Re: one liner for replacing under Windows?
On Sat, Jun 14, 2008 at 9:56 AM, Octavian Rasnita [EMAIL PROTECTED] wrote: Is there a one-liner command that can replace a certain text with another in more files specified with wildcards like *.html that works under Windows cmd? Since you ask this you probably know the Windows shell does not expand wildcards. There's a trick for Perl one-liners though (untested): perl -e @ARGV = glob(qq($ARGV[0])); s/foo/bar/ while *.html You see the idea. -- fxn -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] http://learn.perl.org/
Re: one liner for replacing under Windows?
On Sat, Jun 14, 2008 at 1:05 PM, Octavian Rasnita [EMAIL PROTECTED] wrote:
From: Xavier Noria [EMAIL PROTECTED] On Sat, Jun 14, 2008 at 9:56 AM,
Octavian Rasnita [EMAIL PROTECTED] wrote:
Is there a one-liner command that can replace a certain text with another
in
more files specified with wildcards like *.html that works under Windows
cmd?
Since you ask this you probably know the Windows shell does not expand
wildcards. There's a trick for Perl one-liners though (untested):
perl -e @ARGV = glob(qq($ARGV[0])); s/foo/bar/ while
You see the idea.
Thank you but I don't think it works without writing a full perl program in
a single line that does that.
If I just replace the string in $_, the files are not updated.
If I also use -pi.bak parameter in order to do that, it gives an error
telling that the specified file (*.txt) can't be opened.
So I can't see another solution than opening the files, making the
replacement than writing the new content.
Absolutely, try putting a BEGIN block around (again untested):
perl -pi.bak -e BEGIN { @ARGV = glob(shift) } s/foo/bar/g *.html
-- fxn
--
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
http://learn.perl.org/
Re: what is ^M at the end of a line?
On Wed, May 21, 2008 at 4:20 PM, Remy Guo [EMAIL PROTECTED] wrote: it's really interesting... then how can i match that ^M using regex? i've tried chomp when reading each line but it doesn't work... That's \r everywhere except in Macs before Mac OS X. Some programs display \r as ^M but that's just a way to show it, there's really just one character. I you'd like to order the ideas about how newlines work have a look at this article: http://www.onlamp.com/pub/a/onlamp/2006/08/17/understanding-newlines.html -- fxn -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] http://learn.perl.org/
Re: error message
On Apr 20, 2008, at 21:57 , Levente Kovacs wrote: Can anyone tell me what this error message mean? syntax error at ./dbubdate.pl line 181, near }continue The code around that line in that script is not well-formed. If you don't know what to do please send lines 170 to 190 specifying which one is line 181. -- fxn -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] http://learn.perl.org/
Re: error message
On Apr 20, 2008, at 22:45 , Levente Kovacs wrote: Thank you very much for your answer. I am sitting here for a day seeng what could be wrong, but I can't find it. I post a link to my code here. Thank you for your help in advance. http://logonex.eu/cgi-bin/viewvc/viewvc.cgi/scripts4geda/dbubdate.pl?revision=127 http://logonex.eu/cgi-bin/viewvc/viewvc.cgi/scripts4geda/dbubdate.pl?view=log That script does not seem to contain any occurrence of continue, are you sure that's the one? -- fxn -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] http://learn.perl.org/
Re: Best way to grab lines n thru x of a file
On Apr 10, 2008, at 18:24 , Jonathan Mast wrote: Hi, I have a ~125MB file of which I want to read lines n thru n+x and write those into a separate file. What is the best way to go about this? Tie::File -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] http://learn.perl.org/
Re: Variables initialization
On Mar 17, 2008, at 11:51 , [EMAIL PROTECTED] wrote: How is it possible to initialize variable using 0 or '' and not having undef warning later on using them? Not reassigning to undef. -- fxn -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] http://learn.perl.org/
Re: to test maximum consecutive numbers
On Dec 11, 2007, at 9:53 PM, ciwei2103 wrote:
Give a disk sequence know that is incrementing with base 16, such
as ,
0001
0003
0004
0008
0009
000A
000B
000F
0010
0011
how do I extract the a subset that is have at least 4 consecutives,
such as 0008 0009 000A 000B
I hope this has to do with backups and not homework:
use strict;
use warnings;
