<?xml version="1.0" encoding="utf-8" ?>
<!DOCTYPE cref SYSTEM "cref.dtd">
<cref>
<header>
<copyright>
<year>1998</year><year>2016</year>
<holder>Ericsson AB. All Rights Reserved.</holder>
</copyright>
<legalnotice>
Licensed under the Apache License, Version 2.0 (the "License");
you may not use this file except in compliance with the License.
You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
See the License for the specific language governing permissions and
limitations under the License.
</legalnotice>
<title>registry</title>
<prepared>Gordon Beaton</prepared>
<responsible>Gordon Beaton</responsible>
<docno></docno>
<approved>Gordon Beaton</approved>
<checked>Gordon Beaton</checked>
<date>980707</date>
<rev>A</rev>
<file>registry.sgml</file>
</header>
<lib>registry</lib>
<libsummary>Store and backup key-value pairs</libsummary>
<description>
<p>This module provides support for storing key-value
pairs in a table known as a registry, backing up registries to
Mnesia in an atomic manner, and later restoring the contents of a
registry from Mnesia.</p>
</description>
<funcs>
<func>
<name><ret>ei_reg *</ret><nametext>ei_reg_open(size)</nametext></name>
<fsummary>Create and open a registry</fsummary>
<type>
<v>int size;</v>
</type>
<desc>
<p>Open (create) a registry. The registry will be
initially empty. Use <c><![CDATA[ei_reg_close()]]></c> to close the registry
later.
</p>
<p><c><![CDATA[size]]></c> is the approximate number of objects you intend
to store in the registry. Since the registry uses a hash table
with collision chaining, there is no absolute upper limit on the
number of objects that can be stored in it. However for reasons
of efficiency, it is a good idea to choose a number that is
appropriate for your needs. It is possible to use
<c><![CDATA[ei_reg_resize()]]></c> to change the size later. Note that the
number you provide will be increased to the nearest larger prime
number.
</p>
<p>On success, an empty registry will be returned. On failure, NULL
will be returned.</p>
</desc>
</func>
<func>
<name><ret>int</ret><nametext>ei_reg_resize(reg,newsize)</nametext></name>
<fsummary>Resize a registry</fsummary>
<type>
<v>ei_reg *reg;</v>
<v>int newsize;</v>
</type>
<desc>
<p>Change the size of a registry.
</p>
<p><c><![CDATA[newsize]]></c> is the new size to make the registry. The
number will be increased to the nearest larger prime number.
</p>
<p>On success, the registry will be resized, all contents
rehashed, and the function will return 0. On failure, the
registry will be left unchanged and the function will return -1.</p>
</desc>
</func>
<func>
<name><ret>int</ret><nametext>ei_reg_close(reg)</nametext></name>
<fsummary>Close a registry </fsummary>
<type>
<v>ei_reg *reg;</v>
</type>
<desc>
<p>A registry that has previously been created with
<c><![CDATA[ei_reg_open()]]></c> is closed, and all the objects it contains
are freed.
</p>
<p><c><![CDATA[reg]]></c> is the registry to close.
</p>
<p>The function returns 0.</p>
</desc>
</func>
<func>
<name><ret>int</ret><nametext>ei_reg_setival(reg,key,i)</nametext></name>
<fsummary>Assign an integer object</fsummary>
<type>
<v>ei_reg *reg;</v>
<v>const char *key;</v>
<v>int i;</v>
</type>
<desc>
<p>Create a key-value pair with the specified <c><![CDATA[key]]></c> and integer
value <c><![CDATA[i]]></c>. If an object already existed with the same
<c><![CDATA[key]]></c>, the new value replaces the old one. If the previous
value was a binary or string, it is freed with <c><![CDATA[free()]]></c>.
</p>
<p><c><![CDATA[reg]]></c> is the registry where the object should be placed.
</p>
<p><c><![CDATA[key]]></c> is the name of the object.
</p>
<p><c><![CDATA[i]]></c> is the integer value to assign.
