Re: Proposal: chop() dropped
I already proposed that. Benchmarks show that reading a file with explicit chomp() is easily 20% slower than reading the same file with implicit chomp(), through the -l command line switch. And what, pray tell, do you do about the small matter of wanting to read some files without implicit record-terminator deletion, and others with such? Or, for that matter, applying it to something other than the implicit ARGV reading from -n or -p? Seems like it should be a filehandle property. --tom
Re: Proposal: chop() dropped
Eric Roode writes: Useful functions all, no doubt. But I would lobby heavily for a new set of names -- ones that can be remembered! Quick -- which trims leading spaces, champ, chump, or chimp? My favourite: chafe(). Nat
Re: Proposal: chop() dropped
On Thu, 31 Aug 2000 19:59:31 +0200, you wrotc: tr/\w//dlt # Trim all leading trailing whitespace from $_ Eh, scratch that. Too much caffeine i guess. tr/\n\r\t //dlt; # Trim some whitespace. -DZ -- Tell me your dreams and I will crush them.
Re: Proposal: chop() dropped
On Thu, Aug 31, 2000 at 07:59:31PM +0200, Dan Zetterstrom wrote: Why not use the "function" we already got, tr? Something like: tr///l # Translate only _l_eading characters matching. tr///t # Translate only _t_railing characters matching. With "Only leading" I mean translate from start/end until you find a character not matching. Then you can do nifty things such as: Um, that would radically change the meaning of tr///. Better to use s/^// and s/$//. -Scott -- Jonathan Scott Duff [EMAIL PROTECTED]
Re: Proposal: chop() dropped
tr///l # Translate only _l_eading characters matching. tr///t # Translate only _t_railing characters matching. With "Only leading" I mean translate from start/end until you find a character not matching. Then you can do nifty things such as: tr/\w//dlt # Trim all leading trailing whitespace from $_ tr/// does not admit character classes! The "leading" thing can be effected with \G. For example: s/\G /0/g; # change leading blanks to 0 In general, getting folks to write s/^\s+//s; s/\s+$//s; # XXX: \z is a *good* think. --tom
Re: Proposal: chop() dropped
How would you do:
# Writer insists on blank line between paragraphs, first line indented.
# Publisher insists on one paragraph/line, first word ALL CAPS.
Step 1: Fire the lame publisher. I'm serious. It's amazing
what people tolerate. Some things aren't worth the pane.
{
local $/ = ""; #slurp paragraph at a time.
while (INFILE) {
s/\n//gm;# combine into one line
No, that should be s/\n/ /g, as otherwise you merge two
words and get for example "twowords" on this paragraph.
And that /m does nothing, as there's no ^ nor $ to affect.
s/^\s//; # get rid of indent
Um, surely you mean s/^\s+// there instead.
y/a-z/A-Z/l # upcase first (English) word.
Does that mean capitalize it, or reader all the word's letters
in uppercase? I guess you said ALL CAPS, so, trivially enough,
it's merely
s/(\w+)/\U$1/;
Whatcha doin' with tr anyway? :-) That's what the vi escapes are for!
print OUTFILE;
}
}
--tom
Re: Proposal: chop() dropped
TomC wrote:
In general, getting folks to write
s/^\s+//s;
s/\s+$//s; # XXX: \z
is a *good* think.
Why?
Removing leading/trailing whitespace is a tremendously frequently-
performed task. Perl gives you -l on the command line to strip
newlines on input and add them on output, simply because that's
what you need to do, very often.
Perl allows you to glomp your input one "paragraph" at a time,
by setting $/ to "", because it's quicker and more convenient than
searching for the pattern /\n{2,}/, and because people often want
to do that.
I'm not arguing in favor of the tr/// hack specifically, but
gosh, wouldn't it be nice if there were a thwack() builtin that
stripped leading and trailing spaces?
while ()
{
chomp;
thwack; # or whatever
barf if /^barf/;
last if /!!$/;
print LOG "$_ found\n";
}
--
Eric J. Roode, [EMAIL PROTECTED] print scalar reverse sort
Senior Software Engineer'tona ', 'reh', 'ekca', 'lre',
Myxa Corporation'.r', 'h ', 'uj', 'p ', 'ts';
Re: Proposal: chop() dropped
On Thu, 31 Aug 2000 13:36:10 -0600, Tom Christiansen wrote: I'm not arguing in favor of the tr/// hack specifically, but gosh, wouldn't it be nice if there were a thwack() builtin that stripped leading and trailing spaces? No. People should learn intrinsic mechanisms with which they can construct infinitely many beautiful and powerful effects. This empowers them. Making them learn yet-another-function-call merely hamstrings them with a dead fish; tomorrow, they shall starve. Then let's drop chop(). This function is even far more easily implemented than trim(), thwack(), or whatever you choose to call it. Eh... full circle? -- Bart.
