layer does
something similar on top of QemuOpts). The key fragments are QDict
keys, and the last one's value is updated to VALUE.
* Each key fragment may be up to 127 bytes long. qemu_opts_parse()
limits the entire key to 127 bytes.
* Overlong key fragments are rejected. qemu_opts_parse() silently
truncates them.
* Empty key fragments are rejected. qemu_opts_parse() happily
accepts empty keys.
* It does not store the returned value. qemu_opts_parse() stores it
in the QemuOptsList.
* It does not treat parameter "id" specially. qemu_opts_parse()
ignores all but the first "id", and fails when its value isn't
id_wellformed(), or duplicate (a QemuOpts with the same ID is
already stored). It also screws up when a value contains ",id=".
* Implied value is not supported. qemu_opts_parse() desugars "foo" to
"foo=on", and "nofoo" to "foo=off".
* An implied key's value can't be empty, and can't contain ','.
I intend to grow this into a saner replacement for QemuOpts. It'll
take time, though.
Note: keyval_parse() provides no way to do lists, and its key syntax
is incompatible with the __RFQDN_ prefix convention for downstream
extensions, because it blindly splits at '.', even in __RFQDN_. Both
issues will be addressed later in the series.
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <1488317230-26248-4-git-send-email-armbru@redhat.com>
2017-03-01 00:26:49 +03:00
|
|
|
#include "qapi/qmp/qstring.h"
|
keyval: Support lists
Additionally permit non-negative integers as key components. A
dictionary's keys must either be all integers or none. If all keys
are integers, convert the dictionary to a list. The set of keys must
be [0,N].
Examples:
* list.1=goner,list.0=null,list.1=eins,list.2=zwei
is equivalent to JSON [ "null", "eins", "zwei" ]
* a.b.c=1,a.b.0=2
is inconsistent: a.b.c clashes with a.b.0
* list.0=null,list.2=eins,list.2=zwei
has a hole: list.1 is missing
Similar design flaw as for objects: there is no way to denote an empty
list. While interpreting "key absent" as empty list seems natural
(removing a list member from the input string works when there are
multiple ones, so why not when there's just one), it doesn't work:
"key absent" already means "optional list absent", which isn't the
same as "empty list present".
Update the keyval object visitor to use this a.0 syntax in error
messages rather than the usual a[0].
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <1488317230-26248-25-git-send-email-armbru@redhat.com>
[Off-by-one fix squashed in, as per Kevin's review]
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
2017-03-01 00:27:10 +03:00
|
|
|
#include "qemu/cutils.h"
|
keyval: New keyval_parse()
keyval_parse() parses KEY=VALUE,... into a QDict. Works like
qemu_opts_parse(), except:
* Returns a QDict instead of a QemuOpts (d'oh).
* Supports nesting, unlike QemuOpts: a KEY is split into key
fragments at '.' (dotted key convention; the block layer does
something similar on top of QemuOpts). The key fragments are QDict
keys, and the last one's value is updated to VALUE.
* Each key fragment may be up to 127 bytes long. qemu_opts_parse()
limits the entire key to 127 bytes.
* Overlong key fragments are rejected. qemu_opts_parse() silently
truncates them.
* Empty key fragments are rejected. qemu_opts_parse() happily
accepts empty keys.
* It does not store the returned value. qemu_opts_parse() stores it
in the QemuOptsList.
* It does not treat parameter "id" specially. qemu_opts_parse()
ignores all but the first "id", and fails when its value isn't
id_wellformed(), or duplicate (a QemuOpts with the same ID is
already stored). It also screws up when a value contains ",id=".
* Implied value is not supported. qemu_opts_parse() desugars "foo" to
"foo=on", and "nofoo" to "foo=off".
* An implied key's value can't be empty, and can't contain ','.
I intend to grow this into a saner replacement for QemuOpts. It'll
take time, though.
Note: keyval_parse() provides no way to do lists, and its key syntax
is incompatible with the __RFQDN_ prefix convention for downstream
extensions, because it blindly splits at '.', even in __RFQDN_. Both
issues will be addressed later in the series.
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <1488317230-26248-4-git-send-email-armbru@redhat.com>
2017-03-01 00:26:49 +03:00
|
|
|
#include "qemu/option.h"
|
|
|
|
|
keyval: Support lists
Additionally permit non-negative integers as key components. A
dictionary's keys must either be all integers or none. If all keys
are integers, convert the dictionary to a list. The set of keys must
be [0,N].
Examples:
* list.1=goner,list.0=null,list.1=eins,list.2=zwei
is equivalent to JSON [ "null", "eins", "zwei" ]
* a.b.c=1,a.b.0=2
is inconsistent: a.b.c clashes with a.b.0
* list.0=null,list.2=eins,list.2=zwei
has a hole: list.1 is missing
Similar design flaw as for objects: there is no way to denote an empty
list. While interpreting "key absent" as empty list seems natural
(removing a list member from the input string works when there are
multiple ones, so why not when there's just one), it doesn't work:
"key absent" already means "optional list absent", which isn't the
same as "empty list present".