</p>
<p>The function returns 0 on success, or -1 on failure.</p>
</desc>
</func>
<func>
<name><ret>int</ret><nametext>ei_reg_setfval(reg,key,f)</nametext></name>
<fsummary>Assign a floating point object</fsummary>
<type>
<v>ei_reg *reg;</v>
<v>const char *key;</v>
<v>double f;</v>
</type>
<desc>
<p>Create a key-value pair with the specified <c><![CDATA[key]]></c> and
floating point value <c><![CDATA[f]]></c>. If an object already existed with
the same <c><![CDATA[key]]></c>, the new value replaces the old one. If the
previous value was a binary or string, it is freed with <c><![CDATA[free()]]></c>.
</p>
<p><c><![CDATA[reg]]></c> is the registry where the object should be placed.
</p>
<p><c><![CDATA[key]]></c> is the name of the object.
</p>
<p><c><![CDATA[f]]></c> is the floating point value to assign.
</p>
<p>The function returns 0 on success, or -1 on failure.</p>
</desc>
</func>
<func>
<name><ret>int</ret><nametext>ei_reg_setsval(reg,key,s)</nametext></name>
<fsummary>Assign a string object</fsummary>
<type>
<v>ei_reg *reg;</v>
<v>const char *key;</v>
<v>const char *s;</v>
</type>
<desc>
<p>Create a key-value pair with the specified <c><![CDATA[key]]></c> whose
"value" is the specified string <c><![CDATA[s]]></c>. If an object already
existed with the same <c><![CDATA[key]]></c>, the new value replaces the old
one. If the previous value was a binary or string, it is freed
with <c><![CDATA[free()]]></c>.
</p>
<p><c><![CDATA[reg]]></c> is the registry where the object should be placed.
</p>
<p><c><![CDATA[key]]></c> is the name of the object.
</p>
<p><c><![CDATA[s]]></c> is the string to assign. The string itself
must have been created through a single call to <c><![CDATA[malloc()]]></c> or
similar function, so that the registry can later delete it if
necessary by calling <c><![CDATA[free()]]></c>.
</p>
<p>The function returns 0 on success, or -1 on failure.</p>
</desc>
</func>
<func>
<name><ret>int</ret><nametext>ei_reg_setpval(reg,key,p,size)</nametext></name>
<fsummary>Assign a binary object</fsummary>
<type>
<v>ei_reg *reg;</v>
<v>const char *key;</v>
<v>const void *p;</v>
<v>int size;</v>
</type>
<desc>
<p>Create a key-value pair with the specified <c><![CDATA[key]]></c> whose
"value" is the binary object pointed to by <c><![CDATA[p]]></c>. If an
object already existed with the same <c><![CDATA[key]]></c>, the new value
replaces the old one. If the previous value was a binary or
string, it is freed with <c><![CDATA[free()]]></c>.
</p>
<p><c><![CDATA[reg]]></c> is the registry where the object should be placed.
</p>
<p><c><![CDATA[key]]></c> is the name of the object.
</p>
<p><c><![CDATA[p]]></c> is a pointer to the binary object. The object itself
must have been created through a single call to <c><![CDATA[malloc()]]></c> or
similar function, so that the registry can later delete it if
necessary by calling <c><![CDATA[free()]]></c>.
</p>
<p><c><![CDATA[size]]></c> is the length in bytes of the binary object.
</p>
<p>The function returns 0 on success, or -1 on failure.</p>
</desc>
</func>
<func>
<name><ret>int</ret><nametext>ei_reg_setval(reg,key,flags,v,...)</nametext></name>
<fsummary>Assign a value to any object type</fsummary>
<type>
<v>ei_reg *reg;</v>
<v>const char *key;</v>
<v>int flags;</v>
<v>v (see below)</v>
</type>
<desc>
<p>Create a key-value pair with the specified <c><![CDATA[key]]></c> whose
value is specified by <c><![CDATA[v]]></c>. If an object already
existed with the same <c><![CDATA[key]]></c>, the new value replaces the old
one. If the previous value was a binary or string, it is freed
with <c><![CDATA[free()]]></c>.
