[PATCH] IFP: Suppress warning from static analyzer

[Lukas Slebodnik]

ehlo,

simple patch is attached.

LS

[Pavel Březina]

On 09/23/2015 01:29 PM, Lukas Slebodnik wrote:
> ehlo,
>
> simple patch is attached.
>
> LS
> if (errno != 0) {
>    ret = errno;
> ret = errno;
> if (ret != 0) {
>     goto done;
> }

Hi,
if we are checking ret, it may be nicer to test against EOK since that 
is what we usually do with this name.

[Jakub Hrozek]

On Wed, Sep 23, 2015 at 09:00:04PM +0200, Pavel Březina wrote:
> On 09/23/2015 01:29 PM, Lukas Slebodnik wrote:
> > ehlo,
> >
> > simple patch is attached.
> >
> > LS
> > if (errno != 0) {
> >    ret = errno;
> > ret = errno;
> > if (ret != 0) {
> >    goto done;
> > }
>
> Hi,
> if we are checking ret, it may be nicer to test against EOK since that is
> what we usually do with this name.

We might even go one step further (in master, when we release 1.13.1)
and just convert the function to return a result and error code
separately:
    ret = strtouint32(val, &ret);

[Lukas Slebodnik]

On (23/09/15 21:18), Jakub Hrozek wrote:
> On Wed, Sep 23, 2015 at 09:00:04PM +0200, Pavel Březina wrote:
> > On 09/23/2015 01:29 PM, Lukas Slebodnik wrote:
> > > ehlo,
> > >
> > > simple patch is attached.
> > >
> > > LS
> > > if (errno != 0) {
> > >    ret = errno;
> > > ret = errno;
> > > if (ret != 0) {
> > >    goto done;
> > > }
> >
> > Hi,
> > if we are checking ret, it may be nicer to test against EOK since that is
> > what we usually do with this name.

Updated version is attached.
> We might even go one step further (in master, when we release 1.13.1)
> and just convert the function to return a result and error code
> separately:
>    ret = strtouint32(val, &ret);

It should not be a problem.

LS

[Pavel Březina]

On 09/24/2015 09:36 AM, Lukas Slebodnik wrote:
> On (23/09/15 21:18), Jakub Hrozek wrote:
> > On Wed, Sep 23, 2015 at 09:00:04PM +0200, Pavel Březina wrote:
> > > On 09/23/2015 01:29 PM, Lukas Slebodnik wrote:
> > > > ehlo,
> > > >
> > > > simple patch is attached.
> > > >
> > > > LS
> > > > if (errno != 0) {
> > > >    ret = errno;
> > > > ret = errno;
> > > > if (ret != 0) {
> > > >     goto done;
> > > > }
> > >
> > > Hi,
> > > if we are checking ret, it may be nicer to test against EOK since that is
> > > what we usually do with this name.
>
> Updated version is attached.

Ack.
> > We might even go one step further (in master, when we release 1.13.1)
> > and just convert the function to return a result and error code
> > separately:
> >     ret = strtouint32(val, &ret);
>
> It should not be a problem.

I agree.
> LS

[Lukas Slebodnik]

On (24/09/15 14:10), Pavel Březina wrote:
> On 09/24/2015 09:36 AM, Lukas Slebodnik wrote:
> > On (23/09/15 21:18), Jakub Hrozek wrote:
> > > On Wed, Sep 23, 2015 at 09:00:04PM +0200, Pavel Březina wrote:
> > > > On 09/23/2015 01:29 PM, Lukas Slebodnik wrote:
> > > > > ehlo,
> > > > >
> > > > > simple patch is attached.
> > > > >
> > > > > LS
> > > > > if (errno != 0) {
> > > > >    ret = errno;
> > > > > ret = errno;
> > > > > if (ret != 0) {
> > > > >    goto done;
> > > > > }
> > > >
> > > > Hi,
> > > > if we are checking ret, it may be nicer to test against EOK since that is
> > > > what we usually do with this name.
> >
> > Updated version is attached.
>
> Ack.

http://sssd-ci.duckdns.org/logs/job/27/90/summary.html

* master: afdc0179af0ad8ddbedd67422193ef02dcd2bf84

LS