Update the keyval object visitor to use this a.0 syntax in error
messages rather than the usual a[0].
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <1488317230-26248-25-git-send-email-armbru@redhat.com>
[Off-by-one fix squashed in, as per Kevin's review]
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
2017-03-01 00:27:10 +03:00
|
|
|
/*
|
|
|
|
* Convert @key to a list index.
|
2017-03-20 15:55:45 +03:00
|
|
|
* Convert all leading decimal digits to a (non-negative) number,
|
|
|
|
* capped at INT_MAX.
|
keyval: Support lists
Additionally permit non-negative integers as key components. A
dictionary's keys must either be all integers or none. If all keys
are integers, convert the dictionary to a list. The set of keys must
be [0,N].
Examples:
* list.1=goner,list.0=null,list.1=eins,list.2=zwei
is equivalent to JSON [ "null", "eins", "zwei" ]
* a.b.c=1,a.b.0=2
is inconsistent: a.b.c clashes with a.b.0
* list.0=null,list.2=eins,list.2=zwei
has a hole: list.1 is missing
Similar design flaw as for objects: there is no way to denote an empty
list. While interpreting "key absent" as empty list seems natural
(removing a list member from the input string works when there are
multiple ones, so why not when there's just one), it doesn't work:
"key absent" already means "optional list absent", which isn't the
same as "empty list present".
Update the keyval object visitor to use this a.0 syntax in error
messages rather than the usual a[0].
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <1488317230-26248-25-git-send-email-armbru@redhat.com>
[Off-by-one fix squashed in, as per Kevin's review]
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
2017-03-01 00:27:10 +03:00
|
|
|
* If @end is non-null, assign a pointer to the first character after
|
|
|
|
* the number to *@end.
|
|
|
|
* Else, fail if any characters follow.
|
|
|
|
* On success, return the converted number.
|
|
|
|
* On failure, return a negative value.
|
|
|
|
* Note: since only digits are converted, no two keys can map to the
|
|
|
|
* same number, except by overflow to INT_MAX.
|
|
|
|
*/
|
|
|
|
static int key_to_index(const char *key, const char **end)
|
|
|
|
{
|
|
|
|
int ret;
|
|
|
|
unsigned long index;
|
|
|
|
|
|
|
|
if (*key < '0' || *key > '9') {
|
|
|
|
return -EINVAL;
|
|
|
|
}
|
|
|
|
ret = qemu_strtoul(key, end, 10, &index);
|
|
|
|
if (ret) {
|
|
|
|
return ret == -ERANGE ? INT_MAX : ret;
|
|
|
|
}
|
|
|
|
return index <= INT_MAX ? index : INT_MAX;
|
|
|
|
}
|
|
|
|
|
keyval: New keyval_parse()
keyval_parse() parses KEY=VALUE,... into a QDict. Works like
qemu_opts_parse(), except:
* Returns a QDict instead of a QemuOpts (d'oh).
* Supports nesting, unlike QemuOpts: a KEY is split into key
fragments at '.' (dotted key convention; the block layer does
something similar on top of QemuOpts). The key fragments are QDict
keys, and the last one's value is updated to VALUE.
* Each key fragment may be up to 127 bytes long. qemu_opts_parse()
limits the entire key to 127 bytes.
* Overlong key fragments are rejected. qemu_opts_parse() silently
truncates them.
* Empty key fragments are rejected. qemu_opts_parse() happily
accepts empty keys.
* It does not store the returned value. qemu_opts_parse() stores it
in the QemuOptsList.
* It does not treat parameter "id" specially. qemu_opts_parse()
ignores all but the first "id", and fails when its value isn't
id_wellformed(), or duplicate (a QemuOpts with the same ID is
already stored). It also screws up when a value contains ",id=".
* Implied value is not supported. qemu_opts_parse() desugars "foo" to
"foo=on", and "nofoo" to "foo=off".
* An implied key's value can't be empty, and can't contain ','.
I intend to grow this into a saner replacement for QemuOpts. It'll
take time, though.
Note: keyval_parse() provides no way to do lists, and its key syntax
is incompatible with the __RFQDN_ prefix convention for downstream
extensions, because it blindly splits at '.', even in __RFQDN_. Both
issues will be addressed later in the series.