</p>
<p><c><![CDATA[reg]]></c> is the registry where the object should be placed.
</p>
<p><c><![CDATA[key]]></c> is the name of the object.
</p>
<p><c><![CDATA[flags]]></c> indicates the type of the object specified by
<c><![CDATA[v]]></c>. Flags must be one of EI_INT, EI_FLT, EI_STR and
EI_BIN, indicating whether <c><![CDATA[v]]></c> is <c><![CDATA[int]]></c>, <c><![CDATA[double]]></c>,
<c><![CDATA[char*]]></c> or <c><![CDATA[void*]]></c>. If <c><![CDATA[flags]]></c> is EI_BIN, then a
fifth argument <c><![CDATA[size]]></c> is required, indicating the size
in bytes of the object pointed to by <c><![CDATA[v]]></c>.
</p>
<p>If you wish to store an arbitrary pointer in the registry,
specify a <c><![CDATA[size]]></c> of 0. In this case, the object itself will
not be transferred by an <c><![CDATA[ei_reg_dump()]]></c> operation, just
the pointer value.
</p>
<p>The function returns 0 on success, or -1 on failure.</p>
</desc>
</func>
<func>
<name><ret>int</ret><nametext>ei_reg_getival(reg,key)</nametext></name>
<fsummary>Get an integer object</fsummary>
<type>
<v>ei_reg *reg;</v>
<v>const char *key;</v>
</type>
<desc>
<p>Get the value associated with <c><![CDATA[key]]></c> in the
registry. The value must be an integer.
</p>
<p><c><![CDATA[reg]]></c> is the registry where the object will be looked
up.
</p>
<p><c><![CDATA[key]]></c> is the name of the object to look up.
</p>
<p>On success, the function returns the value associated with <c><![CDATA[key]]></c>.
If the object was not found or it was not an integer
object, -1 is returned. To avoid problems with in-band error
reporting (i.e. if you cannot distinguish between -1 and a
valid result) use the more general function <c><![CDATA[ei_reg_getval()]]></c>
instead.</p>
</desc>
</func>
<func>
<name><ret>double</ret><nametext>ei_reg_getfval(reg,key)</nametext></name>
<fsummary>Get a floating point object</fsummary>
<type>
<v>ei_reg *reg;</v>
<v>const char *key;</v>
</type>
<desc>
<p>Get the value associated with <c><![CDATA[key]]></c> in the
registry. The value must be a floating point type.
</p>
<p><c><![CDATA[reg]]></c> is the registry where the object will be looked
up.
</p>
<p><c><![CDATA[key]]></c> is the name of the object to look up.
</p>
<p>On success, the function returns the value associated with <c><![CDATA[key]]></c>.
If the object was not found or it was not a floating point
object, -1.0 is returned. To avoid problems with in-band error
reporting (i.e. if you cannot distinguish between -1.0 and a
valid result) use the more general function <c><![CDATA[ei_reg_getval()]]></c>
instead.</p>
</desc>
</func>
<func>
<name><ret>const char *</ret><nametext>ei_reg_getsval(reg,key)</nametext></name>
<fsummary>Get a string object</fsummary>
<type>
<v>ei_reg *reg;</v>
<v>const char *key;</v>
</type>
<desc>
<p>Get the value associated with <c><![CDATA[key]]></c> in the
registry. The value must be a string.
</p>
<p><c><![CDATA[reg]]></c> is the registry where the object will be looked
up.
</p>
<p><c><![CDATA[key]]></c> is the name of the object to look up.
