Regarding sssd.conf syntax check, going thru dinglib

[amit kumar]

Hello,

*Present **Behavior*:
# vim  /usr/local/etc/sssd/sssd.conf
[sssd]
services = nss, pam
config_file_version = 2
domains = LDAP

[domain/LDAP]
ldap_search_base = dc=example,dc=com
id_provider = ldap
*auth_provider = ldap9001    **<== '**sssctl config_check' does not
reports this    1*
ldap_uri = ldap://server.example.com
ldap_id_use_start_tls = True
ldap_tls_cacert = /etc/openldap/certs/cacert.asc
*debug_level = tt        **<== 'sssctl config_check' does not reports
this    2
**klala =  1                <== '**sssctl config_check' r**eports
thi*s        3

*My **Interpretation*:
sss_ini_call_validators_errobj()    //sssd/src/util/sss_ini.c
    ini_rules_read_from_file(rules_path, &rules_cfgobj);       
//rules_path=/usr/share/sssd/cfg_rules.ini        {All _keywords on left
side of __assignment__are rules_ which are _read from cfg_rules.ini_
fills in struct ini_cfgobj}

*Why sssd does**reports 3*?    Because rule is not present in cfg_rules.ini
*Why sssd doe**s not report 1,2*?    May be
    - Because there is no such check in ding-lib about values of options.
    - OR Check is broken.    I also find
# cat /root/ding-libs-0.6.0/ini/ini.d/mysssd.conf
[service]
# Options available to all services
debug_level = int, None, false                                    <=Now
what's this syntax. If it takes int, then is this broken.

Please throw some light...

[Lukas Slebodnik]

On (29/03/17 19:13), amit kumar wrote:
> Hello,
>
> *Present **Behavior*:
> # vim  /usr/local/etc/sssd/sssd.conf
> [sssd]
> services = nss, pam
> config_file_version = 2
> domains = LDAP
>
> [domain/LDAP]
> ldap_search_base = dc=example,dc=com
> id_provider = ldap
> *auth_provider = ldap9001    **<== '**sssctl config_check' does not

^^^^^^^^^^^^^^
ATM it reports just an invalid option name
and "auth_provider" is valid.

Validating values need to be implemented.
I think we have ticket somewhere.

LS

[amit kumar]

Hello Lukas,

Thanks for response, yes we have ticket
https://pagure.io/SSSD/sssd/issue/416

But my query was regarding the design *how we parse smb.conf using
ding-lib.*

I am planing to provide a fix so that 'sssctl config-check' reports
something as this incorrect.
debug_level = uu
I found anything on RHS of = inside smb.conf is not validated?
Do we use # cat /root/ding-libs-0.6.0/ini/ini.d/mysssd.conf for
validation of contents, if yes {How}.

Hope I am clear with my question.

Many Thanks in Advance
Amit

On 03/29/2017 10:12 PM, Lukas Slebodnik wrote:
> On (29/03/17 19:13), amit kumar wrote:
> > Hello,
> >
> > *Present **Behavior*:
> > # vim  /usr/local/etc/sssd/sssd.conf
> > [sssd]
> > services = nss, pam
> > config_file_version = 2
> > domains = LDAP
> >
> > [domain/LDAP]
> > ldap_search_base = dc=example,dc=com
> > id_provider = ldap
> > *auth_provider = ldap9001    **<== '**sssctl config_check' does not
>
> ^^^^^^^^^^^^^^
> ATM it reports just an invalid option name
> and "auth_provider" is valid.
>
> Validating values need to be implemented.
> I think we have ticket somewhere.
>
> LS
> _______________________________________________
> sssd-devel mailing list -- [e-mail]
> To unsubscribe send an email to [e-mail]

[Lukas Slebodnik]

On (30/03/17 22:00), amit kumar wrote:
> Hello Lukas,
>
> Thanks for response, yes we have ticket
> https://pagure.io/SSSD/sssd/issue/416
>
> But my query was regarding the design *how we parse smb.conf using
> ding-lib.*

Here is a desing page which describe ding-libs part.
https://fedorahosted.org/sssd/wiki/DesignDocs/libini-config-file-checks

I hope it is up to date. It also describe two ding-libs native validators
and their schemas in ini format.

There is also possible register another (external) validators.
But I think checking values would be usefull for everyone.

IIRC Michal Zidek (mzidek at redhat dot com) had some ideas
about schema for checking values.

LS

BTW the usage in sssd was quite trivial
https://pagure.io/SSSD/sssd/c/8b2a31634764168183506925a4b9f461afdba6f3
+ providing rules /usr/share/sssd/cfg_rules.ini