# Issue with processing nested list without having key

**URL:** <https://dmap-community.ductus.global/t/issue-with-processing-nested-list-without-having-key/3232>\
**Category:** YANG\
**Created:** [August 19, 2020, 2:14pm UTC](https://dmap-community.ductus.global/t/issue-with-processing-nested-list-without-having-key/3232 "2020-08-19T14:14:32Z")\
**Posts on this page:** 12\
**Page:** 1

<div class="post-metadata">

**Author:** ![Biswajit84](https://avatars.discourse-cdn.com/v4/letter/b/13edae/32.png) [@Biswajit84](https://dmap-community.ductus.global/u/Biswajit84)\
**Post date:** [August 19, 2020, 2:14pm UTC](https://dmap-community.ductus.global/t/issue-with-processing-nested-list-without-having-key/3232/1 "2020-08-19T14:14:32Z")

</div>

Hi,  
I have a nested list without key as below:

 ![image](https://canada1.discourse-cdn.com/flex010/uploads/ductus_dmap/original/1X/f16c5eba9635bbb52bdb7061f36157b0f4585e7d.png)

I have created a callpoint for active-alarms.  
Currently I am facing two issue:  
1.get element is not getting called  
2.nested list which is affected-objects also not getting called.  
Both doesn’t have key.  
I have followed 5-c\_stats example. for the same.  
If you can review the below code and find the issue then it will really be helpful.

```c
     int fm_status_get_elem(struct confd_trans_ctx *tctx, confd_hkeypath_t * kp)
     {

    confd_value_t v;

     struct confd_datetime eventTime;

    switch( CONFD_GET_XMLTAG(&kp->v[0][0]))
    {

        case o_ran_fm_fault_id:
            {

		alwaysLog("o_ran_fm_fault_id");
                CONFD_SET_UINT16(&v,100);
                confd_data_reply_value(tctx, &v);
            }
            break;
        case o_ran_fm_fault_source:
            {

		alwaysLog("o_ran_fm_fault_source");
                CONFD_SET_STR(&v,"SOURCE");
                confd_data_reply_value(tctx, &v);
            }
            break;
        case o_ran_fm_fault_severity:
            {

		alwaysLog("o_ran_fm_fault_severity");
                CONFD_SET_ENUM_VALUE(&v,1);
                confd_data_reply_value(tctx, &v);
            }
            break;
        case o_ran_fm_is_cleared:
            {

				alwaysLog("o_ran_fm_is_cleared");
                CONFD_SET_BOOL(&v,1);
                confd_data_reply_value(tctx, &v);
            }
            break;
        case o_ran_fm_fault_text:
            {

				alwaysLog("o_ran_fm_fault_text");
                CONFD_SET_STR(&v,"TEXT");
                confd_data_reply_value(tctx, &v);
            }
            break;
        case o_ran_fm_event_time:
            {
				alwaysLog("o_ran_fm_event_time");
				getdatetime(&eventTime);
				CONFD_SET_DATETIME(&v,eventTime);
            }
            break;
    }

    return CONFD_OK;

}

      Int fm_status_get_next(struct confd_trans_ctx *tctx,
        confd_hkeypath_t * kp, long next)
     {
           confd_value_t v[2];
          int pos;
                  alwaysLog("GET NEXT PRINT");

    switch (CONFD_GET_XMLTAG(&kp->v[0][0])) {
        case o_ran_fm_active_alarms:
            alwaysLog("ACTIVE");
            if (next == -1) { /* first call */
                pos = 0;
            } else {
                pos = (int)next;
            }
            if (pos >= 6) { /* We have reached the end of the list*/
                confd_data_reply_next_key(tctx, NULL, -1, -1);
                return CONFD_OK;
            }
        CONFD_SET_UINT16(&v[0],11);
            confd_data_reply_next_key(tctx, v,1, (long)(pos + 1));
            break;
        case o_ran_fm_affected_objects:
            alwaysLog("OBJECT");
            if (next == -1) { /* first call */
                pos = 0;
                confd_data_reply_next_key(tctx, v, 1, 1);
            } else {
                // pos = (int)next;
                confd_data_reply_next_key(tctx, NULL, -1, -1);
            }
 
            break;
       default:
            return CONFD_ERR;

    }

    return CONFD_OK;

}

```

fm\_status\_get\_next callback to get\_next and fm\_status\_get\_elem callback to get\_elem  
and get\_next and passed to confd\_register\_trans\_cb(dctx, &tcb) through tcb object.

