| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
| |
The patch contains two major pieces of work:
* The beginning of the new INI interface
significantly re-thought and reworked.
* The implementation of the file object.
Parsing was adjusted to use this object.
|
|
|
|
|
|
|
|
|
|
| |
I realized that the new high level interface should be
defined in one file instead of being scattered in many headers.
I would have to eventuallu eliminate ini_configobj.h too
and move everything to the new interface in ini_config.h.
This patch renames couple functions to be consistent
with the naming convention and removes ini_parse.h header.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The parser is added to the existing module.
The old parsing functuion will be removed when we
switch to the new interface.
Parser logic:
* There is one high level function that wraps the
parser interface. It is at the bottom of the module.
ini_parse_config();
* Internally it creates a perser object
and then runs parser on it.
* At the end parser object is destroyed.
* This object stores the state of the parser.
* The parser has an action queue
* There are several actions that parser can perform
- read line
- inspect read line
- record an error
- process last comment in the file (POST)
* Each action handler determines what to do next
depending upon what has happened.
* Read handler reads lines and enqueues inspection action
in case of success or error action in case of failure.
* Inspection action parses last read line and
treats it either is a:
* Comment
* Section
* New key + value
* Continuation of the value
In case of error the error action is enqueued.
* Error can be fatal or non fatal. It depend on the error_level
flag passed in. If the error is non fatal the read action is
enqueued otherwise parser stops.
* The POST action is a special action to handle comment at
the bottom of the file. The comment is stored with the
value it preceeds so in case there is a comment at
the bottom of the file a special value needs to be created
to hold just the comment.
|
|
|
|
|
|
|
|
| |
* ini_config_priv.h - private header
* ini_configobj.h - future public header for the new interface
* ini_configobj.c - just constructor/decructor for now
* ini_serialize.c - code to serialize the configuration
into a buffer so it can be written somewhere.
|
|
|
|
|
|
|
|
|
|
|
|
| |
* Removed the annoyong type casting in the serialization
since there is a new function in the simplebuffer to
addstring.
* Changed the serialization code to not create simple
buffer object but rather add to the passed in one.
* Added serialization of the value that stores section
* Started cleaning constants but then stopped since
this effor should be done a bit later when the new
high level ini config interface is better baked.
|
| |
|
|
|
|
|
| |
Take advantage of the already known string length instead
of counting bytes once more.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Value object is an object that combines
a potentially multiline value
comment and some statistical information
regarding a configuration value.
Patch includes:
Source
Header
Unit test
Makefile changes
|
|
|
|
| |
Comment object should store lengths of the lines.
|
|
|
|
|
|
|
|
| |
Addressing coverity issue. Ticket #513.
The memory was really leaked when
the configuration value was empty.
Added unit test that confirmed the bug.
MANUAL MERGE.
|
|
|
|
|
|
|
|
| |
Addressing ticket #504
Conflicts:
common/ini/ini_config.c
|
| |
|
|
|
|
|
| |
This is a reworked patch to add support for explicit
32 and 64 bit values in the config files.
|
|
|
|
|
|
|
|
|
| |
Comment object will store
the comments found in the INI file.
It is based on the ref_array
interface.
Fixing review comments for comment obj.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Time came to split ini_config.c
into many much smaller pieces.
1) ini_parse.c - will have parsing functions
2) ini_get_value.c - will have single value
interpretation functions
3) ini_get_array.c - will have array interpretation
functions.
4) ini_print.c - error printing
5) ini_defines.h - common constants
6) ini_parse.h header for parsing functions
7) ini_list.c - will have list processing functions
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
When running 'make distcheck', the entire source directory is set
to read-only, to ensure that the build process only has write
access to $builddir. As a result, this was causing the unit test
for file mode to fail, since the file it was testing resides in
the $srcdir.
This patch guarantees that the test file has the correct
permissions prior to running the access test.
|
|
|
|
| |
[TRACE] Adding macros for signed numbers
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
1) Fixed the issue that metadata was saved
as numbers. Was supposed to be saved as strings.
2) Added two functions. One is to check permissions
on the config file. Another to check if the file
has changed and thus the cinfiguration needs
to be reread.
3) Added unit test will sample code
and comments how to use the functions.
4) Added doxygen description in the comments.
5) Fixed couple typos and ommisions here and there.
[INI] Fixing crash detected on 64-bit system
This patch corrects original code to be
more on the safe side and check parameters
before using.
Instead of dereferencing metadata it is now
passed as reference to the next level.
It is not used there yet so no other new changes
needed so far.
[INI] Addressing review comments
[INI] Addressing comments.
|
| |
|
|
|
|
|
|
| |
This patch implements function that collects
stats and saves them in the ACCESS section
inside metadata.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This patch:
1) Adds the definition of the metadata interface
to the header file. The functions that were exposed
for no good reason are now hidden.