</p>
<p>On success, the function returns the value associated with
<c><![CDATA[key]]></c>. If the object was not found or it was not a string,
NULL is returned. To avoid problems with in-band error
reporting (i.e. if you cannot distinguish between NULL and a
valid result) use the more general function <c><![CDATA[ei_reg_getval()]]></c>
instead.</p>
</desc>
</func>
<func>
<name><ret>const void *</ret><nametext>ei_reg_getpval(reg,key,size)</nametext></name>
<fsummary>Get a binary object</fsummary>
<type>
<v>ei_reg *reg;</v>
<v>const char *key;</v>
<v>int size;</v>
</type>
<desc>
<p>Get the value associated with <c><![CDATA[key]]></c> in the
registry. The value must be a binary (pointer) type.
</p>
<p><c><![CDATA[reg]]></c> is the registry where the object will be looked
up.
</p>
<p><c><![CDATA[key]]></c> is the name of the object to look up.
</p>
<p><c><![CDATA[size]]></c> will be initialized to contain the length in
bytes of the object, if it is found.
</p>
<p>On success, the function returns the value associated with
<c><![CDATA[key]]></c> and indicates its length in <c><![CDATA[size]]></c>.
If the object was not found or it was not a binary object,
NULL is returned. To avoid problems with in-band error
reporting (i.e. if you cannot distinguish between NULL and a
valid result) use the more general function <c><![CDATA[ei_reg_getval()]]></c>
instead.</p>
</desc>
</func>
<func>
<name><ret>int</ret><nametext>ei_reg_getval(reg,key,flags,v,...)</nametext></name>
<fsummary>Get any object</fsummary>
<type>
<v>ei_reg *reg;</v>
<v>const char *key;</v>
<v>int flags;</v>
<v>void *v (see below)</v>
</type>
<desc>
<p>This is a general function for retrieving any kind of
object from the registry.
</p>
<p><c><![CDATA[reg]]></c> is the registry where the object will be looked
up.
</p>
<p><c><![CDATA[key]]></c> is the name of the object to look up.
</p>
<p><c><![CDATA[flags]]></c> indicates the type of object that you are
looking for. If <c><![CDATA[flags]]></c> is 0, then any kind of object will
be returned. If <c><![CDATA[flags]]></c> is one of EI_INT, EI_FLT, EI_STR or
EI_BIN, then only values of that kind will be returned. The
buffer pointed to by <c><![CDATA[v]]></c> must be large enough to hold the return
data, i.e. it must be a pointer to one of <c><![CDATA[int]]></c>,
<c><![CDATA[double]]></c>, <c><![CDATA[char*]]></c> or <c><![CDATA[void*]]></c>, respectively. Also,
if <c><![CDATA[flags]]></c> is EI_BIN, then a fifth argument <c><![CDATA[int *size]]></c> is required, so that the size of the object can be
returned.
</p>
<p>If the function succeeds, <c><![CDATA[v]]></c> (and <c><![CDATA[size]]></c> if the
object is binary) will be initialized with the value associated
with <c><![CDATA[key]]></c>, and the function will return one of EI_INT,
EI_FLT, EI_STR or EI_BIN, indicating the type of object. On failure the
function will return -1 and the arguments will not be updated.</p>
</desc>
</func>
<func>
<name><ret>int</ret><nametext>ei_reg_markdirty(reg,key)</nametext></name>
<fsummary>Mark an object as dirty </fsummary>
<type>
<v>ei_reg *reg;</v>
<v>const char *key;</v>
</type>
<desc>
<p>Mark a registry object as dirty. This will ensure that
it is included in the next backup to Mnesia. Normally this
operation will not be necessary since all of the normal registry
'set' functions do this automatically. However if you have
retrieved the value of a string or binary object from the
registry and modified the contents, then the change will be
invisible to the registry and the object will be assumed to be
unmodified. This function allows you to make such modifications
and then let the registry know about them.
</p>
<p><c><![CDATA[reg]]></c> is the registry containing the object.
</p>
<p><c><![CDATA[key]]></c> is the name of the object to mark.