use constant AT_LEAST_THESE_MANY = 4;
my @disks = qw(
0001
0003
0004
0008
0009
000A
000B
000F
0010
0011
);
my @codes = map hex, @disks;
my $start;
my $nsucc = 0;
for (my $i = 0; $i $#codes; ++$i) {
my $diff = $codes[$i+1] - $codes[$i];
if ($diff == 1 $i + 1 $#codes) {
++$nsucc;
} elsif ($nsucc + 1 = AT_LEAST_THESE_MANY) {
$start = $i - $nsucc;
last;
} else {
$nsucc = 0;
}
}
print @disks[$start..($start+$nsucc)]\n if $start;
-- fxn
--
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
http://learn.perl.org/
Re: to test maximum consecutive numbers
On Dec 11, 2007, at 9:53 PM, ciwei2103 wrote:
Give a disk sequence know that is incrementing with base 16, such
as ,
0001
0003
0004
0008
0009
000A
000B
000F
0010
0011
how do I extract the a subset that is have at least 4 consecutives,
such as 0008 0009 000A 000B
Sorry, the previous code missed an edge case, I think this one is
correct:
use strict;
use warnings;
use constant AT_LEAST_THESE_MANY = 4;
my @disks = qw(
0001
0003
0004
0008
0009
000A
000B
000F
0010
0011
);
my @codes = map hex, @disks;
my $start;
my $nsucc = 0;
for (my $i = 0; $i $#codes; ++$i) {
my $diff = $codes[$i+1] - $codes[$i];
++$nsucc if $diff == 1;
if ($nsucc + 1 = AT_LEAST_THESE_MANY ($diff != 1 || $i + 1 ==
$#codes)) {
$start = $i - $nsucc;
++$start if $diff == 1;
last;
} elsif ($diff != 1) {
$nsucc = 0;
}
}
print @disks[$start..($start+$nsucc)]\n if $start;
-- fxn
--
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
http://learn.perl.org/
Re: fixed list combinatorics
On Nov 29, 2007, at 2:06 AM, Xavier Noria wrote: Indeed, the iterator provided by Algorithm::Combinatorics is faster only for lists of sizes = 7. (And gets to be twice as fast for size 16.) Certainly there's room for improvement here. For the archives, I copied the iterator in List::PowerSet and rewrote it in XS for Algorithm::Combinatorics, that's up in the 0.25. -- fxn -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] http://learn.perl.org/
Re: fixed list combinatorics
On Nov 28, 2007, at 8:58 PM, yitzle wrote: In a personal email conversation, he realized what he was actually looking for is the power set. List::PowerSet http://search.cpan.org/~nikc/List-PowerSet-0.01/lib/List/PowerSet.pm If speed is an issue Algorith::Combinatorics provides subsets() in XS: http://search.cpan.org/~fxn/Algorithm-Combinatorics-0.24/Combinatorics.pm -- fxn Disclaimer: That one is mine but that does not matter. -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] http://learn.perl.org/
Re: fixed list combinatorics
On Nov 28, 2007, at 9:54 PM, Xavier Noria wrote: On Nov 28, 2007, at 8:58 PM, yitzle wrote: In a personal email conversation, he realized what he was actually looking for is the power set. List::PowerSet http://search.cpan.org/~nikc/List-PowerSet-0.01/lib/List/PowerSet.pm If speed is an issue Algorith::Combinatorics provides subsets() in XS: I had done some benchmarks which are included in the distribution but not this one. I've found that List::PowerSet outperforms Algorithm::Combinatorics' subset(). Time to revise that subroutine. -- fxn -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] http://learn.perl.org/
Re: fixed list combinatorics
On Nov 28, 2007, at 10:37 PM, Xavier Noria wrote:
On Nov 28, 2007, at 9:54 PM, Xavier Noria wrote:
On Nov 28, 2007, at 8:58 PM, yitzle wrote:
In a personal email conversation, he realized what he was actually
looking for is the power set.
List::PowerSet
http://search.cpan.org/~nikc/List-PowerSet-0.01/lib/List/PowerSet.pm
If speed is an issue Algorith::Combinatorics provides subsets() in
XS:
I had done some benchmarks which are included in the distribution
but not this one.
I've found that List::PowerSet outperforms Algorithm::Combinatorics'
subset(). Time to revise that subroutine.
Indeed, the iterator provided by Algorithm::Combinatorics is faster
only for lists of sizes = 7. (And gets to be twice as fast for size
16.)
The subroutine that gives you all the subsets in one shot has its own
implementation in List::PowerSet (it is not a wrapper around the
iterator). It is recursive and fast, and clever! I think it is that
fast because it actually runs basically in C:
[ map { [$first, @$_ ], [ @$_] } @$pow ];
given that map is an opcode.
For that call Algorithm::Combinatorics matches and performs slighly
better than List::PowerSet from size 14 on. I guess that's because it
creates less intermediate stuff.
Certainly there's room for improvement here.
