Bug#1086753: spandsp: FTBFS on armhf and armel

[Dylan Aïssi]

Source: spandsp
Version: 0.0.6+dfsg-2.1
Severity: serious
Tags: ftbfs
Justification: fails to build from source (but built successfully in the past)

During the recent rebuild of packages to enable PAC and BTI support on arm64,
spandsp fails to build on armel [1] and armel [2] with errors like:

v17rx.c:1265:40: error: passing argument 1 of
'vec_circular_dot_prodi16' from incompatible pointer type
[-Wincompatible-pointer-types]
 1265 |         vi = vec_circular_dot_prodi16(s->rrc_filter,
rx_pulseshaper_re[step], V17_RX_FILTER_STEPS, s->rrc_filter_step);
      |                                       ~^~~~~~~~~~~~
      |                                        |
      |                                        float *
In file included from v17rx.c:54:
spandsp/vector_int.h:111:62: note: expected 'const int16_t *' {aka
'const short int *'} but argument is of type 'float *'
  111 | SPAN_DECLARE(int32_t) vec_circular_dot_prodi16(const int16_t
x[], const int16_t y[], int n, int pos);
      |                                                ~~~~~~~~~~~~~~^~~
v17rx.c:1265:71: error: passing argument 2 of
'vec_circular_dot_prodi16' from incompatible pointer type
[-Wincompatible-pointer-types]
 1265 |         vi = vec_circular_dot_prodi16(s->rrc_filter,
rx_pulseshaper_re[step], V17_RX_FILTER_STEPS, s->rrc_filter_step);
      |
~~~~~~~~~~~~~~~~~^~~~~~
      |                                                                       |
      |
       const float *
spandsp/vector_int.h:111:81: note: expected 'const int16_t *' {aka
'const short int *'} but argument is of type 'const float *'
  111 | SPAN_DECLARE(int32_t) vec_circular_dot_prodi16(const int16_t
x[], const int16_t y[], int n, int pos);
      |
   ~~~~~~~~~~~~~~^~~
v17rx.c:1306:44: error: passing argument 1 of
'vec_circular_dot_prodi16' from incompatible pointer type
[-Wincompatible-pointer-types]
 1306 |             vi = vec_circular_dot_prodi16(s->rrc_filter,
rx_pulseshaper_im[step], V17_RX_FILTER_STEPS, s->rrc_filter_step);
      |                                           ~^~~~~~~~~~~~
      |                                            |
      |                                            float *
spandsp/vector_int.h:111:62: note: expected 'const int16_t *' {aka
'const short int *'} but argument is of type 'float *'
  111 | SPAN_DECLARE(int32_t) vec_circular_dot_prodi16(const int16_t
x[], const int16_t y[], int n, int pos);
      |                                                ~~~~~~~~~~~~~~^~~
v17rx.c:1306:75: error: passing argument 2 of
'vec_circular_dot_prodi16' from incompatible pointer type
[-Wincompatible-pointer-types]
 1306 |             vi = vec_circular_dot_prodi16(s->rrc_filter,
rx_pulseshaper_im[step], V17_RX_FILTER_STEPS, s->rrc_filter_step);
      |
~~~~~~~~~~~~~~~~~^~~~~~
      |
           |
      |
           const float *
spandsp/vector_int.h:111:81: note: expected 'const int16_t *' {aka
'const short int *'} but argument is of type 'const float *'
  111 | SPAN_DECLARE(int32_t) vec_circular_dot_prodi16(const int16_t
x[], const int16_t y[], int n, int pos);
      |
   ~~~~~~~~~~~~~~^~~
v17rx.c: In function 'v17_rx_restart':
v17rx.c:1419:18: error: passing argument 1 of 'vec_zeroi16' from
incompatible pointer type [-Wincompatible-pointer-types]
 1419 |     vec_zeroi16(s->rrc_filter,
sizeof(s->rrc_filter)/sizeof(s->rrc_filter[0]));
      |                 ~^~~~~~~~~~~~
      |                  |
      |                  float *
spandsp/vector_int.h:58:44: note: expected 'int16_t *' {aka 'short int
*'} but argument is of type 'float *'
   58 | static __inline__ void vec_zeroi16(int16_t z[], int n)
      |                                    ~~~~~~~~^~~
make[4]: *** [Makefile:978: v17rx.lo] Error 1

Best regards,
Dylan

[1] https://buildd.debian.org/status/fetch.php?pkg=spandsp&arch=armel&ver=0.0.6%2Bdfsg-2.1%2Bb1&stamp=1730324213&raw=0
[2] https://buildd.debian.org/status/fetch.php?pkg=spandsp&arch=armhf&ver=0.0.6%2Bdfsg-2.1%2Bb1&stamp=1730287421&raw=0