Currently, only once the outer list i.e active-alarms of fm\_status\_get\_next is printing.

Regards,  
Biswajit

---

<div class="post-metadata">

**Author:** ![mnovak](https://yyz2.discourse-cdn.com/flex010/user_avatar/dmap-community.ductus.global/mnovak/32/183_2.png) [@mnovak](https://dmap-community.ductus.global/u/mnovak)\
**Post date:** [August 20, 2020, 7:20am UTC](https://dmap-community.ductus.global/t/issue-with-processing-nested-list-without-having-key/3232/2 "2020-08-20T07:20:10Z")

</div>

By list without `key`, do you mean `keyless` lists? (If so see `8-c_stats_no_key`) Example.  
`get_next` and `get_elem` should be passed to `confd_register_data_cb`, not to `confd_register_trans_cb` as you write (you can see it in the example).

---

<div class="post-metadata">

**Author:** ![Biswajit84](https://avatars.discourse-cdn.com/v4/letter/b/13edae/32.png) [@Biswajit84](https://dmap-community.ductus.global/u/Biswajit84)\
**Post date:** [August 20, 2020, 7:28am UTC](https://dmap-community.ductus.global/t/issue-with-processing-nested-list-without-having-key/3232/3 "2020-08-20T07:28:31Z")

</div>

Hi,  
Sorry it was a Typo error.  
Yes,both the outer list as well as inner list doesn’t have a key.

```auto
    #define CALLPOINT_DECL(mib, cp_id, get_elem_cb, get_next_cb,get_object_cb) \
                {.callpoint = CALLPOINT_ID(mib, cp_id), \
                                             .get_elem = get_elem_cb, \
                                             .get_next = get_next_cb, \
                                             .get_object = get_object_cb }

  
     void fm_setup(struct confd_daemon_ctx *dctx, int rsock)
    {
             struct confd_data_cbs dcb[] = {
        CALLPOINT_DECL(o_ran_fm,cb_op_fm,fm_status_get_elem,
                fm_status_get_next,NULL)
                };

      int i;

    /* Register data callbacks */
    for (i = 0; i < sizeof(dcb) / sizeof(struct confd_data_cbs); i++) {
        if (confd_register_data_cb(dctx, &dcb[i]) != CONFD_OK)
            alwaysLog("Failed to register callpoint \"%s\"", dcb[i].callpoint);
       }

    } 

```

As i said earlier, only once the get\_next callback i.e fm\_status\_get\_next is getting called once.  
Not sure why get\_elem callback is not getting called.

Regards,  
Biswajit

---

<div class="post-metadata">

**Author:** ![josephm](https://avatars.discourse-cdn.com/v4/letter/j/a88e57/32.png) [@josephm](https://dmap-community.ductus.global/u/josephm)\
**Post date:** [August 20, 2020, 9:32am UTC](https://dmap-community.ductus.global/t/issue-with-processing-nested-list-without-having-key/3232/4 "2020-08-20T09:32:30Z")

</div>

See guidelines in User Guide - section **Operational data lists without keys** …

Your code returns same key value repeatedly (`11`) for sub-sequent get-next() calls of parent key-less list. This conflicts with requirement that key must allow to retrieve specific list entry values identified by pseudo-key…

Your copy&paste does not show `exists_optional()` implementation - U.G. states that this is needed for servicing key-less lists for list entry existence checks…

Make sure you enable developer log in ConfD, and trace log in confd-lib in your application.  
It usually gives more hints on when specifically something goes wrong or give better traceability -\> lets us see sequence of operations that will help identify what is wrong…

---

<div class="post-metadata">

**Author:** ![Biswajit84](https://avatars.discourse-cdn.com/v4/letter/b/13edae/32.png) [@Biswajit84](https://dmap-community.ductus.global/u/Biswajit84)\
**Post date:** [August 20, 2020, 10:56am UTC](https://dmap-community.ductus.global/t/issue-with-processing-nested-list-without-having-key/3232/5 "2020-08-20T10:56:45Z")

