Re: Initdb failure

On Thu, Jul 25, 2019 at 8:39 PM Rafia Sabih <rafia(dot)pghackers(at)gmail(dot)com> wrote:
>
> On Thu, 25 Jul 2019 at 16:44, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:
> >
> > Rafia Sabih <rafia(dot)pghackers(at)gmail(dot)com> writes:
> > > On Thu, 25 Jul 2019 at 13:50, vignesh C <vignesh21(at)gmail(dot)com> wrote:
> > >>> Initdb fails when following path is provided as input:
> > >>> datasadfasfdsafddsdatasadfasfdsafddsdatasadfasfdsafddsdatasadfasfdsafddsdatasadfasfdsafddsdatasadfasfdsafddsdatasadfasfdsafddsdatasadfasfdsafddsdatasadfasfdsafddsdatasadfasfdsafddsdatasadfasfdsafddsdatasadfasfdsafdds/datasadfasfdsafddsdatasadfasfdsafddsdatasadfasfdsafddsdatasadfasfdsafddsdatasadfasfdsafddsdatasadfasfdsafddsdatasadfasfdsafddsdatasadfasfdsafddsdatasadfasfdsafddsdatasadfasfdsafddsdatasadfasfdsafddsdatasadfasfdsafdds/datasadfasfdsafddsdatasadfasfdsafddsdatasadfasfdsafddsdatasadfasfdsafddsdatasadfasfdsafddsdatasadfasfdsafddsdatasadfasfdsafddsdatasadfasfdsafddsdatasadfasfdsafddsdatasadfasfdsafddsdatasadfasfdsafddsdatasadfasfdsafdds/datasadfasfdsafddsdatasadfasfdsafddsdatasadfasfdsafddsdatasadfasfdsafddsdatasadfasfdsafddsdatasadfasfdsafddsdatasadfasfdsafddsdatasadfasfdsafddsdatasadfasfdsafddsdatasadfasfdsafddsdatasadfasfdsafddsdatasadfasfdsafdds/datasadfasfdsafddsdatasadfasfdsafddsdatasadfasfdsafddsdatasadfasfdsafddsdatasadfasfdsafddsdatasadfasfdsafddsdatasadfasfdsafddsdatasadfasfdsafddsdatasadfasfdsafddsdatasadfasfdsafddsdatasadfasfdsafddsdatasadfasfdsafdds/datasadfasfdsafddsdatasadfasfdsafddsdatasadfasfdsafddsdatasadfasfdsafddsdatasadfasfdsafddsdatasadfasfdsafddsdatasadfasfdsafddsdatasadfasfdsafddsdatasadfasfdsafddsdatasadfasfdsafddsdatasadfasfdsafddsdatasadfasfdsafdds/
> >
> > > Now that you say this, it does make sense to atleast inform about the
> > > correct error and that too earlier. Something like the attached patch
> > > would make sense.
> >
> > I am not terribly excited about putting effort into this at all, because
> > I don't think that any actual user anywhere will ever get any benefit.
> > The proposed test case is just silly.
>
> That I totally agree upon!
> But on the other hand emitting the right error message atleast would
> be good for the sake of correctness if nothing else. But yes that
> definitely should be weighed against what is the effort required for
> this.
>
Thanks Tom for your opinion.
Thanks Rafia for your thoughts and effort in making the patch.

I'm not sure if we are planning to fix this.
If we are planning to fix, one suggestion from my side we can
choose a safe length which would include the subdirectories
and file paths. I think one of these will be the longest:
base/database_oid/tables
pg_wal/archive_status/
pg_wal/archive_file

Fix can be something like:
MAXPGPATH - (LONGEST_PATH_FROM_ABOVE)

Regards,
Vignesh
EnterpriseDB: http://www.enterprisedb.com