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <1488317230-26248-4-git-send-email-armbru@redhat.com>
2017-03-01 00:26:49 +03:00
|
|
|
/*
|
|
|
|
* Ensure @cur maps @key_in_cur the right way.
|
|
|
|
* If @value is null, it needs to map to a QDict, else to this
|
|
|
|
* QString.
|
|
|
|
* If @cur doesn't have @key_in_cur, put an empty QDict or @value,
|
|
|
|
* respectively.
|
|
|
|
* Else, if it needs to map to a QDict, and already does, do nothing.
|
|
|
|
* Else, if it needs to map to this QString, and already maps to a
|
|
|
|
* QString, replace it by @value.
|
|
|
|
* Else, fail because we have conflicting needs on how to map
|
|
|
|
* @key_in_cur.
|
|
|
|
* In any case, take over the reference to @value, i.e. if the caller
|
|
|
|
* wants to hold on to a reference, it needs to QINCREF().
|
|
|
|
* Use @key up to @key_cursor to identify the key in error messages.
|
|
|
|
* On success, return the mapped value.
|
|
|
|
* On failure, store an error through @errp and return NULL.
|
|
|
|
*/
|
|
|
|
static QObject *keyval_parse_put(QDict *cur,
|
|
|
|
const char *key_in_cur, QString *value,
|
|
|
|
const char *key, const char *key_cursor,
|
|
|
|
Error **errp)
|
|
|
|
{
|
|
|
|
QObject *old, *new;
|
|
|
|
|
|
|
|
old = qdict_get(cur, key_in_cur);
|
|
|
|
if (old) {
|
|
|
|
if (qobject_type(old) != (value ? QTYPE_QSTRING : QTYPE_QDICT)) {
|
|
|
|
error_setg(errp, "Parameters '%.*s.*' used inconsistently",
|
|
|
|
(int)(key_cursor - key), key);
|
|
|
|
QDECREF(value);
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
if (!value) {
|
|
|
|
return old; /* already QDict, do nothing */
|
|
|
|
}
|
|
|
|
new = QOBJECT(value); /* replacement */
|
|
|
|
} else {
|
|
|
|
new = value ? QOBJECT(value) : QOBJECT(qdict_new());
|
|
|
|
}
|
|
|
|
qdict_put_obj(cur, key_in_cur, new);
|
|
|
|
return new;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Parse one KEY=VALUE from @params, store result in @qdict.
|
|
|
|
* The first fragment of KEY applies to @qdict. Subsequent fragments
|
|
|
|
* apply to nested QDicts, which are created on demand. @implied_key
|
|
|
|
* is as in keyval_parse().
|
|
|
|
* On success, return a pointer to the next KEY=VALUE, or else to '\0'.
|
|
|
|
* On failure, return NULL.
|
|
|
|
*/
|
|
|
|
static const char *keyval_parse_one(QDict *qdict, const char *params,
|
|
|
|
const char *implied_key,
|
|
|
|
Error **errp)
|
|
|
|
{
|
keyval: Support lists
Additionally permit non-negative integers as key components. A
dictionary's keys must either be all integers or none. If all keys
are integers, convert the dictionary to a list. The set of keys must
be [0,N].
Examples:
* list.1=goner,list.0=null,list.1=eins,list.2=zwei
is equivalent to JSON [ "null", "eins", "zwei" ]
* a.b.c=1,a.b.0=2
is inconsistent: a.b.c clashes with a.b.0
* list.0=null,list.2=eins,list.2=zwei
has a hole: list.1 is missing
Similar design flaw as for objects: there is no way to denote an empty
list. While interpreting "key absent" as empty list seems natural
(removing a list member from the input string works when there are
multiple ones, so why not when there's just one), it doesn't work:
"key absent" already means "optional list absent", which isn't the
same as "empty list present".
Update the keyval object visitor to use this a.0 syntax in error
messages rather than the usual a[0].
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <1488317230-26248-25-git-send-email-armbru@redhat.com>
[Off-by-one fix squashed in, as per Kevin's review]
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
2017-03-01 00:27:10 +03:00
|
|
|
const char *key, *key_end, *s, *end;
|
keyval: New keyval_parse()
keyval_parse() parses KEY=VALUE,... into a QDict. Works like
qemu_opts_parse(), except:
* Returns a QDict instead of a QemuOpts (d'oh).
* Supports nesting, unlike QemuOpts: a KEY is split into key
fragments at '.' (dotted key convention; the block layer does
something similar on top of QemuOpts). The key fragments are QDict
keys, and the last one's value is updated to VALUE.
* Each key fragment may be up to 127 bytes long. qemu_opts_parse()
limits the entire key to 127 bytes.
* Overlong key fragments are rejected. qemu_opts_parse() silently
truncates them.
* Empty key fragments are rejected. qemu_opts_parse() happily
accepts empty keys.
* It does not store the returned value. qemu_opts_parse() stores it
in the QemuOptsList.
* It does not treat parameter "id" specially. qemu_opts_parse()
ignores all but the first "id", and fails when its value isn't
id_wellformed(), or duplicate (a QemuOpts with the same ID is
already stored). It also screws up when a value contains ",id=".
