[Libguestfs] [PATCH nbdkit v2 3/4] common/utils: Add C string quoting function

[Laszlo Ersek]

On 5/9/23 11:51, Richard W.M. Jones wrote:
> eg. { '\r', '\n' } -> { '\\', 'n', '\\', 'r' }
> ---
>  common/utils/utils.h |  1 +
>  common/utils/quote.c | 58 ++++++++++++++++++++++++++++++++++++++++++++
>  2 files changed, 59 insertions(+)
> 
> diff --git a/common/utils/utils.h b/common/utils/utils.h
> index 42288a5cd..1e6d8972b 100644
> --- a/common/utils/utils.h
> +++ b/common/utils/utils.h
> @@ -35,6 +35,7 @@
>  
>  extern void shell_quote (const char *str, FILE *fp);
>  extern void uri_quote (const char *str, FILE *fp);
> +extern void c_string_quote (const char *str, FILE *fp);
>  extern int exit_status_to_nbd_error (int status, const char *cmd);
>  extern int set_cloexec (int fd);
>  extern int set_nonblock (int fd);
> diff --git a/common/utils/quote.c b/common/utils/quote.c
> index 863e08a8e..877035387 100644
> --- a/common/utils/quote.c
> +++ b/common/utils/quote.c
> @@ -37,6 +37,9 @@
>  #include <string.h>
>  #include <unistd.h>
>  
> +#include "ascii-ctype.h"
> +#include "hexdigit.h"
> +
>  /* Print str to fp, shell quoting if necessary.  This comes from
>   * libguestfs, but was written by me so I'm relicensing it to a BSD
>   * license for nbdkit.
> @@ -98,3 +101,58 @@ uri_quote (const char *str, FILE *fp)
>        fprintf (fp, "%%%02X", str[i] & 0xff);
>    }
>  }
> +
> +/* Print str to fp, quoting in a way similar to C strings, for example
> + * '\n' -> "\n".
> + *
> + * Note that we do not emit quotes around the string, and double
> + * quotes within the string are not escaped.
> + */
> +void
> +c_string_quote (const char *str, FILE *fp)
> +{
> +  size_t i;
> +  char c;
> +
> +  for (i = 0; c = str[i], c != '\0'; ++i) {

For the controlling expression, do we like

  (c = str[i]) != '\0'

more?

Another question regarding brace style:


> +    if (ascii_isprint (c))
> +      fputc (c, fp);
> +    else {
> +      switch (c) {
> +      case '\a':
> +        fputc ('\\', fp);
> +        fputc ('a', fp);
> +        break;
> +      case '\b':
> +        fputc ('\\', fp);
> +        fputc ('b', fp);
> +        break;
> +      case '\f':
> +        fputc ('\\', fp);
> +        fputc ('f', fp);
> +        break;
> +      case '\n':
> +        fputc ('\\', fp);
> +        fputc ('n', fp);
> +        break;
> +      case '\r':
> +        fputc ('\\', fp);
> +        fputc ('r', fp);
> +        break;
> +      case '\t':
> +        fputc ('\\', fp);
> +        fputc ('t', fp);
> +        break;
> +      case '\v':
> +        fputc ('\\', fp);
> +        fputc ('v', fp);
> +        break;
> +      default:
> +        fputc ('\\', fp);
> +        fputc ('x', fp);
> +        fputc (hexchar (c >> 4), fp);
> +        fputc (hexchar (c), fp);
> +      }
> +    }
> +  }

In theory, we don't have to use any brace characters here except within
the "switch" statement -- the "for" and "else" parts don't need braces.
libnbd / nbdkit usually (?) forbids using braces unless we *must*
braces. But this looks like an exception. What's the rule?

(Sorry if I asked the same question before -- I don't remember exactly,
but I might have. It just doesn't congeal into a consistent rule for me.)

Looks OK otherwise of course.

Thanks!
Laszlo

> +}