Re^2: Extracting /regex/mg in a while loop

I expect the /dev/null on line 14 will need changing

The "bit bucket" on Windows is the psuedo file called NUL. This is a reserved file name and you cannot create a file named that. On command line: type someFile > NUL reads someFile and sends it to nowhere. You can open a filehandle to NUL and write to it.

Comment on Re^2: Extracting /regex/mg in a while loop Download Code

Replies are listed 'Best First'.
Re^3: Extracting /regex/mg in a while loop by kcott (Archbishop) on Oct 10, 2023 at 23:04 UTC
G'day Marshall, ++haj already pointed me to a portable solution, using devnull() from File::Spec, which I incorporated into my code. `... use File::Spec (); ... open $null_fh, '>', File::Spec::->devnull(); ...` [download] While writing my original code, I had used `state` variables so that the "`# Increase multiplier for benchmarking`" (from NetWallah's OP) would only be called once for each of the multitude of calls by `Benchmark::cmpthese()`. Unfortunately, subsequently adding code (prior to posting) to test each of the subroutines to be benchmarked ("`say 'Test', ...`") caused major problems. I didn't spot that until after I posted: I pretty much rewrote the code to fix these problems and added Updates to explain what I'd done. In order to avoid invalidating what haj had written (as per "How do I change/delete my post?") I left "CAVEAT for MSWin users: ..." but immediately followed it with "Update: ... CAVEAT is no longer valid ..." I appreciate this ended up all a bit messy. My apologies if it caused confusion. — Ken	[reply] [d/l] [select]

Replies are listed 'Best First'.

Re^3: Extracting /regex/mg in a while loop
by kcott (Archbishop) on Oct 10, 2023 at 23:04 UTC

G'day Marshall,

++haj already pointed me to a portable solution, using devnull() from File::Spec, which I incorporated into my code.

...
use File::Spec ();
...
    open $null_fh, '>', File::Spec::->devnull();
...
[download]

While writing my original code, I had used state variables so that the "# Increase multiplier for benchmarking" (from NetWallah's OP) would only be called once for each of the multitude of calls by Benchmark::cmpthese(). Unfortunately, subsequently adding code (prior to posting) to test each of the subroutines to be benchmarked ("say 'Test', ...") caused major problems. I didn't spot that until after I posted: I pretty much rewrote the code to fix these problems and added Updates to explain what I'd done.

In order to avoid invalidating what haj had written (as per "How do I change/delete my post?") I left "CAVEAT for MSWin users: ..." but immediately followed it with "Update: ... CAVEAT is no longer valid ..."

I appreciate this ended up all a bit messy. My apologies if it caused confusion.

— Ken

[reply]
[d/l]
[select]