Re: [Csnd-dev] [EXTERNAL] [Csnd-dev] Spaces in csoundSetOption

The problem was with calls of the form

res = csoundSetOption(csound, "-odac:Built-in Audio");

or similar, which are common in linux systems running pipewire. I see
similar problems with midi devices with spaces.

On Thu, Jan 9, 2025 at 12:29 PM vlz <[email protected]> wrote:

> Multiple whitespaces are now properly accounted for in #2078.
>
> > On 9 Jan 2025, at 10:12, vlz <[email protected]> wrote:
> >
> > I checked and this code works
> >
> > res = csoundSetOption(csound, "-o dac -m 6");
> >
> > so you can have spaces between the arguments. There is a problem with
> more than a single space, which I
> > can look into.
> >
> >> On 9 Jan 2025, at 09:55, vlz <[email protected]> wrote:
> >>
> >> *Warning*
> >> This email originated from outside of Maynooth University's Mail
> System. Do not reply, click links or open attachments unless you recognise
> the sender and know the content is safe.
> >> I wrote this new code, because the 6.x version was problematic. I can
> look into improving it.
> >>
> >> Prof. Victor Lazzarini
> >> Maynooth University
> >> Ireland
> >>
> >>> On 9 Jan 2025, at 09:17, Eduardo Moguillansky <
> [email protected]> wrote:
> >>>
> >>>  Hi,
> >>>
> >>> Directly using the API in c (csound 7) I don't seem to be able to set
> an option which includes a space. In particular, setting the -o or --output
> option with a device which includes a space does not seem to work in any
> case. Looking at the source code it seems that a very naive string
> splitting code was added to csound 7.
> >>>
> >>> csound 6
> >>>
> >>> PUBLIC int csoundSetOption(CSOUND *csound, const char *option){
> >>>    /* if already compiled and running, return */
> >>>    if (csound->engineStatus & CS_STATE_COMP) return 1;
> >>>    else {
> >>>    const char *args[2] = {"csound", option};
> >>>    csound->info_message_request = 1;
> >>>    return (argdecode(csound, 1, args) ? 0 : 1);
> >>>    }
> >>> }
> >>>
> >>> csound 7
> >>>
> >>> /* count whitespaces */
> >>>    while (*sp++ != '\0') {
> >>>      if (*sp == ' ') {
> >>>        cnt++;
> >>>        *sp = '\0';
> >>>        sp++;
> >>>      }
> >>>    }
> >>>    args = (char **)mcalloc(csound, sizeof(char *) * (cnt + 2));
> >>>    args[0] = "csound";
> >>>    args[1] = sp = options;
> >>>    cnt = 1;
> >>>    /* split into separate args */
> >>>    while (*opt) {
> >>>      if (*opt == ' ') {
> >>>        args[++cnt] = sp + 1;
> >>>      }
> >>>      sp++;
> >>>      opt++;
> >>>    }
> >>>
> >>>    ret = argdecode(csound, cnt, (const char **)args);
> >>>    mfree(csound, args);
> >>>    mfree(csound, options);
> >>>
> >>> I don't know the rationale behind this inclusion but I think that at
> least some form of escaping or quotation should be added to the splitting.
> For me the actual solution would be to remove the splitting at the API
> level.
> >>>
> >>> cheers
> >>> Eduardo
>