-- fxn
--
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
http://learn.perl.org/
Re: Writing DOS CRLF via Unix Perl
On Nov 8, 2007, at 9:33 PM, C.R. wrote: I run a script on unix Perl to write a text file. By default, when Perl writes \n it writes a line ending sequence which is native to the current OS. How do I force this particular script to always write DOS CRLF line endings? A good approach is to hard-code CRLF: my $CRLF = \015\012; and then output that by hand the same way you would append \n: print foo bar baz$CRLF; You don't mention the runtime platform. If you are sure the runtime platform uses Unix line-ending conventions you're done. If you need the script to be portable or don't want to leave there that brittle assumption you need to work in binmode. Otherwise, since CRLF has a LF, on Windows you'd end up having triple CRCRLF on disk, because the I/O layer translates LF - CRLF on writing no matter the surrounding characters. Even if I weren't the author :-) I'd add a pointer to this article: Understanding Newlines http://www.onlamp.com/pub/a/onlamp/2006/08/17/understanding-newlines.html -- fxn -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] http://learn.perl.org/
Re: peek next line in file
On Sep 27, 2007, at 1:29 AM, Mahurshi Akilla wrote:
Is there a way in perl to peek the next line in the file while keeping
the line pointer the same?
I want to do something like this:
while (INFILE)
{
//do some stuff
//?? peek next line ?? and enter conditional block//
//do some more stuff
}
You could do this:
my $pos = tell $fh;
my $next_line = $fh;
seek $fh, $pos, 0;
-- fxn
--
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
http://learn.perl.org/
Re: replace existing text within multiple text files: How?
On Sep 11, 2007, at 9:43 PM, Gerald Wheeler wrote:
I have about 400 text files I need to replace the contents. All files
are in the current directory as the perl script runs out of
my $newText = This is my new text.. anybody's text goes here;
open(INFILE, $plants)
while ()
{
print $newText;
}
I need help..
You mean $newText contains the whole content of a single file, and
the 400 files end up having that very same content (that is $newText)?
-- fxn
--
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
http://learn.perl.org/
Re: replace existing text within multiple text files: How?
On Sep 11, 2007, at 10:05 PM, Gerald Wheeler wrote:
Correct
That can be done with a one-liner:
perl -0777 -pi.bak -e '$_ = q{new text goes here}' *.txt
The options -p, -i, and -e are documented in perlrun. The flag -0777
has the side-effect of slurping the whole file into $_ (one file at a
time), it is documented in perlrun as well, under -0.
-- fxn
--
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
http://learn.perl.org/
Re: String question
On Sep 7, 2007, at 10:07 AM, Santana wrote: Hei all, i'am a newbie in PERL and i find a solution for this problem : I have a string xxx , i want put ones(1) on left of string, if this string dont have a length of 20 character. Example : if i have th string HELLO and woul like get this ; 111HELLO How i can this ??? Is this homework? -- fxn -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] http://learn.perl.org/
Re: how to remove ^M character from every line
Those ^Ms are \015s, which is \r practically everywhere. Note that line-oriented scripts almost work on CRLF text files on Unix because CRLF has a trailing \n by chance. I think this is unfortunate because instead of just breaking your program there's the annoying situation where you still read lines but there's something spurious at the end you don't completely understand. If the script is line-oriented, to read CRLF text files under Unix it is enough to set $/ = \r\n; and chomp each line as usual. If you want to preserve newlines instead of chomping then s/\r\n/\n/ instead. -- fxn -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] http://learn.perl.org/
Re: substitution key-value
On Aug 27, 2007, at 10:59 AM, Petra Vide Ogrin wrote:
Hi all,
I have a hash and some prose text and want my perl to identify the
keys of
the hash in this text and replace them with the corresponding
values of
the keys.
I tried the following
foreach (keys %expan) {
if ($sbl =~ m/$_/g) {
$sbl =~ s/$_/$expan{$_}/g;
No need for the previous check, if it can't perform the substitution
it won't.
That assumes no key is part of another key's value, if it was the
case you could end up doing double/triple/... substitutions. I would
switch to a one-shot (untested) substitution:
my $re = join '|', map quotemeta, keys %expan;
$sbl =~ s/\b($re)\b/$expan{$1}/g;
I don't know why are you applying that substitution, but smells like
a use case for some templating system like Text::Template. If it is
please consider basing the script on it.
