Libguestfs - Jan 2017 - Re: [PATCH v6 2/3] mllib: modify nsplit to take optional noempty and count arguments

On Mon, Jan 30, 2017 at 10:43:15PM +0100, Tomáš Golembiovský
wrote:
> Added two new optional arguments to nsplit:
> 
> * keep_empty: if set to false empty elements are not stored in the
>   returned list. The default is to keep the empty elements
The ?keep_empty flag is pointless.  It's simpler and more
clear to write:

  List.filter ((<>) "") (nsplit ...)

when you don't want empty elements.

However the ?count flag is fine, since that's what Perl's split
function also does.
> -    let rec nsplit sep str > +    let rec nsplit ?(keep_empty = true)
?(count = -1) sep str >        let len = length str in
>        let seplen = length sep in
>        let i = find str sep in
> -      if i = -1 then [str]
> +      if i = -1 || count = 0 then
> +        if str = "" && not keep_empty then [] else [str]
>        else (
>          let s' = sub str 0 i in
>          let s'' = sub str (i+seplen) (len-i-seplen) in
> -        s' :: nsplit sep s''
> +        let elem, count > +          if s' = ""
&& not keep_empty then
> +            [], count
> +          else
> +            [s'], if count > 0 then count-1 else count
> +        in
> +        elem @ nsplit ~keep_empty:keep_empty ~count:count sep s''
You can write this line as:

    elem @ nsplit ~count sep s''

(after dropping keep_empty).

Rich.

-- 
Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones
Read my programming and virtualization blog: http://rwmj.wordpress.com
libguestfs lets you edit virtual machines.  Supports shell scripting,
bindings from many languages.  http://libguestfs.org

On Tue, Jan 31, 2017 at 11:35:01AM +0000, Richard W.M. Jones
wrote:
> You can write this line as:
> 
>     elem @ nsplit ~count sep s''
For why, see:

https://www.redhat.com/archives/libguestfs/2017-January/msg00020.html

and follow-ups.

Rich.

-- 
Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones
Read my programming and virtualization blog: http://rwmj.wordpress.com
virt-p2v converts physical machines to virtual machines.  Boot with a
live CD or over the network (PXE) and turn machines into KVM guests.
http://libguestfs.org/virt-v2v

On Tue, 31 Jan 2017 11:35:01 +0000
"Richard W.M. Jones" <rjones@redhat.com> wrote:
> On Mon, Jan 30, 2017 at 10:43:15PM +0100, Tomáš Golembiovský wrote:
> > Added two new optional arguments to nsplit:
> > 
> > * keep_empty: if set to false empty elements are not stored in the
> >   returned list. The default is to keep the empty elements  
> 
> The ?keep_empty flag is pointless.  It's simpler and more
> clear to write:
> 
>   List.filter ((<>) "") (nsplit ...)
> 
> when you don't want empty elements.
That's not entirely true when keep_empty is used in combination with
count. When keep_empty is false then the split is not considered a
(successful) split and internal state of count is not decreased.

Example:

# List.filter ((<>) "") (nsplit  ~count:3 ","
"a,,,b,,,c,,d" ) ;;
  - : string list = ["a"; "b,,,c,,d"]

# nsplit ~keep_empty:false ~count:3 "," "a,,,b,,,c,,d" ;;
  - : string list = ["a"; "b"; "c";
",d"]
> 
> However the ?count flag is fine, since that's what Perl's split
> function also does.
> 
> > -    let rec nsplit sep str > > +    let rec nsplit ?(keep_empty
= true) ?(count = -1) sep str > >        let len = length str in
> >        let seplen = length sep in
> >        let i = find str sep in
> > -      if i = -1 then [str]
> > +      if i = -1 || count = 0 then
> > +        if str = "" && not keep_empty then [] else
[str]
> >        else (
> >          let s' = sub str 0 i in
> >          let s'' = sub str (i+seplen) (len-i-seplen) in
> > -        s' :: nsplit sep s''
> > +        let elem, count > > +          if s' = ""
&& not keep_empty then
> > +            [], count
> > +          else
> > +            [s'], if count > 0 then count-1 else count
> > +        in
> > +        elem @ nsplit ~keep_empty:keep_empty ~count:count sep
s''
> 
> You can write this line as:
> 
>     elem @ nsplit ~count sep s''
> 
> (after dropping keep_empty).
> 
> Rich.
> 
> -- 
> Richard Jones, Virtualization Group, Red Hat
http://people.redhat.com/~rjones
> Read my programming and virtualization blog: http://rwmj.wordpress.com
> libguestfs lets you edit virtual machines.  Supports shell scripting,
> bindings from many languages.  http://libguestfs.org

-- 
Tomáš Golembiovský <tgolembi@redhat.com>

On Tue, Jan 31, 2017 at 01:17:12PM +0100, Tomáš Golembiovský
wrote:
> On Tue, 31 Jan 2017 11:35:01 +0000
> "Richard W.M. Jones" <rjones@redhat.com> wrote:
> 
> > On Mon, Jan 30, 2017 at 10:43:15PM +0100, Tomáš Golembiovský wrote:
> > > Added two new optional arguments to nsplit:
> > > 
> > > * keep_empty: if set to false empty elements are not stored in
the
> > >   returned list. The default is to keep the empty elements  
> > 
> > The ?keep_empty flag is pointless.  It's simpler and more
> > clear to write:
> > 
> >   List.filter ((<>) "") (nsplit ...)
> > 
> > when you don't want empty elements.
> 
> That's not entirely true when keep_empty is used in combination with
> count. When keep_empty is false then the split is not considered a
> (successful) split and internal state of count is not decreased.
> 
> Example:
> 
> # List.filter ((<>) "") (nsplit  ~count:3 ","
"a,,,b,,,c,,d" ) ;;
>   - : string list = ["a"; "b,,,c,,d"]
> 
> # nsplit ~keep_empty:false ~count:3 "," "a,,,b,,,c,,d"
;;
>   - : string list = ["a"; "b"; "c";
",d"]
IMHO that shows how ?keep_empty is a very complex interface that's
difficult for ordinary programmers to understand.

Rich.

-- 
Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones
Read my programming and virtualization blog: http://rwmj.wordpress.com
Fedora Windows cross-compiler. Compile Windows programs, test, and
build Windows installers. Over 100 libraries supported.
http://fedoraproject.org/wiki/MinGW