</p>
<p>The function returns 0 on success, or -1 on failure.</p>
</desc>
</func>
<func>
<name><ret>int</ret><nametext>ei_reg_delete(reg,key)</nametext></name>
<fsummary>Delete an object from the registry</fsummary>
<type>
<v>ei_reg *reg;</v>
<v>const char *key;</v>
</type>
<desc>
<p>Delete an object from the registry. The object is not
actually removed from the registry, it is only marked for later
removal so that on subsequent backups to Mnesia, the
corresponding object can be removed from the Mnesia table as
well. If another object is later created with the same key, the
object will be reused.
</p>
<p>The object will be removed from the registry after a call to
<c><![CDATA[ei_reg_dump()]]></c> or <c><![CDATA[ei_reg_purge()]]></c>.
</p>
<p><c><![CDATA[reg]]></c> is the registry containing <c><![CDATA[key]]></c>.
</p>
<p><c><![CDATA[key]]></c> is the object to remove.
</p>
<p>If the object was found, the function returns 0 indicating
success. Otherwise the function returns -1.</p>
</desc>
</func>
<func>
<name><ret>int</ret><nametext>ei_reg_stat(reg,key,obuf)</nametext></name>
<fsummary>Get object information</fsummary>
<type>
<v>ei_reg *reg;</v>
<v>const char *key;</v>
<v>struct ei_reg_stat *obuf;</v>
</type>
<desc>
<p>Return information about an object.
</p>
<p><c><![CDATA[reg]]></c> is the registry containing the object.
</p>
<p><c><![CDATA[key]]></c> is the name of the object.
</p>
<p><c><![CDATA[obuf]]></c> is a pointer to an <c><![CDATA[ei_reg_stat]]></c> structure,
defined below:
</p>
<code type="none"><![CDATA[
struct ei_reg_stat {
int attr;
int size;
};
]]></code>
<p>In <c><![CDATA[attr]]></c> the object's attributes are stored as the logical
OR of its type (one of EI_INT, EI_FLT, EI_BIN and EI_STR),
whether it is marked for deletion (EI_DELET) and whether it has
been modified since the last backup to Mnesia (EI_DIRTY).
</p>
<p>The <c><![CDATA[size]]></c> field indicates the size in bytes required to store
EI_STR (including the terminating 0) and EI_BIN objects, or 0
for EI_INT and EI_FLT.
</p>
<p>The function returns 0 and initializes <c><![CDATA[obuf]]></c> on
success, or returns -1 on failure.</p>
</desc>
</func>
<func>
<name><ret>int</ret><nametext>ei_reg_tabstat(reg,obuf)</nametext></name>
<fsummary>Get registry information</fsummary>
<type>
<v>ei_reg *reg;</v>
<v>struct ei_reg_tabstat *obuf;</v>
</type>
<desc>
<p>Return information about a registry. Using information
returned by this function, you can see whether the size of the
registry is suitable for the amount of data it contains.
</p>
<p><c><![CDATA[reg]]></c> is the registry to return information about.
</p>
<p><c><![CDATA[obuf]]></c> is a pointer to an <c><![CDATA[ei_reg_tabstat]]></c> structure,
defined below:
</p>
<code type="none"><![CDATA[
struct ei_reg_tabstat {
int size;
int nelem;
int npos;
int collisions;
};
]]></code>
<p>The <c><![CDATA[size]]></c> field indicates the number of hash positions
in the registry. This is the number you provided when you
created or last resized the registry, rounded up to the nearest
prime.
</p>
<p><c><![CDATA[nelem]]></c> indicates the number of elements stored in the
registry. It includes objects that are deleted but not purged.
</p>
<p><c><![CDATA[npos]]></c> indicates the number of unique positions that are
occupied in the registry.
</p>
<p><c><![CDATA[collisions]]></c> indicates how many elements are sharing
positions in the registry.