</div>

Hi,  
Thanks for figuring out the mistake. But I am not getting what I am supposed to add in the .exists\_optional call back.

```
   struct aentry *find_ae(confd_hkeypath_t *keypath, struct arpdata *dp)
{
 return (struct aentry *)CONFD_GET_INT64(&keypath->v[1][0]);
}

static int exists(struct confd_trans_ctx *tctx,
              confd_hkeypath_t *keypath)
{
struct aentry *ae = find_ae(keypath, tctx->t_opaque);
if (ae == NULL) {
    confd_data_reply_not_found(tctx);
} else {
    confd_data_reply_found(tctx);
}
return CONFD_OK;
}

```

Can you please help me in modifying the same.

Regards,  
Biswajit

---

<div class="post-metadata">

**Author:** ![josephm](https://avatars.discourse-cdn.com/v4/letter/j/a88e57/32.png) [@josephm](https://dmap-community.ductus.global/u/josephm)\
**Post date:** [August 20, 2020, 11:18am UTC](https://dmap-community.ductus.global/t/issue-with-processing-nested-list-without-having-key/3232/6 "2020-08-20T11:18:06Z")

</div>

`exists_optional()` is used to answer the ConfD’s question: “Does element with this keypath exist?”.

1. Input argument of the callback asks about some element existence (in this case it would be something like `/active-alarm-list{11}`.

2. Your code must decide whether such oper. data exists or not.  
(`confd_data_reply_found()` or `confd_data_reply_not_found()` )  
This is up to your specific code conditions -\> how/where you have oper data etc.

The example code you pasted is from ConfD example, and it looks into internal memory representation of oper. data that example data provider implement.

In this case, you need to replace this code and get needed info from your source of oper. data…

If you tell ConfD that some data exists, it will continue with asking for nested children/leaves etc., if you tell it it does not, it will not continue with callbacks invocation for nested elements…

---

<div class="post-metadata">

**Author:** ![Biswajit84](https://avatars.discourse-cdn.com/v4/letter/b/13edae/32.png) [@Biswajit84](https://dmap-community.ductus.global/u/Biswajit84)\
**Post date:** [August 20, 2020, 8:33pm UTC](https://dmap-community.ductus.global/t/issue-with-processing-nested-list-without-having-key/3232/7 "2020-08-20T20:33:48Z")

</div>

Hi ,  
Thanks a lot for your valuable input. I have added call back to exists\_optional and  
Now I am able to fetch all the values from confd server through netconf-console(`netconf-console --host=10.10.11.101 -port=2023 --get -x /active-alarm-list`).

But still, I am facing few issues like :

1. server application is getting crashed when I the releasing memory call back of finish\_data ie in fm\_finish\_data function.

```auto
free(): invalid pointer
Aborted

```

The developer log shows :

1. `20-Aug-2020::19:36:53.830 FXT01200014 confd[3806]: devel-c close_trans/"Operational" error {external, ""}`

2. The callback to exists\_optional is not getting called.

Not sure if Issues 1 & 2 are interlinked or not.

