Re: [PATCH 4/4] cut: Optionally treat multiple consecutive delimiters as one

[Dragan Simic]

On 2023-08-01 18:27, Rob Landley wrote:
> On 8/1/23 04:07, Dragan Simic wrote:
> > Add new command-line option and the required logic that allow multiple
> > consecutive delimiters to be treated as a single delimiter.  Of 
> > course,
> > this option is valid only with the cut's field mode.
> >
> > This new feature should make cut much more usable in various 
> > real-world
> > applications, some of which are already mentioned in the gotchas.  For
> > example, merging the consecutive delimiters is very useful when cut is
> > used to process the outputs of various commands.
>
> A year and a half ago I submitted cut -D -F -O to this list, which 
> included a
> regex mode for delimiters that would allow you to do that in a generic 
> way:
>
>   -F like -f but the -d delimiter is interpreted as an extended regex 
> string
>   -D disables field sort/collate so -f 3,2,1-3 prints fields 3 2 1 2 3
>   -O specifies the output separator (default ' ' for -F, input delim 
> for -f).
>
> The -d default value also changes, the toybox code for which is:
>
>   if (!TT.d) TT.d = FLAG(F) ? "[[:space:]][[:space:]]*" : "\t";
>
> I.E. run of whitespace for -F, same old \t for -f. (There should be no 
> change to
> any existing behavior unless you specify one or more of the new 
> options.)
>
> These options have been in toybox since 2017 (commit 3eb69d412e79, I 
> think it
> missed the Android 8.1 cutoff and wound up in 9? Um, "Oreo" and "Pie"
> respectively.) They were merged into busybox in 2021 (commit 
> 0068ce2fa0e3).
>
> The coreutils developers showed interest in merging support, came up 
> with a
> patch, and discussed it (defining --longopts for everything because 
> gnu, which I
> added to toybox in commit 9b108ab90b63), but the discussion petered out 
> here:
>
>   https://lists.gnu.org/archive/html/coreutils/2022-01/msg00050.html
>
> When I asked again after the next release, they said it would be in 
> 9.2:
>
>   https://lists.gnu.org/archive/html/coreutils/2022-04/msg00010.html
>
> In February of this year I asked again, and it was on the review list 
> for the
> upcoming release:
>
>   https://lists.gnu.org/archive/html/coreutils/2023-02/msg00012.html
>
> The coreutils 9.3 page on the web doesn't include the -DFO options:
>
>   
> https://www.gnu.org/software/coreutils/manual/coreutils.html#cut-invocation
>
> Which means I still need to use "awk '{print $3}'" instead of "cut -DF 
> 3" in
> scripts I want to run on the host.

Thanks for a detailed summary!  I'd have to perform a detailed review of 
the code changes to say "yay" or "nay" to the merging of your patches, 
but the described "-F" and "-D" options look good and I think they would 
be very useful!

By the way, isn't the described "-O" option already covered by the 
"--output-delimiter" option?  However, having "-O" as the short alias 
would be very useful.  Can we have at least that merged?