</p>
<p>On success, the function returns 0 and <c><![CDATA[obuf]]></c> is
initialized to contain table statistics. On failure, the function
returns -1.</p>
</desc>
</func>
<func>
<name><ret>int</ret><nametext>ei_reg_dump(fd,reg,mntab,flags)</nametext></name>
<fsummary>Back up a registry to Mnesia</fsummary>
<type>
<v>int fd;</v>
<v>ei_reg *reg;</v>
<v>const char *mntab;</v>
<v>int flags;</v>
</type>
<desc>
<p>Dump the contents of a registry to a Mnesia table in an
atomic manner, i.e. either all data will be updated, or none of
it will. If any errors are encountered while backing up
the data, the entire operation is aborted.
</p>
<p><c><![CDATA[fd]]></c> is an open connection to Erlang.
Mnesia 3.0 or later must be running on the Erlang node.
</p>
<p><c><![CDATA[reg]]></c> is the registry to back up.
</p>
<p><c><![CDATA[mntab]]></c> is the name of the Mnesia table where the backed
up data should be placed. If the table does not exist, it will
be created automatically using configurable defaults. See your
Mnesia documentation for information about configuring this
behaviour.
</p>
<p>If <c><![CDATA[flags]]></c> is 0, the backup will include only those
objects which have been created, modified or deleted since the
last backup or restore (i.e. an incremental backup). After the
backup, any objects that were marked dirty are now clean, and any
objects that had been marked for deletion are deleted.
</p>
<p>Alternatively, setting flags to EI_FORCE will cause a full
backup to be done, and EI_NOPURGE will cause the deleted objects
to be left in the registry afterwards. These can be bitwise ORed
together if both behaviours are desired. If EI_NOPURGE was
specified, you can use <c><![CDATA[ei_reg_purge()]]></c> to explicitly remove
the deleted items from the registry later.
</p>
<p>The function returns 0 on success, or -1 on failure.</p>
</desc>
</func>
<func>
<name><ret>int</ret><nametext>ei_reg_restore(fd,reg,mntab)</nametext></name>
<fsummary>Restore a registry from Mnesia</fsummary>
<type>
<v>int fd;</v>
<v>ei_reg *reg;</v>
<v>const char *mntab;</v>
</type>
<desc>
<p>The contents of a Mnesia table are read into the
registry.
</p>
<p><c><![CDATA[fd]]></c> is an open connection to Erlang.
Mnesia 3.0 or later must be running on the Erlang node.
</p>
<p><c><![CDATA[reg]]></c> is the registry where the data should be placed.
</p>
<p><c><![CDATA[mntab]]></c> is the name of the Mnesia table to read data
from.
</p>
<p>Note that only tables of a certain format can be
restored, i.e. those that have been created and backed up to
with <c><![CDATA[ei_reg_dump()]]></c>. If the registry was not empty before
the operation, then the contents of the table are added to the
contents of the registry. If the table contains objects with the
same keys as those already in the registry, the registry objects
will be overwritten with the new values. If the registry
contains objects that were not in the table, they will be
unchanged by this operation.
</p>
<p>After the restore operation, the entire contents of the
registry is marked as unmodified. Note that this includes any
objects that were modified before the restore and not
overwritten by the restore.
</p>
<p>The function returns 0 on success, or -1 on failure.</p>
</desc>
</func>
<func>
<name><ret>int</ret><nametext>ei_reg_purge(reg)</nametext></name>
<fsummary>Remove deleted objects</fsummary>
<type>
<v>ei_reg *reg;</v>
</type>
<desc>
<p>Remove all objects marked for deletion. When objects
are deleted with <c><![CDATA[ei_reg_delete()]]></c> they are not actually
removed from the registry, only marked for later removal. This
is so that on a subsequent backup to Mnesia, the
objects can also be removed from the Mnesia table. If you are
not backing up to Mnesia then you may wish to remove the objects
manually with this function.
</p>
<p><c><![CDATA[reg]]></c> is a registry containing objects marked for
deletion.
</p>
<p>The function returns 0 on success, or -1 on failure.</p>
</desc>
</func>
</funcs>
</cref>