Please find the complete updated code below :

```auto
     typedef struct alarmstatus
      {
        int alarmid;
       }activeAlarm;

    static int fm_init_data(struct confd_trans_ctx *tctx )
    {
    alwaysLog("INSIDE fm_init_data");
    activeAlarm *dp ;
    if ((dp = malloc(2 *sizeof(activeAlarm))) == NULL)
        return CONFD_ERR;
    memset(dp, 0, 2 * sizeof(activeAlarm));
    tctx->t_opaque =dp;
    return CONFD_OK;
    }

    static int fm_status_get_next(struct confd_trans_ctx *tctx,
        confd_hkeypath_t * kp, long next)
    {
    confd_value_t v[2];
    int pos=0;
    alwaysLog("GET NEXT PRINT");
    activeAlarm *dp=tctx->t_opaque;
    if (next == -1) { /* first call */
        pos = 0;
        alwaysLog(" *****FIRST CALL*****");
    } else {
        pos = (int)next;
        alwaysLog("Inside else");
    }
    if (pos >= 2) { /* We have reached the end of the list*/
        alwaysLog(" ****** POS:%d",pos);
        confd_data_reply_next_key(tctx, NULL, -1, -1);
        return CONFD_OK;
    }
        CONFD_SET_INT64(&v[0],(int64_t )dp);
        confd_data_reply_next_key(tctx, &v[0],1, (long)(pos +1));
        return CONFD_OK;

      }
     int fm_status_get_elem(struct confd_trans_ctx *tctx, confd_hkeypath_t * kp)
     {
       confd_value_t v;
     struct confd_datetime eventTime;
     alwaysLog("Inside fm_status_get_elem");
      activeAlarm *ae = find_ae(kp, tctx->t_opaque);
    if (ae == NULL ) {
        alwaysLog("GET ELE NOT FOUND");
        confd_data_reply_not_found(tctx);
        return CONFD_OK;
      }

    switch( CONFD_GET_XMLTAG(&kp->v[0][0]))
    {
        case o_ran_fm_fault_id:
            {
           alwaysLog("o_ran_fm_fault_id");
                CONFD_SET_UINT16(&v,100);
            }
            break;
           case o_ran_fm_fault_source:
            {

           alwaysLog("o_ran_fm_fault_source");
                CONFD_SET_STR(&v,"SOURCE");
            }
            break;
        case o_ran_fm_fault_severity:
            {

           alwaysLog("o_ran_fm_fault_severity");
                CONFD_SET_ENUM_VALUE(&v,1);
            }
            break;
        case o_ran_fm_is_cleared:
            {

           alwaysLog("o_ran_fm_is_cleared");
                CONFD_SET_BOOL(&v,1);
            }
            break;
        case o_ran_fm_fault_text:
            {

           alwaysLog("o_ran_fm_fault_text");
                CONFD_SET_STR(&v,"TEXT");
            }
            break;
        case o_ran_fm_event_time:
            {
           alwaysLog("o_ran_fm_event_time");
                    getdatetime(&eventTime);
      CONFD_SET_DATETIME(&v,eventTime);
            }
            break;
        case o_ran_fm_name:
            alwaysLog("o_ran_fm_name");
                CONFD_SET_STR(&v,"NAME");
            break;
      }
    confd_data_reply_value(tctx, &v);
    return CONFD_OK;

    }

     static int exists(struct confd_trans_ctx *tctx,confd_hkeypath_t * keypath)
    {
       alwaysLog("INSIDE EXISTS");
     activeAlarm *ae = find_ae(keypath, tctx->t_opaque);
    if (ae == NULL ) {
        alwaysLog("NOT FOUND");
        confd_data_reply_not_found(tctx);
    } else {
        confd_data_reply_found(tctx);
        alwaysLog(" FOUND");
    }
    return CONFD_OK;

     }

    activeAlarm *find_ae(confd_hkeypath_t *keypath, activeAlarm *dp)
    {
        return (activeAlarm *)(CONFD_GET_INT64(&keypath->v[1][0]));
    }

    static int fm_finish_data(struct confd_trans_ctx *tctx)
    {
    alwaysLog("Inside fm_finish_data");
    activeAlarm *dp = tctx->t_opaque;
    if(dp != NULL)
     free(dp);
    return CONFD_OK;
    }

    #define CALLPOINT_DECL(mib, cp_id, get_elem_cb, 
    get_next_cb,get_object_cb,exists_opt_cb) \
                {.callpoint = CALLPOINT_ID(mib, cp_id), \
                                             .get_elem = get_elem_cb, \
                                             .get_next = get_next_cb, \
                                             .get_object = get_object_cb, \
                                             .exists_optional = exists_opt_cb }

    void fm_setup(struct confd_daemon_ctx *dctx, int rsock)
    {
     struct confd_data_cbs dcb[] = {
        CALLPOINT_DECL(o_ran_fm,cb_op_fm,fm_status_get_elem,fm_status_get_next,NULL,exists)
    };

    int i;

    /* Register data callbacks */
    for (i = 0; i < sizeof(dcb) / sizeof(struct confd_data_cbs); i++) {
        if (confd_register_data_cb(dctx, &dcb[i]) != CONFD_OK)
            alwaysLog("Failed to register callpoint \"%s\"", dcb[i].callpoint);
     }

     }

