[asterisk-users] libpri compile ubuntu 22.04

[asterisk at phreaknet.org]

On 9/26/2022 8:25 PM, Jerry Geis wrote:
> On Mon, Sep 26, 2022 at 8:09 PM <asterisk at phreaknet.org 
> <mailto:asterisk at phreaknet.org>> wrote:
>
>     On 9/26/2022 8:00 PM, Jerry Geis wrote:
>     > I am getting a compile error:
>     >
>     > gcc -g  -Wall -Werror -Wstrict-prototypes -Wmissing-prototypes
>     > -Werror=zero-length-bounds -fPIC  -O2  -MD -MT q921.o -MF
>     .q921.o.d -MP -c
>     > -o q921.o q921.c
>     > q921.c: In function ‘q921_dump’:
>     > q921.c:1333:85: error: array subscript 0 is outside the bounds of an
>     > interior zero-length array ‘u_int8_t[0]’ {aka ‘unsigned char[]’}
>     > [-Werror=zero-length-bounds]
>     >   1333 |                 if ((h->u.ft == 3) && (h->u.m3 == 0) &&
>     (h->u.m2 ==
>     > 0) && (h->u.data[0] == 0x0f)) {
>     >        |
>     >         ~~~~~~~~~^~~
>     >
>     > EVEN though I added the "-Werror=zero-length-bounds" to the
>     makefile as you
>     > can see.
>     > How do I get libpri-1.6.0 to compile on Ubuntu 22.04 LTS ?
>     Just informationally, there is an open issue for this[1], but it
>     hasn't
>     been worked on.
>     I am not sure if changing the [0] to [] will do the trick as I
>     haven't
>     been able to look at this yet, but you could try that.
>
>     [1] https://issues.asterisk.org/jira/browse/PRI-189
>
>
> Problem is it references [0], [1], [2], [3], and [4]. Not just [0], so 
> I don't think so.
> Thank you.

[0] is a notation used for flexible array members, placed at the end of 
a struct. It does not actually mean the array is of size 0. It seems 
that [0] may have fallen out of use now in favor of [] which is how it 
is defined in the C99 standard.
I have not tested this in any way whatsoever, but here is a quick 
patch[1] that makes it compile again, that you could try. If you do try 
it, let us know how it works.

[1] https://gerrit.asterisk.org/c/libpri/+/19311