* Implied value is not supported. qemu_opts_parse() desugars "foo" to
"foo=on", and "nofoo" to "foo=off".
* An implied key's value can't be empty, and can't contain ','.
I intend to grow this into a saner replacement for QemuOpts. It'll
take time, though.
Note: keyval_parse() provides no way to do lists, and its key syntax
is incompatible with the __RFQDN_ prefix convention for downstream
extensions, because it blindly splits at '.', even in __RFQDN_. Both
issues will be addressed later in the series.
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <1488317230-26248-4-git-send-email-armbru@redhat.com>
2017-03-01 00:26:49 +03:00
|
|
|
size_t len;
|
|
|
|
char key_in_cur[128];
|
|
|
|
QDict *cur;
|
2017-03-01 00:27:05 +03:00
|
|
|
int ret;
|
keyval: New keyval_parse()
keyval_parse() parses KEY=VALUE,... into a QDict. Works like
qemu_opts_parse(), except:
* Returns a QDict instead of a QemuOpts (d'oh).
* Supports nesting, unlike QemuOpts: a KEY is split into key
fragments at '.' (dotted key convention; the block layer does
something similar on top of QemuOpts). The key fragments are QDict
keys, and the last one's value is updated to VALUE.
* Each key fragment may be up to 127 bytes long. qemu_opts_parse()
limits the entire key to 127 bytes.
* Overlong key fragments are rejected. qemu_opts_parse() silently
truncates them.
* Empty key fragments are rejected. qemu_opts_parse() happily
accepts empty keys.
* It does not store the returned value. qemu_opts_parse() stores it
in the QemuOptsList.
* It does not treat parameter "id" specially. qemu_opts_parse()
ignores all but the first "id", and fails when its value isn't
id_wellformed(), or duplicate (a QemuOpts with the same ID is
already stored). It also screws up when a value contains ",id=".
* Implied value is not supported. qemu_opts_parse() desugars "foo" to
"foo=on", and "nofoo" to "foo=off".
* An implied key's value can't be empty, and can't contain ','.
I intend to grow this into a saner replacement for QemuOpts. It'll
take time, though.
Note: keyval_parse() provides no way to do lists, and its key syntax
is incompatible with the __RFQDN_ prefix convention for downstream
extensions, because it blindly splits at '.', even in __RFQDN_. Both
issues will be addressed later in the series.
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <1488317230-26248-4-git-send-email-armbru@redhat.com>
2017-03-01 00:26:49 +03:00
|
|
|
QObject *next;
|
|
|
|
QString *val;
|
|
|
|
|
|
|
|
key = params;
|
|
|
|
len = strcspn(params, "=,");
|
|
|
|
if (implied_key && len && key[len] != '=') {
|
|
|
|
/* Desugar implied key */
|
|
|
|
key = implied_key;
|
|
|
|
len = strlen(implied_key);
|
|
|
|
}
|
|
|
|
key_end = key + len;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Loop over key fragments: @s points to current fragment, it
|
|
|
|
* applies to @cur. @key_in_cur[] holds the previous fragment.
|
|
|
|
*/
|
|
|
|
cur = qdict;
|
|
|
|
s = key;
|
|
|
|
for (;;) {
|
keyval: Support lists
Additionally permit non-negative integers as key components. A
dictionary's keys must either be all integers or none. If all keys
are integers, convert the dictionary to a list. The set of keys must
be [0,N].
Examples:
* list.1=goner,list.0=null,list.1=eins,list.2=zwei
is equivalent to JSON [ "null", "eins", "zwei" ]
* a.b.c=1,a.b.0=2
is inconsistent: a.b.c clashes with a.b.0
* list.0=null,list.2=eins,list.2=zwei
has a hole: list.1 is missing
Similar design flaw as for objects: there is no way to denote an empty
list. While interpreting "key absent" as empty list seems natural
(removing a list member from the input string works when there are
multiple ones, so why not when there's just one), it doesn't work:
"key absent" already means "optional list absent", which isn't the
same as "empty list present".
Update the keyval object visitor to use this a.0 syntax in error
messages rather than the usual a[0].
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <1488317230-26248-25-git-send-email-armbru@redhat.com>
[Off-by-one fix squashed in, as per Kevin's review]
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
2017-03-01 00:27:10 +03:00
|
|
|
/* Want a key index (unless it's first) or a QAPI name */
|
|
|
|
if (s != key && key_to_index(s, &end) >= 0) {
|
|
|
|
len = end - s;
|
|
|
|
} else {
|
|
|
|
ret = parse_qapi_name(s, false);
|
|
|
|
len = ret < 0 ? 0 : ret;
|
|
|
|
}
|
2017-03-01 00:27:05 +03:00
|
|
|
assert(s + len <= key_end);
|
|
|
|
if (!len || (s + len < key_end && s[len] != '.')) {
|
keyval: New keyval_parse()
keyval_parse() parses KEY=VALUE,... into a QDict. Works like
qemu_opts_parse(), except:
* Returns a QDict instead of a QemuOpts (d'oh).