```

---

<div class="post-metadata">

**Author:** ![josephm](https://avatars.discourse-cdn.com/v4/letter/j/a88e57/32.png) [@josephm](https://dmap-community.ductus.global/u/josephm)\
**Post date:** [August 21, 2020, 8:27am UTC](https://dmap-community.ductus.global/t/issue-with-processing-nested-list-without-having-key/3232/8 "2020-08-21T08:27:47Z")

</div>

1. Looks like double free or un-intialized variable problem.  
Make sure that you don’t invoke e.g. finish function explicitly yourself (except for passing it to transaction callback).  
Regular approach → see ConfD developer log, confd-lib trace logs in app, and try to spot the suspect.  
(e.g. can you see `init()` invoked before `finish()` for the transaction? are there any errors preceding the “free()” problem etc.)  
If problem still evades identification, you can try to run data provider app with [Valgrind](https://valgrind.org/) to see some more/less hidden error/risk areas… this is my favorite helper for debugging C apps 🙂

2. this is most probably just related to/caused by 1.

3. if you mean after observing 1. / 2., then again most probably yes → is a consequence → you can check whether callbacks are still registered at this point due to previous errors e.g. via `confd --status`

---

<div class="post-metadata">

**Author:** ![Biswajit84](https://avatars.discourse-cdn.com/v4/letter/b/13edae/32.png) [@Biswajit84](https://dmap-community.ductus.global/u/Biswajit84)\
**Post date:** [August 21, 2020, 10:58am UTC](https://dmap-community.ductus.global/t/issue-with-processing-nested-list-without-having-key/3232/9 "2020-08-21T10:58:03Z")

</div>

Hi,  
I am freeing the memory in finish\_data function only which is a callback.I am not explicitly calling the function.  
Also i have seen the confd --status and i could see .exists\_optional callback is registered(id=cb\_op\_fm daemonId=18 daemonName=Operational callbacks=get\_next,get\_elem,exists\_optional).

I have tried to print the address holding by dp(pointer to structure) after dynamically allocated memory as below:

```
 static int fm_init_data(struct confd_trans_ctx *tctx )
{
 activeAlarm *dp ;
if ((dp = malloc(2 *sizeof(activeAlarm))) == NULL)
    return CONFD_ERR;
    alwaysLog("ADDRESS:%p",dp);
memset(dp, 0, 2* sizeof(activeAlarm));
  free(dp);
  return CONFD_ERR;
 }

  tctx->t_opaque =dp;
return CONFD_OK;

}

```

Similarly, I try to print in get\_next and get\_elem call back function also. As below:

```
 activeAlarm *dp=tctx->t_opaque;
  alwaysLog(" GET NEXT ADDRESS:%p",dp);

```

I found the value collected from init\_data is different then other.  
ADDRESS:0x4c33d0 —\>message from fm\_init\_data  
GET NEXT ADDRESS:0x4836e8  
GET ELEM ADDRESS:0x4836e8  
FINISH ADDRESS: 0x4836e8

I am not sure if I am missing anything.  
devel.log looks OK .  
TRACE LOG:

TRACE CALL trans init(thandle=5868,mode=“r”,db=running)2020-08-21 10:50:30 -6- as-confd[11676] 62151.385261 1.461278 oran-fm.c:220 ADDRESS:0x4c33d0  
TRACE CALL data get\_next(thandle=5868, /active-alarm-list/active-alarms, -1)2020-08-21 10:50:30 -6- as-confd[11676] 62151.732625 0.334174 oran-fm.c:239 GET NEXT ADDRESS:0x4836e8  
–\> CONFD\_OK  
TRACE CALL data get\_elem(thandle=5868,/active-alarm-list/active-alarms{4732648}/fault-id) (100) --\> CONFD\_OK  
TRACE CALL data get\_elem(thandle=5868,/active-alarm-list/active-alarms{4732648}/fault-source) (“SOURCE”) --\> CONFD\_OK  
TRACE CALL data get\_next(thandle=5868, /active-alarm-list/active-alarms{4732648}/affected-objects, -1)2020-08-21 10:50:30 -6- as-confd[11676] 62151.735993 0.003368 oran-fm.c:239 GET NEXT ADDRESS:0x4836e8  
–\> CONFD\_OK  
TRACE CALL data get\_elem(thandle=5868,/active-alarm-list/active-alarms{4732648}/affected-objects{4732648}/name) (“NAME”) --\> CONFD\_OK  
TRACE CALL data get\_next(thandle=5868, /active-alarm-list/active-alarms{4732648}/affected-objects, 1)2020-08-21 10:50:30 -6- as-confd[11676] 62151.736844 0.000853 oran-fm.c:239 GET NEXT ADDRESS:0x4836e8  
–\> CONFD\_OK  
TRACE CALL data get\_elem(thandle=5868,/active-alarm-list/active-alarms{4732648}/fault-severity) (enum\<1\>) --\> CONFD\_OK  
TRACE CALL data get\_elem(thandle=5868,/active-alarm-list/active-alarms{4732648}/is-cleared) (true) --\> CONFD\_OK  
TRACE CALL data get\_elem(thandle=5868,/active-alarm-list/active-alarms{4732648}/fault-text) (“TEXT”) --\> CONFD\_OK  
TRACE CALL data get\_elem(thandle=5868,/active-alarm-list/active-alarms{4732648}/event-time) (2020-08-21T10:50:30.610692+00:00) --\> CONFD\_OK  
TRACE CALL data get\_next(thandle=5868, /active-alarm-list/active-alarms, 1)2020-08-21 10:50:30 -6- as-confd[11676] 62151.738953 0.002108 oran-fm.c:239 GET NEXT ADDRESS:0x4836e8

