GCC Built-in functions

[Rahul Patel]

Recently, come across usage of Built-in functions.

https://gcc.gnu.org/onlinedocs/gcc/Other-Builtins.html

Surprisingly, Replacing below piece of code with usage of Built-in functions has reduced elf size by half.

 

#define FAIL_UNLESS(fmt, value, cmp, expect, description...) \

  do { \

    __typeof(0 ? (value) : (expect)) _m_value = (value); \

    __typeof(0 ? (value) : (expect)) _m_expect = (expect); \

    _testCondition(fmt,value,expect,cmp);\

    if(strcmp((char *)fmt,"%s")==0)\

    {\

      (strcmp((char *)_m_value,(char *)_m_expect) cmp 0) \

        ? _tHarnessTestCondition(1, __FILENAME__, __FUNCTION__, __LINE__,description) \

        : _tHarnessTestCondition(0, __FILENAME__, __FUNCTION__, __LINE__,description);\

    }\

    else\

    {\

      ((_m_value) cmp (_m_expect)) \

          ? _tHarnessTestCondition(1, __FILENAME__, __FUNCTION__, __LINE__,description) \

          : _tHarnessTestCondition(0, __FILENAME__, __FUNCTION__, __LINE__,description);\

    }\

  } while (0)

  

  

#define FAIL_UNLESS(fmt, value, cmp, expect, ...)                                     \

  do {                                                                                \

    __typeof(0 ? (value) : (expect)) _m_value  = (value);                             \

    __typeof(0 ? (value) : (expect)) _m_expect = (expect);                            \

    _testCondition(fmt, value, expect, cmp);                                          \

    bool _m_result =                                                                  \

        /* Don't include code to process strings if the argument isn't a string. */   \

        __builtin_choose_expr(                                                        \

            __builtin_types_compatible_p(__typeof(_m_value), const char*),            \

            /* With __builtin_strcmp the compiler should optimize out the check. */   \

            ((__builtin_strcmp((const char*)fmt, "%s") == 0)                          \

                /* The double-typecasts are in case the value can't be converted  */  \

                /* directly to a pointer (e.g. float). Conversion from float to   */  \

                /* pointer causes an error even if this expression is not chosen. */  \

                /* In the final code _m_value is always compatible with (char*).  */  \

                ? (strcmp((const char*)(uint32_t)_m_value,                            \

                          (const char*)(uint32_t)_m_expect) cmp 0)                    \

                : ((_m_value) cmp (_m_expect))),                                      \

            ((_m_value) cmp (_m_expect)));                                            \

    _tHarnessTestCondition(_m_result,__FILENAME__,__func__,__LINE__,__VA_ARGS__);     \

  } while (0)

-- Rahul Patel

[Vishal Thanki]

Good to know this. I wonder how strcmp can be optimized further ;-) 

--
-- Owner of this group do not accept ANY liability for the views expressed in the email or for the consequences of any computer viruses that may be transmitted with this email. Owner is not responsible for any kind of copyright violations or infringements caused/posted by any user of the group. Messages are posted through moderation also does not include any kind of verification against copyright violations or infringements.
---
You received this message because you are subscribed to the Google Groups "insane" group.
To unsubscribe from this group and stop receiving emails from it, send an email to fundamental-discussion+unsub...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Viral Mehta]

Ohhh ! Our group is still active. I am missing them under Forum tabs.

And believe me, I dont understand a bit of this code :(

To unsubscribe from this group and stop receiving emails from it, send an email to fundamental-discussion+unsubscri...@googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

--
-- Owner of this group do not accept ANY liability for the views expressed in the email or for the consequences of any computer viruses that may be transmitted with this email. Owner is not responsible for any kind of copyright violations or infringements caused/posted by any user of the group. Messages are posted through moderation also does not include any kind of verification against copyright violations or infringements.
---
You received this message because you are subscribed to the Google Groups "insane" group.
To unsubscribe from this group and stop receiving emails from it, send an email to fundamental-discussion+unsub...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

--

Thanks,
Viral Mehta

[Rahul Patel]

Viral, you are beyond such code. you focus on complex code which can be patented. :)

--

Thanks,
Viral Mehta

--
-- Owner of this group do not accept ANY liability for the views expressed in the email or for the consequences of any computer viruses that may be transmitted with this email. Owner is not responsible for any kind of copyright violations or infringements caused/posted by any user of the group. Messages are posted through moderation also does not include any kind of verification against copyright violations or infringements.
---
You received this message because you are subscribed to the Google Groups "insane" group.
To unsubscribe from this group and stop receiving emails from it, send an email to fundamental-discussion+unsub...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

--

Rahul Patel

[Tanmay Upadhyay]

😀👍

[Viral Mehta]

Ha ha ! That was a joke I know ;)

Thanks,
Viral Mehta