* Supports nesting, unlike QemuOpts: a KEY is split into key
fragments at '.' (dotted key convention; the block layer does
something similar on top of QemuOpts). The key fragments are QDict
keys, and the last one's value is updated to VALUE.
* Each key fragment may be up to 127 bytes long. qemu_opts_parse()
limits the entire key to 127 bytes.
* Overlong key fragments are rejected. qemu_opts_parse() silently
truncates them.
* Empty key fragments are rejected. qemu_opts_parse() happily
accepts empty keys.
* It does not store the returned value. qemu_opts_parse() stores it
in the QemuOptsList.
* It does not treat parameter "id" specially. qemu_opts_parse()
ignores all but the first "id", and fails when its value isn't
id_wellformed(), or duplicate (a QemuOpts with the same ID is
already stored). It also screws up when a value contains ",id=".
* Implied value is not supported. qemu_opts_parse() desugars "foo" to
"foo=on", and "nofoo" to "foo=off".
* An implied key's value can't be empty, and can't contain ','.
I intend to grow this into a saner replacement for QemuOpts. It'll
take time, though.
Note: keyval_parse() provides no way to do lists, and its key syntax
is incompatible with the __RFQDN_ prefix convention for downstream
extensions, because it blindly splits at '.', even in __RFQDN_. Both
issues will be addressed later in the series.
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <1488317230-26248-4-git-send-email-armbru@redhat.com>
2017-03-01 00:26:49 +03:00
|
|
|
assert(key != implied_key);
|
|
|
|
error_setg(errp, "Invalid parameter '%.*s'",
|
|
|
|
(int)(key_end - key), key);
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
if (len >= sizeof(key_in_cur)) {
|
|
|
|
assert(key != implied_key);
|
|
|
|
error_setg(errp, "Parameter%s '%.*s' is too long",
|
|
|
|
s != key || s + len != key_end ? " fragment" : "",
|
|
|
|
(int)len, s);
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (s != key) {
|
|
|
|
next = keyval_parse_put(cur, key_in_cur, NULL,
|
|
|
|
key, s - 1, errp);
|
|
|
|
if (!next) {
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
cur = qobject_to_qdict(next);
|
|
|
|
assert(cur);
|
|
|
|
}
|
|
|
|
|
|
|
|
memcpy(key_in_cur, s, len);
|
|
|
|
key_in_cur[len] = 0;
|
|
|
|
s += len;
|
|
|
|
|
|
|
|
if (*s != '.') {
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
s++;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (key == implied_key) {
|
|
|
|
assert(!*s);
|
|
|
|
s = params;
|
|
|
|
} else {
|
|
|
|
if (*s != '=') {
|
|
|
|
error_setg(errp, "Expected '=' after parameter '%.*s'",
|
|
|
|
(int)(s - key), key);
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
s++;
|
|
|
|
}
|
|
|
|
|
|
|
|
val = qstring_new();
|
|
|
|
for (;;) {
|
|
|
|
if (!*s) {
|
|
|
|
break;
|
|
|
|
} else if (*s == ',') {
|
|
|
|
s++;
|
|
|
|
if (*s != ',') {
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
qstring_append_chr(val, *s++);
|
|
|
|
}
|
|
|
|
|
|
|
|
if (!keyval_parse_put(cur, key_in_cur, val, key, key_end, errp)) {
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
return s;
|
|
|
|
}
|
|
|
|
|
keyval: Support lists
Additionally permit non-negative integers as key components. A
dictionary's keys must either be all integers or none. If all keys
are integers, convert the dictionary to a list. The set of keys must
be [0,N].
Examples:
* list.1=goner,list.0=null,list.1=eins,list.2=zwei
is equivalent to JSON [ "null", "eins", "zwei" ]
* a.b.c=1,a.b.0=2
is inconsistent: a.b.c clashes with a.b.0
* list.0=null,list.2=eins,list.2=zwei
has a hole: list.1 is missing
Similar design flaw as for objects: there is no way to denote an empty
list. While interpreting "key absent" as empty list seems natural
(removing a list member from the input string works when there are
multiple ones, so why not when there's just one), it doesn't work:
"key absent" already means "optional list absent", which isn't the
same as "empty list present".
Update the keyval object visitor to use this a.0 syntax in error
messages rather than the usual a[0].