2) Previously exposed functions and their descriptions
are removed from the public header and placed into
the source code for now.
3) The function that reads the config file no longer
tries to close file in case of error.
4) Lines collection is still passed in into the reading
function but as a collection itself not as a pointer
to it.
5) All the parts related to processing lines are currently
ifdefed using HAVE_VALIDATION that is currently is not defined.
This is done to disable creation of the lines collection
utill it is actually needed. I did not want to blindly remove
it though and loose already done work that will be useful
in future.
6) Version of the library and interface is updated
7) New header and source modules are introduced to hold functions
related to the meta data. They are mostly stubbed out.
This is incomplete patch. It builds and make check runs.
It is created just to simplify the review a bit.
|
|
|
|
| |
Additional changes.
|
| |
|
| |
|
| |
|
| |
|
|
|
|
|
|
| |
The null-terminator would have been written one byte past the end
of the array (and there may have been an extra garbage character
in the index before it)
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
The original implementation was compressing the list,
throwing away empty strings.
The function that did that was pretty brain damaging.
I cleaned it up and adjusted so that it could return
list with empty values and without them.
The old function was turned into a wrapper and a new
high level function was intorduced to provide
ability to get both empty and non empty strings.
|
|
|
|
|
|
| |
The previous patch included "config.h"
into the public header which caused all sorts
of redefinitions and warnings.
|
|
|
|
|
|
|
| |
Added configurable key length.
Changed comments for the functions that
are currently not used and reserved for
future functionality.
|
|
|
|
|
|
|
| |
Patch adds ability to read
configuration using already open
file descriptor.
Started by Steve G and refined a bit by me.
|
|
|
|
| |
Necessary for RPM builds on RHEL5
|
|
|
|
|
|
|
|
| |
Inspired by issue #173 I reviewed the
other function of the interface and
found a lot of problems with its
error handling.
Also made INI use collection public interfaces.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This patch creates the infrastructure for
logging of the event from the top of the interface
to the bottom. It is a start. A lot of functionality
is left aside.
The attempt of this patch is pass event from caller
of the ELAPI interface via targets to sinks
then to providers and do serialization creating
entity that is ready to be written to a file.
It also implements more specific provider related
configuration parameters.
Also it addresses couple suggestions that were
brought up against previous patch.
ELAPI Correcting issues
This patch addresses the issues found during the
review of the previous patches and addresses
ticket #166.
|
|
|
|
|
| |
- replaced mailing list address
- let sssd base components read version from VERSION
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Started looking at the ticket #107 related to
traverse functions. Realized that the return values
are not consistent. That ovelapped with the work
that I wanted to do for ticket #103 - errno cleanup.
So I (across collection, INI and ELAPI):
* Made the return codes consistent (where found)
* Removed errno where it is not needed
While was testing used valgrind and found a nasty
problem when the value was added to collection with
overwriting duplicates the count was decreased improperly.
Fixing collection.c to not decrease count made
valgrind happy. While I was debugging this
I also spotted several build warnings in trace
statements when the " exp ? v1 : v2 " was used.
Fixed those.
In ini_config.c there was a trace stament that used
variable after it was freed. Removed trace stament.
|
|
|
|
| |
Also fixed a typo in the header file.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
I started to cleanup the unit tests from the type cust around NULL and
found several problems that I had to address:
1) The choice of the "." as a search separator
turned out to be a poor choice.
The problem was that the file name has "."
and INI was relaying on files to be used as
property names.
I corrected that part in the INI but after discussion
with Simo we decided to switch from "." to "!"
as special symbol anyways.
2) Found that the property rename was not
reinitializing the hash. Corrected.
Added ticket to add unit tests around it (#83).
|
|
|
|
|
| |
Also adds support for running ini_config_ut under a parallel build
directory.
|
|
|
|
|
|
|
| |
The inefficient function from ini_config.c is
replaced with the function provided now by the
collection interface. Also the unit test is updated
to run from the local directory.
|
|
|
|
|
|
| |
Previously, only the SSSD server components were being built with
compile-time warnings enabled. This patch will ensure that all
components in common and sss_client are also built the same way.
|
|
|
|
| |
Fix consts and function declarations
|
|
|
|
| |
See ticket #37 in sssd track.
|
|
|
|
|
|
|
| |
Patch prepends prefix "col_" to all
functions related to collection.
This caused some formatiing issues
so the alignement was addressed too.
|
|
|
|
|
|
|
|
|
| |
If automake 1.11 or higher is available, use of the configure
option '--enable-silent-rules' will suppress most build messages
and replace them with lines such as
CC sysdb.o
This affects libtool as well.
|