Re: Proposal: chop() dropped
On Wed, Aug 30, 2000 at 02:31:00PM -0600, Nathan Torkington wrote: chomp() is best used for chop()s main raison d'etre, removing $/ from a string. I say we drop chop(). I'll second that motion. We already have lots of ways of removing the last character of a string if that's what we really need. -Scott -- Jonathan Scott Duff [EMAIL PROTECTED]
Re: Proposal: chop() dropped
chomp() is best used for chop()s main raison d'etre, removing $/
from a string. I say we drop chop().
So code that says
chop($k,$v)
will need to say
for ($k,$v) { s/.\z//s }
or else something like:
for ($k, $v) { substr($_, length() - 1) = '' }
I'm not sure I find either of those more legible. And they certainly
won't be faster. chop() has been around since perl1, too.
Then again, yes, people do tend to use it when they oughtn't. Hm...
---tom
Re: Proposal: chop() dropped
the reason that they're duplicatable with other features, while I want to drop chop because its main purpose has now been replaced with the far superior chomp. Except that chomp() relies upon the ueberglobal $/ variable, irrespective of the source of the data being chomped. --tom
Re: Proposal: chop() dropped
Tom Christiansen writes:
So code that says
chop($k,$v)
will need to say
for ($k,$v) { s/.\z//s }
or else something like:
for ($k, $v) { substr($_, length() - 1) = '' }
I don't think chop() is an operation that's done often enough for
either of the things above to be a problem.
I'm not sure I find either of those more legible. And they certainly
won't be faster. chop() has been around since perl1, too.
Yes, but chop()'s original purpose was what chomp() is now used for.
I doubt Larry would really have put in a function to remove the last
char in a string just to have a function that removes the last char in
a string. It was the chomp()-like itch I think he was trying to
scratch.
Then again, yes, people do tend to use it when they oughtn't. Hm...
Exactly.
Nat
Re: Proposal: chop() dropped
At 04:36 PM 8/30/00 -0600, Tom Christiansen wrote: the reason that they're duplicatable with other features, while I want to drop chop because its main purpose has now been replaced with the far superior chomp. Except that chomp() relies upon the ueberglobal $/ variable, irrespective of the source of the data being chomped. I presume that line disciplines will be object-oriented and inherit from some base class; therefore a bare chomp will use the line terminator from that base class and for the others, you could do something like $fh-chomp($line) to do chomping specific to a particular filehandle. Make sense? -- Peter Scott Pacific Systems Design Technologies
Re: Proposal: chop() dropped
On Wed, 30 Aug 2000 16:14:35 -0600, Tom Christiansen wrote:
I say we drop chop().
So code that says
chop($k,$v)
will need to say
for ($k,$v) { s/.\z//s }
or else something like:
for ($k, $v) { substr($_, length() - 1) = '' }
I'm not sure I find either of those more legible.
I'm sure. It's not. But either of this works.
substr($_, -1) = '';
$chopped = substr $_, -1, 1, '';
chop() is a speed optimization for a (currently) pretty uncommon task:
removing the last character of a string. Why not the first character,
for example? The string equivalent of shift()?
Only people who insist that ALL lines in a text file should end in a
newline, would still prefer use it.
--
Bart.
Re: Proposal: chop() dropped
I would actualy like to see chop expanded to allow a variable number of
characters to be removed and a sister function to cut the head off. Yes I
know you can do this with substr but sometimes when you want the performance
and need to cut up a string into fields. Lets say the new function is called
'take'.
while($str =F)
{
$head = take($str,40);
$tail = chop($str,40);
$middle = $str;
#assuming $str had 120 characters in it
}
Just a thought
The current syntax is more like chop(@), not chop($;$).
--tom
Re: Proposal: chop() dropped
On Wed, Aug 30, 2000 at 02:31:00PM -0600, Nathan Torkington wrote: chomp() is best used for chop()s main raison d'etre, removing $/ from a string. I say we drop chop(). Works for me. Are you going to RFC it? K. -- Kirrily Robert -- [EMAIL PROTECTED] -- http://netizen.com.au/ Open Source development, consulting and solutions Level 10, 500 Collins St, Melbourne VIC 3000 Phone: +61 3 9614 0949 Fax: +61 3 9614 0948 Mobile: +61 410 664 994
Re: Proposal: chop() dropped
Tom Christiansen wrote: I'll second that motion. We already have lots of ways of removing the last character of a string if that's what we really need. But they're slow and hard to read. I think the word "drop" should be clarified as "dropped from the core binary". In a very cool email, Bryan Warnock talked about half a dozen different chop-like functions that are useful but not in core: http://www.mail-archive.com/perl6-language@perl.org/msg01522.html I think chop, champ, chip, and friends should be available via the use String::Chop; or related module. -Nate
