Re: Add support for %0d format specifier to base m4

On Sun, Dec 05, 2021 at 09:51:48AM +0000, Stuart Henderson wrote:
> It may work to symlink gm4 to ${WRKDIR}/bin/m4 in post-extract. But this has
> come up in other places too so it would be helpful to have support in base
> m4.
>
> --
> Sent from a phone, apologies for poor formatting.
>
> On 4 December 2021 12:47:30 Christopher Zimmermann <chrisz@openbsd.org> wrote:
>
> > Hi,
> >
> > current ocaml uses %02d format specifier in its autoconf. This is not
> > supported by our base m4. I had no idea how to force autoconf to use
> > devel/m4. So I added support for this flag to base m4. OK?
> > Alternatively suggestions are welcome how to force use of
> > /usr/local/bin/gm4 instead of /usr/bin/m4.
> >
> > Christopher
> >
> >
> > Index: gnum4.c
> > ===================================================================
> > RCS file: /cvs/src/usr.bin/m4/gnum4.c,v
> > retrieving revision 1.52
> > diff -u -p -r1.52 gnum4.c
> > --- gnum4.c 21 Aug 2017 21:41:13 -0000 1.52
> > +++ gnum4.c 4 Dec 2021 12:41:43 -0000
> > @@ -542,12 +542,12 @@ doformat(const char *argv[], int argc)
> > {
> > const char *format = argv[2];
> > int pos = 3;
> > - int left_padded;
> > - long width;
> > + int left_padded, fill_0 = 0;
> > + int width;
> > size_t l;
> > const char *thisarg;
> > - char temp[2];
> > - long extra;
> > + char temp[32];
> > + long extra, number;
> >
> > while (*format != 0) {
> > if (*format != '%') {
> > @@ -564,15 +564,19 @@ doformat(const char *argv[], int argc)
> > addchar('%');
> > break;
> > }
> > + if (*format == '0') {
> > + format++;
> > + fill_0 = 1;
> > + }
> >
> > if (*format == '*') {
> > format++;
> > if (pos >= argc)
> > m4errx(1,
> > "Format with too many format specifiers.");
> > - width = strtol(argv[pos++], NULL, 10);
> > + width = (int)strtol(argv[pos++], NULL, 10);
> > } else {
> > - width = strtol(format, (char **)&format, 10);
> > + width = (int)strtol(format, (char **)&format, 10);
> > }
> > if (width < 0) {
> > left_padded = 1;
> > @@ -599,6 +603,18 @@ doformat(const char *argv[], int argc)
> > switch(*format) {
> > case 's':
> > thisarg = argv[pos++];
> > + break;
> > + case 'd':
> > + number = strtol(argv[pos], (char **)&thisarg, 0);
> > + if (*argv[pos] == '\0' || *thisarg != '\0')
> > + m4errx(1,
> > + "Unexpected argument to %%d format specifier: %s.",
> > + argv[pos]);
> > + pos ++;
> > + snprintf(temp, sizeof(temp),
> > + fill_0 ? "%0*ld" : "%*ld",
> > + left_padded ? -width : width, number);
> > + thisarg = temp;
> > break;
> > case 'c':
> > temp[0] = strtoul(argv[pos++], NULL, 10);
> > Index: m4.1
> > ===================================================================
> > RCS file: /cvs/src/usr.bin/m4/m4.1,v
> > retrieving revision 1.65
> > diff -u -p -r1.65 m4.1
> > --- m4.1 8 Mar 2021 02:47:28 -0000 1.65
> > +++ m4.1 4 Dec 2021 12:41:43 -0000
> > @@ -290,8 +290,8 @@ and following arguments, in a way simila
> > .Xr printf 3 .
> > This built-in is only available in GNU-m4 compatibility mode, and the only
> > parameters implemented are there for autoconf compatibility:
> > -left-padding flag, an optional field width, a maximum field width,
> > -*-specified field widths, and the %s and %c data type.
> > +left-padding flag, pad with leading zeros flag, an optional field width, a
> > +maximum field width, *-specified field widths, and the %d, %s and %c
> > data type.
> > .It Fn ifdef name yes no
> > If the macro named by the first argument is defined then return the second
> > argument, otherwise the third.
>
>
I'll have a look at supporting this properly.

I wasn't even aware this came up in other ports!

Talk to me guys ;)