This is the mail archive of the libc-alpha@sourceware.org mailing list for the glibc project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

SV: [PATCH] Remove __wur attribute for printf/scanf functions



Hello!

> you're confusing asprintf as a benign printf

ok I guess so.


> as for the scanf family, see BZ #14254 and the thread from last time where we rejected this proposal:

thanks. however I am still not convinced. If I trust the input or think that 'garbage in => garbage out' is fine then ignoring the return value from scanf is ok.

I suggest a conservative approach that avoids FP; only use __wur if it's pointless or leads to UB to ignore the return value. calling strlen without using the return value would be pointless. calling fopen without taking the return value would probably lead to a resource leak.

If you still think that __wur should be used.. then let's keep it. I can live with the FP.

Best regards,
Daniel Marjamäki

..................................................................................................................
Daniel Marjamäki Senior Engineer
Evidente ES East AB  Warfvinges väg 34  SE-112 51 Stockholm  Sweden

Mobile:                 +46 (0)709 12 42 62
E-mail:                 Daniel.Marjamaki@evidente.se

www.evidente.se

________________________________________
Från: Mike Frysinger [vapier@gentoo.org]
Skickat: den 12 augusti 2015 11:59
Till: Daniel Marjamäki
Kopia: libc-alpha@sourceware.org
Ämne: Re: [PATCH] Remove __wur attribute for printf/scanf functions

On 12 Aug 2015 09:38, Daniel Marjamäki wrote:
> Imho the __wur attribute should not be used on printf/scanf functions.
>
> It is not insane to ignore the return value from these. these functions will perform the wanted I/O even though the return value is not used.

you're confusing asprintf as a benign printf.  it is insane to not be checking
the return value of that function.  it does memory allocation among other
things.

as for the scanf family, see BZ #14254 and the thread from last time where we
rejected this proposal:
https://sourceware.org/bugzilla/show_bug.cgi?id=14254
https://sourceware.org/ml/libc-alpha/2013-10/msg00290.html
-mike

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]