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <1488317230-26248-25-git-send-email-armbru@redhat.com>
[Off-by-one fix squashed in, as per Kevin's review]
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
2017-03-01 00:27:10 +03:00
|
|
|
static char *reassemble_key(GSList *key)
|
|
|
|
{
|
|
|
|
GString *s = g_string_new("");
|
|
|
|
GSList *p;
|
|
|
|
|
|
|
|
for (p = key; p; p = p->next) {
|
|
|
|
g_string_prepend_c(s, '.');
|
|
|
|
g_string_prepend(s, (char *)p->data);
|
|
|
|
}
|
|
|
|
|
|
|
|
return g_string_free(s, FALSE);
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Listify @cur recursively.
|
|
|
|
* Replace QDicts whose keys are all valid list indexes by QLists.
|
|
|
|
* @key_of_cur is the list of key fragments leading up to @cur.
|
|
|
|
* On success, return either @cur or its replacement.
|
|
|
|
* On failure, store an error through @errp and return NULL.
|
|
|
|
*/
|
|
|
|
static QObject *keyval_listify(QDict *cur, GSList *key_of_cur, Error **errp)
|
|
|
|
{
|
|
|
|
GSList key_node;
|
|
|
|
bool has_index, has_member;
|
|
|
|
const QDictEntry *ent;
|
|
|
|
QDict *qdict;
|
|
|
|
QObject *val;
|
|
|
|
char *key;
|
|
|
|
size_t nelt;
|
|
|
|
QObject **elt;
|
|
|
|
int index, max_index, i;
|
|
|
|
QList *list;
|
|
|
|
|
|
|
|
key_node.next = key_of_cur;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Recursively listify @cur's members, and figure out whether @cur
|
|
|
|
* itself is to be listified.
|
|
|
|
*/
|
|
|
|
has_index = false;
|
|
|
|
has_member = false;
|
|
|
|
for (ent = qdict_first(cur); ent; ent = qdict_next(cur, ent)) {
|
|
|
|
if (key_to_index(ent->key, NULL) >= 0) {
|
|
|
|
has_index = true;
|
|
|
|
} else {
|
|
|
|
has_member = true;
|
|
|
|
}
|
|
|
|
|
|
|
|
qdict = qobject_to_qdict(ent->value);
|
|
|
|
if (!qdict) {
|
|
|
|
continue;
|
|
|
|
}
|
|
|
|
|
|
|
|
key_node.data = ent->key;
|
|
|
|
val = keyval_listify(qdict, &key_node, errp);
|
|
|
|
if (!val) {
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
if (val != ent->value) {
|
|
|
|
qdict_put_obj(cur, ent->key, val);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
if (has_index && has_member) {
|
|
|
|
key = reassemble_key(key_of_cur);
|
|
|
|
error_setg(errp, "Parameters '%s*' used inconsistently", key);
|
|
|
|
g_free(key);
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
if (!has_index) {
|
|
|
|
return QOBJECT(cur);
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Copy @cur's values to @elt[] */
|
|
|
|
nelt = qdict_size(cur) + 1; /* one extra, for use as sentinel */
|
|
|
|
elt = g_new0(QObject *, nelt);
|
|
|
|
max_index = -1;
|
|
|
|
for (ent = qdict_first(cur); ent; ent = qdict_next(cur, ent)) {
|
|
|
|
index = key_to_index(ent->key, NULL);
|
|
|
|
assert(index >= 0);
|
|
|
|
if (index > max_index) {
|
|
|
|
max_index = index;
|
|
|
|
}
|
|
|
|
/*
|
|
|
|
* We iterate @nelt times. If we get one exceeding @nelt
|
|
|
|
* here, we will put less than @nelt values into @elt[],
|
|
|
|
* triggering the error in the next loop.
|
|
|
|
*/
|
|
|
|
if ((size_t)index >= nelt - 1) {
|
|
|
|
continue;
|
|
|
|
}
|
|
|
|
/* Even though dict keys are distinct, indexes need not be */
|
|
|
|
elt[index] = ent->value;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
2017-03-20 15:55:45 +03:00
|
|
|
* Make a list from @elt[], reporting the first missing element,
|
|
|
|
* if any.
|
keyval: Support lists
Additionally permit non-negative integers as key components. A
dictionary's keys must either be all integers or none. If all keys
are integers, convert the dictionary to a list. The set of keys must
be [0,N].
Examples:
* list.1=goner,list.0=null,list.1=eins,list.2=zwei
is equivalent to JSON [ "null", "eins", "zwei" ]
* a.b.c=1,a.b.0=2
is inconsistent: a.b.c clashes with a.b.0
* list.0=null,list.2=eins,list.2=zwei
has a hole: list.1 is missing
Similar design flaw as for objects: there is no way to denote an empty
list. While interpreting "key absent" as empty list seems natural
(removing a list member from the input string works when there are
multiple ones, so why not when there's just one), it doesn't work:
"key absent" already means "optional list absent", which isn't the
same as "empty list present".
Update the keyval object visitor to use this a.0 syntax in error
messages rather than the usual a[0].
