Odp: Re: [PATCH] fputs return the number of characters written

[Paul Eggert]

I'm not convinced it'd be useful to modify puts / fputs, due to the 
INT_MAX and portability issues mentioned. That being said....

[putting my standards-nerd hat on; apologies in advance]

On 9/20/21 11:39 AM, Siddhesh Poyarekar wrote:
> AFAICT, the return value is considered implementation defined[1],

[1] doesn't say the return value of 'puts' is implementation-defined. It 
says merely that 'puts' returns a non-negative number on success. This 
is in contrast to the specs for functions like 'malloc'[2], where the 
spec is careful to say that the behavior of 'malloc (0)' is 
implementation-defined on success.

The current glibc manual doesn't say what 'puts' returns, which is 
surely a defect in the manual. The manual should use return-value 
wording similar to that of 'fputs'.

Although the current glibc manual does say that 'fputs' returns a 
non-negative value on success, it doesn't say which value, and this 
vagueness is OK as far as POSIX goes, as POSIX doesn't say the behavior 
is implementation-defined. So from what I can see, changing a successful 
'puts' to return some other non-negative value is merely of pragmatic 
concern rather than being an API or ABI event (as I understand these 
things, anyway).

> [1] 
> https://pubs.opengroup.org/onlinepubs/9699919799/functions/puts.html#tag_16_468

[2] https://pubs.opengroup.org/onlinepubs/9699919799/functions/malloc.html