TRACE CALL data get\_elem(thandle=5868,/active-alarm-list/active-alarms{4732648}/fault-id) (100) --\> CONFD\_OK  
TRACE CALL data get\_elem(thandle=5868,/active-alarm-list/active-alarms{4732648}/fault-source) (“SOURCE”) --\> CONFD\_OK  
TRACE CALL data get\_next(thandle=5868, /active-alarm-list/active-alarms{4732648}/affected-objects, -1)2020-08-21 10:50:30 -6- as-confd[11676] 62151.740280 0.001306 oran-fm.c:239 GET NEXT ADDRESS:0x4836e8  
–\> CONFD\_OK  
TRACE CALL data get\_elem(thandle=5868,/active-alarm-list/active-alarms{4732648}/affected-objects{4732648}/name) (“NAME”) --\> CONFD\_OK  
TRACE CALL data get\_next(thandle=5868, /active-alarm-list/active-alarms{4732648}/affected-objects, 1)2020-08-21 10:50:30 -6- as-confd[11676] 62151.741128 0.000849 oran-fm.c:239 GET NEXT ADDRESS:0x4836e8  
–\> CONFD\_OK  
TRACE CALL data get\_elem(thandle=5868,/active-alarm-list/active-alarms{4732648}/fault-severity) (enum\<1\>) --\> CONFD\_OK  
TRACE CALL data get\_elem(thandle=5868,/active-alarm-list/active-alarms{4732648}/is-cleared) (true) --\> CONFD\_OK  
TRACE CALL data get\_elem(thandle=5868,/active-alarm-list/active-alarms{4732648}/fault-text) (“TEXT”) --\> CONFD\_OK  
TRACE CALL data get\_elem(thandle=5868,/active-alarm-list/active-alarms{4732648}/event-time) (2020-08-21T10:50:30.615115+00:00) --\> CONFD\_OK  
TRACE CALL data get\_next(thandle=5868, /active-alarm-list/active-alarms, 2)2020-08-21 10:50:30 -6- as-confd[11676] 62151.743273 0.002145 oran-fm.c:239 GET NEXT ADDRESS:0x4836e8  
–\> CONFD\_OK  
TRACE CALL trans finish(thandle=5868)2020-08-21 10:50:30 -6- as-confd[11676] 62151.747292 0.003992 oran-soft-mgm.c:187 s\_finish called  
2020-08-21 10:50:30 -6- as-confd[11676] 62151.747434 0.000144 oran-fm.c:122 Inside fm\_finish\_data  
2020-08-21 10:50:30 -6- as-confd[11676] 62151.747455 0.000020 oran-fm.c:124 FINISH ADDRESS: 0x4836e8  
free(): invalid pointer

If you can help me in figuring out the root cause it will be really helpful.