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <1488317230-26248-25-git-send-email-armbru@redhat.com>
[Off-by-one fix squashed in, as per Kevin's review]
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
2017-03-01 00:27:10 +03:00
|
|
|
* If we dropped an index >= nelt in the previous loop, this loop
|
|
|
|
* will run into the sentinel and report index @nelt missing.
|
|
|
|
*/
|
|
|
|
list = qlist_new();
|
|
|
|
assert(!elt[nelt-1]); /* need the sentinel to be null */
|
|
|
|
for (i = 0; i < MIN(nelt, max_index + 1); i++) {
|
|
|
|
if (!elt[i]) {
|
|
|
|
key = reassemble_key(key_of_cur);
|
|
|
|
error_setg(errp, "Parameter '%s%d' missing", key, i);
|
|
|
|
g_free(key);
|
|
|
|
g_free(elt);
|
|
|
|
QDECREF(list);
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
qobject_incref(elt[i]);
|
|
|
|
qlist_append_obj(list, elt[i]);
|
|
|
|
}
|
|
|
|
|
|
|
|
g_free(elt);
|
|
|
|
return QOBJECT(list);
|
|
|
|
}
|
|
|
|
|
keyval: New keyval_parse()
keyval_parse() parses KEY=VALUE,... into a QDict. Works like
qemu_opts_parse(), except:
* Returns a QDict instead of a QemuOpts (d'oh).
* Supports nesting, unlike QemuOpts: a KEY is split into key
fragments at '.' (dotted key convention; the block layer does
something similar on top of QemuOpts). The key fragments are QDict
keys, and the last one's value is updated to VALUE.
* Each key fragment may be up to 127 bytes long. qemu_opts_parse()
limits the entire key to 127 bytes.
* Overlong key fragments are rejected. qemu_opts_parse() silently
truncates them.
* Empty key fragments are rejected. qemu_opts_parse() happily
accepts empty keys.
* It does not store the returned value. qemu_opts_parse() stores it
in the QemuOptsList.
* It does not treat parameter "id" specially. qemu_opts_parse()
ignores all but the first "id", and fails when its value isn't
id_wellformed(), or duplicate (a QemuOpts with the same ID is
already stored). It also screws up when a value contains ",id=".
* Implied value is not supported. qemu_opts_parse() desugars "foo" to
"foo=on", and "nofoo" to "foo=off".
* An implied key's value can't be empty, and can't contain ','.
I intend to grow this into a saner replacement for QemuOpts. It'll
take time, though.
Note: keyval_parse() provides no way to do lists, and its key syntax
is incompatible with the __RFQDN_ prefix convention for downstream
extensions, because it blindly splits at '.', even in __RFQDN_. Both
issues will be addressed later in the series.
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <1488317230-26248-4-git-send-email-armbru@redhat.com>
2017-03-01 00:26:49 +03:00
|
|
|
/*
|
|
|
|
* Parse @params in QEMU's traditional KEY=VALUE,... syntax.
|
|
|
|
* If @implied_key, the first KEY= can be omitted. @implied_key is
|
|
|
|
* implied then, and VALUE can't be empty or contain ',' or '='.
|
|
|
|
* On success, return a dictionary of the parsed keys and values.
|
|
|
|
* On failure, store an error through @errp and return NULL.
|
|
|
|
*/
|
|
|
|
QDict *keyval_parse(const char *params, const char *implied_key,
|
|
|
|
Error **errp)
|
|
|
|
{
|
|
|
|
QDict *qdict = qdict_new();
|
keyval: Support lists
Additionally permit non-negative integers as key components. A
dictionary's keys must either be all integers or none. If all keys
are integers, convert the dictionary to a list. The set of keys must
be [0,N].
Examples:
* list.1=goner,list.0=null,list.1=eins,list.2=zwei
is equivalent to JSON [ "null", "eins", "zwei" ]
* a.b.c=1,a.b.0=2
is inconsistent: a.b.c clashes with a.b.0
* list.0=null,list.2=eins,list.2=zwei
has a hole: list.1 is missing
Similar design flaw as for objects: there is no way to denote an empty
list. While interpreting "key absent" as empty list seems natural
(removing a list member from the input string works when there are
multiple ones, so why not when there's just one), it doesn't work:
"key absent" already means "optional list absent", which isn't the
same as "empty list present".
Update the keyval object visitor to use this a.0 syntax in error
messages rather than the usual a[0].
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <1488317230-26248-25-git-send-email-armbru@redhat.com>
[Off-by-one fix squashed in, as per Kevin's review]
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
2017-03-01 00:27:10 +03:00
|
|
|
QObject *listified;
|
keyval: New keyval_parse()
keyval_parse() parses KEY=VALUE,... into a QDict. Works like
qemu_opts_parse(), except:
* Returns a QDict instead of a QemuOpts (d'oh).
