[9fans] APE vsnprintf return value bug

[9fans] APE vsnprintf return value bug

[Nathaniel W Filardo]

[-- Attachment #1: Type: text/plain, Size: 1097 bytes --]

There seems to be some disagreement about vsnprintf()'s return code, in a
way that's breaking my attempts to get git building on Plan 9.

The manual on my Linux box says:

> The  functions snprintf() and vsnprintf() do not write more than size
> bytes (including the trailing '\0').  If the output was truncated due to
> this limit then the return value is the number of characters (not
> including the trailing '\0') which would have been written to the final
> string if enough space had been available.  Thus, a return value of size
> or more means that the output was truncated.  (See also below under
> NOTES.)

While this strikes me as funny (though I suppose I see the merit), git
depends upon it being this way...  so when
/sys/src/ape/lib/ap/stdio/vfprintf.c says:

> int
> vfprintf(FILE *f, const char *s, va_list args)
> {
...
>   return ferror(f)? -1: nprint;;
> }

It appears that all hope is lost.  If I were to submit a patch which renamed
vfprintf to _vfprintf and pushed the ferror() check down to its callers,
would such a thing be accepted?

--nwf;

[-- Attachment #2: Type: application/pgp-signature, Size: 204 bytes --]

Re: [9fans] APE vsnprintf return value bug

[Steve Simon]

I don't have a copy of the ISO 9899:1990 but Plauger's The Standard C Library
has excerpts from it. In one of these it states:

	``The vfprintf function returns the number of characters transmitted,
	or a negative value if an output error occured''

As APE does take great pains to be standards conformant its a shame to break it here.

How about a _GLIBC_SOURCE define in the headers and a glibc_vfprintf() function
in the library?

Just my 2 cents worth.

-Steve

Re: [9fans] APE vsnprintf return value bug

[Charles Forsyth]

vfprintf should return the number of characters transmitted.
v?snprintf should return the number needed (excluding the nul terminator),
but the actual result varies (incorrectly) between systems.