Regards,  
Biswajit

---

<div class="post-metadata">

**Author:** ![josephm](https://avatars.discourse-cdn.com/v4/letter/j/a88e57/32.png) [@josephm](https://dmap-community.ductus.global/u/josephm)\
**Post date:** [August 21, 2020, 12:20pm UTC](https://dmap-community.ductus.global/t/issue-with-processing-nested-list-without-having-key/3232/10 "2020-08-21T12:20:36Z")

</div>

unfortunately we cannot debug/finetune custom C applications here, rather give hints/tips on confd (confd-lib)…

e.g. not sure if copy&paste error, but the code you pasted in last comment calls free() right after malloc, and thus leads to double free C issue in the end… this line was not there in previous code printout…

i can give pointers no problem for existing examples in confd, or with minor updates/extensions of those, but not for full custom implementations…

> TRACE CALL data get\_elem(thandle=5868,/active-alarm-list/active-alarms{4732648}/affected-objects{4732648}/name)

There is same key value in both lists on keypath (both `active-alams{}` and `affected-object{}`).

This seems strange / possibly wrong, assuming you don’t use same memory to identify oper. data of two different lists (resp. their rows - one parent, one child)… again, this depends on how you want to represent/manage your memory/data in the “backend”… Your pasted code of get\_next() does not take input keypath into account, and does not act “differently” depending on which list ConfD is asking about…

---

<div class="post-metadata">

**Author:** ![Biswajit84](https://avatars.discourse-cdn.com/v4/letter/b/13edae/32.png) [@Biswajit84](https://dmap-community.ductus.global/u/Biswajit84)\
**Post date:** [August 21, 2020, 2:16pm UTC](https://dmap-community.ductus.global/t/issue-with-processing-nested-list-without-having-key/3232/11 "2020-08-21T14:16:08Z")

</div>

Hi Joseph,  
Yes correct.I have provided same key for both outer and inner.  
Can you please help me in providing different key for both outer list and inner list.

Because in inner list when I am trying to set key in get\_next :  
CONFD\_SET\_STR(&v[0],“OBJ”);  
Then its not working.

Also I am following the way linuxcfg application is following to subscribe init\_data,finish\_data.

```
 static int init_data(struct confd_trans_ctx *tctx)
{
      int i, ret = CONFD_OK;

confd_trans_set_fd(tctx, opworkersock.fd);

for (i = 0; i < num_components && ret == CONFD_OK; i++) {
    if (components[i]->init_data != NULL)
        ret = (*components[i]->init_data)(tctx);
}

if (ret != CONFD_OK) {
    /* Need to call finish for successfully called inits */
    for (i -= 2; i >= 0; i--) {
        if (components[i]->finish_data != NULL)
            (*components[i]->finish_data)(tctx);
    }
}
return ret;
}

static int finish_data(struct confd_trans_ctx *tctx)
{
int i;

for (i = num_components - 1; i >= 0; i--) {
    if (components[i]->finish_data != NULL)
        (*components[i]->finish_data)(tctx);
}
return CONFD_OK;
}

static void register_transpoints()
{

struct confd_trans_cbs tcb;
memset(&tcb, 0, sizeof(tcb));
tcb.init = init_data;
tcb.finish = finish_data;
confd_register_trans_cb(dctx, &tcb);
}

```

So i hope, the subscription part is not having any issue.Please review this part also.

Regards,  
Biswajit

Regards,  
Biswajit

---

<div class="post-metadata">

**Author:** ![mnovak](https://yyz2.discourse-cdn.com/flex010/user_avatar/dmap-community.ductus.global/mnovak/32/183_2.png) [@mnovak](https://dmap-community.ductus.global/u/mnovak)\
**Post date:** [August 25, 2020, 8:06am UTC](https://dmap-community.ductus.global/t/issue-with-processing-nested-list-without-having-key/3232/12 "2020-08-25T08:06:38Z")

</div>

```auto
 CONFD_SET_INT64(&v[0],(int64_t )dp);
 confd_data_reply_next_key(tctx, &v[0],1, (long)(pos +1));

```

Is this correct? Do you change `dp` (which is `tctx->t_opaque`) in next `get_next` call?

I suggest you create very simple application (e.g. based on `examples.confd/intro/8-c_stats_no_key/arpstat.c`) and see how it behaves (perhaps by adding `traceh.h` - see e.g. `examples.confd/intro/11-c_hooks/hooks.c` how `examples.confd/include/traceh.h` is used.