* Supports nesting, unlike QemuOpts: a KEY is split into key
fragments at '.' (dotted key convention; the block layer does
something similar on top of QemuOpts). The key fragments are QDict
keys, and the last one's value is updated to VALUE.
* Each key fragment may be up to 127 bytes long. qemu_opts_parse()
limits the entire key to 127 bytes.
* Overlong key fragments are rejected. qemu_opts_parse() silently
truncates them.
* Empty key fragments are rejected. qemu_opts_parse() happily
accepts empty keys.
* It does not store the returned value. qemu_opts_parse() stores it
in the QemuOptsList.
* It does not treat parameter "id" specially. qemu_opts_parse()
ignores all but the first "id", and fails when its value isn't
id_wellformed(), or duplicate (a QemuOpts with the same ID is
already stored). It also screws up when a value contains ",id=".
* Implied value is not supported. qemu_opts_parse() desugars "foo" to
"foo=on", and "nofoo" to "foo=off".
* An implied key's value can't be empty, and can't contain ','.
I intend to grow this into a saner replacement for QemuOpts. It'll
take time, though.
Note: keyval_parse() provides no way to do lists, and its key syntax
is incompatible with the __RFQDN_ prefix convention for downstream
extensions, because it blindly splits at '.', even in __RFQDN_. Both
issues will be addressed later in the series.
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <1488317230-26248-4-git-send-email-armbru@redhat.com>
2017-03-01 00:26:49 +03:00
|
|
|
const char *s;
|
|
|
|
|
|
|
|
s = params;
|
|
|
|
while (*s) {
|
|
|
|
s = keyval_parse_one(qdict, s, implied_key, errp);
|
|
|
|
if (!s) {
|
|
|
|
QDECREF(qdict);
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
implied_key = NULL;
|
|
|
|
}
|
|
|
|
|
keyval: Support lists
Additionally permit non-negative integers as key components. A
dictionary's keys must either be all integers or none. If all keys
are integers, convert the dictionary to a list. The set of keys must
be [0,N].
Examples:
* list.1=goner,list.0=null,list.1=eins,list.2=zwei
is equivalent to JSON [ "null", "eins", "zwei" ]
* a.b.c=1,a.b.0=2
is inconsistent: a.b.c clashes with a.b.0
* list.0=null,list.2=eins,list.2=zwei
has a hole: list.1 is missing
Similar design flaw as for objects: there is no way to denote an empty
list. While interpreting "key absent" as empty list seems natural
(removing a list member from the input string works when there are
multiple ones, so why not when there's just one), it doesn't work:
"key absent" already means "optional list absent", which isn't the
same as "empty list present".
Update the keyval object visitor to use this a.0 syntax in error
messages rather than the usual a[0].
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <1488317230-26248-25-git-send-email-armbru@redhat.com>
[Off-by-one fix squashed in, as per Kevin's review]
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
2017-03-01 00:27:10 +03:00
|
|
|
listified = keyval_listify(qdict, NULL, errp);
|
|
|
|
if (!listified) {
|
|
|
|
QDECREF(qdict);
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
assert(listified == QOBJECT(qdict));
|
keyval: New keyval_parse()
keyval_parse() parses KEY=VALUE,... into a QDict. Works like
qemu_opts_parse(), except:
* Returns a QDict instead of a QemuOpts (d'oh).
* Supports nesting, unlike QemuOpts: a KEY is split into key
fragments at '.' (dotted key convention; the block layer does
something similar on top of QemuOpts). The key fragments are QDict
keys, and the last one's value is updated to VALUE.
* Each key fragment may be up to 127 bytes long. qemu_opts_parse()
limits the entire key to 127 bytes.
* Overlong key fragments are rejected. qemu_opts_parse() silently
truncates them.
* Empty key fragments are rejected. qemu_opts_parse() happily
accepts empty keys.
* It does not store the returned value. qemu_opts_parse() stores it
in the QemuOptsList.
* It does not treat parameter "id" specially. qemu_opts_parse()
ignores all but the first "id", and fails when its value isn't
id_wellformed(), or duplicate (a QemuOpts with the same ID is
already stored). It also screws up when a value contains ",id=".
* Implied value is not supported. qemu_opts_parse() desugars "foo" to
"foo=on", and "nofoo" to "foo=off".
* An implied key's value can't be empty, and can't contain ','.
I intend to grow this into a saner replacement for QemuOpts. It'll
take time, though.
Note: keyval_parse() provides no way to do lists, and its key syntax
is incompatible with the __RFQDN_ prefix convention for downstream
extensions, because it blindly splits at '.', even in __RFQDN_. Both
issues will be addressed later in the series.
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <1488317230-26248-4-git-send-email-armbru@redhat.com>
2017-03-01 00:26:49 +03:00
|
|
|
return qdict;
|
|
|
|
}
|