[asterisk-dev] [Code Review] 4533: clang compiler warning: -Wtautological-compare

rmudgett reviewboard at asterisk.org
Tue Mar 31 20:16:32 CDT 2015


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviewboard.asterisk.org/r/4533/#review15003
-----------------------------------------------------------



/branches/13/channels/chan_skinny.c
<https://reviewboard.asterisk.org/r/4533/#comment25666>

    Since one of the definitions of letohl() defined above could return a signed value, it is best to assign the returned value to an unsigned variable and then compare.
    
    unsigned long len;
    
    len = letohl(req->len)
    if (SKINNY_MAX_PACKET < len) {
    }



/branches/13/funcs/func_curl.c
<https://reviewboard.asterisk.org/r/4533/#comment25672>

    Try defining this as:
    #define CURLOPT_SPECIAL_HASHCOMPAT ((CURLoption) -500)
    
    This should shut-up the compiler without changing the binary value.



/branches/13/include/asterisk/app.h
<https://reviewboard.asterisk.org/r/4533/#comment25650>

    This changes ABI by changing the other enum values.
    
    Change to:
    AST_RECORD_IF_EXISTS_ERROR = -1,
    
    This takes the place of plain -1 when it is used in conjunction with this enum.



/branches/13/include/asterisk/cel.h
<https://reviewboard.asterisk.org/r/4533/#comment25652>

    You are missing two values:
    AST_CEL_INVALID_VALUE = -1,
    AST_CEL_ALL = 0,
    
    Use AST_CEL_INVALID_VALUE in place of AST_CEL_SENTINEL.



/branches/13/include/asterisk/utils.h
<https://reviewboard.asterisk.org/r/4533/#comment25654>

    You should be able to do this cast change unconditionally.



/branches/13/include/asterisk/utils.h
<https://reviewboard.asterisk.org/r/4533/#comment25653>

    Put parens around the use of v to cast what you think you are casting.



/branches/13/main/app.c
<https://reviewboard.asterisk.org/r/4533/#comment25660>

    This should never happen so replace with an
    ast_assert(0);



/branches/13/main/cel.c
<https://reviewboard.asterisk.org/r/4533/#comment25656>

    change 0 to AST_CEL_ALL



/branches/13/main/cel.c
<https://reviewboard.asterisk.org/r/4533/#comment25655>

    You have now duplicated this error message in ast_cel_str_to_event_type().



/branches/13/res/res_pjsip_exten_state.c
<https://reviewboard.asterisk.org/r/4533/#comment25658>

    Guidelines: No declarations within code area.  Only at beginning of a statement block.



/branches/13/res/res_pjsip_exten_state.c
<https://reviewboard.asterisk.org/r/4533/#comment25659>

    Pull the assignment within the test out.
    
    presence_state = ast_hint_presence_state();
    if (presence_state == -1) {
    }
    
    I have a feeling that presence_state should be checked for AST_PRESENCE_INVALID as an error in addition to -1.  However, I'm not sure.


- rmudgett


On March 30, 2015, 6:08 a.m., Diederik de Groot wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviewboard.asterisk.org/r/4533/
> -----------------------------------------------------------
> 
> (Updated March 30, 2015, 6:08 a.m.)
> 
> 
> Review request for Asterisk Developers.
> 
> 
> Bugs: ASTERISK-24917
>     https://issues.asterisk.org/jira/browse/ASTERISK-24917
> 
> 
> Repository: Asterisk
> 
> 
> Description
> -------
> 
> clang's static analyzer will throw quite a number warnings / errors during compilation, some of which can be very helpfull in finding corner-case bugs
> 
> clang compiler warning:-Wtautological-compare
> 
> Changes:
> /branches/13/channels/pjsip/dialplan_functions.c:869
> len is of type size_t, which is unsigned. It will not be able to hold a value < 0
> 
> /branches/13/funcs/func_curl.c:174
> Not a 100% sure how to do this correctly. But assiging a negative value is problematic. Extending the enum in curl/curl.h is not possible either. I opted to use the enum last entry (CURL_LAST) which is currently not used for any thing. Another option would be to use one of the OBSOLETE VALUES like 16. Neither way is very nice though.
> 
> /branches/13/include/asterisk/app.h:988
> Needed to convey the error state returned by res/res_stasis_recording.c:stasis_app_recording_if_exists_parse
> 
> /branches/13/include/asterisk/cel.h:77
> Added to convey not-found or error state. Not sure which name would be prefered for such an enum value.
> 
> /branches/13/main/cel.c:536
> Return actual enum instead of -1,l which can not be conveyed by this enum.
> 
> /branches/13/main/enum.c:260
> dn_expand return signed int
> 
> /branches/13/main/event.c:202
> enum type cannot be < 0
> 
> /branches/13/main/indications.c:362
> tone_data.freq1 and freq2 are unsigned int's so no need to check if < 0. Not sure what should happend when freq1 / freq2 are 0 already... (needs recheck by source owner)
> 
> /branches/13/main/presencestate.c:293
> Should use the actual enum value for INVALID State
> 
> /branches/13/main/security_events.c:432/890/1176/
> enum event_type cannot be <0
> 
> /branches/13/main/udptl.c:365/649/661
> encode_length returns and unsigned int, so checking if < 0 does not make sence. Not 100% if encode_length has side effects, so left the actual call the this function in place. (Needs to be rechecked by code-owner)
> 
> /branches/13/res/res_pjsip_exten_state.c:411
> Used a temporary int variable to be able to check the return value from ast_hint_presence_state.. Not very nice, but did not want to change the signature of this function.
> 
> /branches/13/res/res_stasis_playback.c:634
> operation is enum and cannot be < 0
> 
> /branches/13/res/res_stasis_recording.c:598/604
> recording->state / operation is enum and cannot be < 0
> 
> /branches/13/res/res_security_log.c:992
> enum event_type cannot be <0
> 
> use of ARRAY_IN_BOUNDS with enum results in tautological comparison for the lower limit which is always true.
> 
> 
> Diffs
> -----
> 
>   /branches/13/res/res_stasis_recording.c 433444 
>   /branches/13/res/res_stasis_playback.c 433444 
>   /branches/13/res/res_pjsip_exten_state.c 433444 
>   /branches/13/res/ari/resource_channels.c 433444 
>   /branches/13/res/ari/resource_bridges.c 433444 
>   /branches/13/main/udptl.c 433444 
>   /branches/13/main/security_events.c 433444 
>   /branches/13/main/presencestate.c 433444 
>   /branches/13/main/indications.c 433444 
>   /branches/13/main/event.c 433444 
>   /branches/13/main/enum.c 433444 
>   /branches/13/main/cel.c 433444 
>   /branches/13/main/app.c 433444 
>   /branches/13/include/asterisk/utils.h 433444 
>   /branches/13/include/asterisk/logger.h 433444 
>   /branches/13/include/asterisk/cel.h 433444 
>   /branches/13/include/asterisk/app.h 433444 
>   /branches/13/funcs/func_curl.c 433444 
>   /branches/13/channels/pjsip/dialplan_functions.c 433444 
>   /branches/13/channels/chan_skinny.c 433444 
> 
> Diff: https://reviewboard.asterisk.org/r/4533/diff/
> 
> 
> Testing
> -------
> 
> 
> Thanks,
> 
> Diederik de Groot
> 
>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.digium.com/pipermail/asterisk-dev/attachments/20150401/9b0f0ebf/attachment-0001.html>


More information about the asterisk-dev mailing list