-- fxn
--
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
http://learn.perl.org/
Re: substitution key-value
On Aug 27, 2007, at 1:29 PM, Chas Owens wrote: Bad idea*, at least until Perl 5.10 (and maybe not even then). Well, it may or may not be a bad idea. On the one hand that performance penalty may be negligible for the OP problem and thus it just does not matter. On the other hand your solution assumes you can loop through the keys and perform global substitutions, working on the last modified string in each iteration. Assuming you don't want higher-order substitutions (so to speak) which is suggested by the fact that the OP worked in a list of keys in an unkown order, that solutions needs there are no possible clashes between values and keys. And I think you don't need to go to the trouble of checking that or even depend on that, the OR-based solution is robust to that automatically. -- fxn -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] http://learn.perl.org/
Re: Convert int for string
On Aug 26, 2007, at 3:22 AM, Randal L. Schwartz wrote: Perl is a very strongly typed language. The problem is that people keep thinking number or string is a type in Perl. It isn't. The type is called scalar. Other types are array and hash and filehandle and dirhandle and built-in object and user-defined object. In that line, I think Perl is not as dynamically typed as other scripting languages, in the sense that sigils do put types in the source code somehow. Since Perl tries to give meaning almost to anything and context allows to add a scalar to an array, the compiler does not complain that much in practice. But from a formal point of view there are types in the code. -- fxn -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] http://learn.perl.org/
Re: Convert int for string
On Aug 26, 2007, at 3:52 PM, Peter Scott wrote: The term strong typing is so ill-defined as to make this an angels-dancing-on-a-pinhead discussion and unlikely to lead to any enlightenment. Even the Wikipedia definition says that there *is* no accepted definition and some of the common usages contradict each other. I remember MJD has a nice slice about it. I don't know whether it's online though. -- fxn -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] http://learn.perl.org/
Re: Convert from unix newlines to dos newlines
On Aug 25, 2007, at 4:54 AM, Yoyoyo Yoyoyoyo wrote: I use a mac and I was wondering if there was a way to convert unix newlines in a text file to dos newlines. Yeah, with a Perl one-liner it would be perl -pi -we 's/\n/\r\n/' file.txt or perl -pi.bak -we 's/\n/\r\n/' file.txt if you want a backup of the original file. -- fxn PS: For the archives, note those solutions assume the runtime platform is Unix. -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] http://learn.perl.org/
Re: Convert from unix newlines to dos newlines
On Aug 25, 2007, at 4:53 PM, Yoyoyo Yoyoyoyo wrote: Thanks, but quick question. If I do it from the command line it works fine. But if I add: `perl -pi -we 's/\n/\r\n/' ./student.csv`; to my perl script it doesn't make the change to the file. Is there a reason for this? Sure, the literal between backquotes has the same semantics as strings in double quotes, there's interpolation and \t, \n, and friends are translated to their actual meaning. So, the shell sees a hard newline after s/, as if you pressed the return key in the command-line. To fix that double the backslashes, the shell needs to see a \n verbatim, so you need to write \\n in that command. (That's regular English quotes, not Perl double quotes.) If you are already in a Perl script you could do that much more easily without shelling out, for example using Tie::File: use Tie::File; tie my @csv, 'Tie::File', 'student.csv' or die $!; $_ .= \r for @csv; untie @csv; Shelling out is OK though, Perl's about glueing, even to itself :-), you can choose the approach you prefer. -- fxn -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] http://learn.perl.org/
Re: How to create a package level variable
On Aug 24, 2007, at 11:58 AM, Sundeep Gupta wrote: The problem was that this module (XYZ) module uses another module, which again has the use statement to load XYZ. I don't know if the perl loads the module again and might be this caused to reset the value back to 0. When I commented the use statement of that module, it just works fine. If you load a module with use(), it will only get loaded once, no matter how many use()s for that module are executed throughout the source tree. In particular the package variable is not reset: [EMAIL PROTECTED]:~/tmp$ cat Foo.pm use strict; package Foo; our $x = 0; 1; [EMAIL PROTECTED]:~/tmp$ perl -MFoo -wle '$Foo::x = 1; use Foo; print $Foo::x' 1 -- fxn -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] http://learn.perl.org/
Re: How to create a package level variable
On Aug 24, 2007, at 8:13 PM, Dr.Ruud wrote: Chas Owens schreef: [$db_initiali.ed] Unless you have some funky source filter installed that normalizes spelling variants Perl is going to have a problem. Yes, I like the idea: use autocorect; I think there's an Acme:: module that did something like that, it used some heuristic (edit distance or whatever) and chose an already existing indentifier. Something like that. But I can't find it. -- fxn -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] http://learn.perl.org/
Re: perl include dependencies?
On Aug 21, 2007, at 3:37 PM, infobank wrote: I'm trying to embed djabberd (perl5 net-im) onto a m0n0wall base (FreeBSD). I found scandeps, and ran it on DJAbberd and found some useful information about the modules it relies upon. Is there any way to find out which files it relies upon? You mean the files corresponding to those modules? If yes, why you need them? -- fxn -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] http://learn.perl.org/
Re: Array modification
On Aug 16, 2007, at 11:47 AM, Sayed, Irfan (Irfan) wrote: I have one array which stores some data after executing specific command. Depends on situation , command has different output at different time. sometime array may store 4 values or it may store 5 values. Now my req. is that I need to assign no. to those values. for example: if array is @array1=(data1,data2,data3,data4); now i need to assign 4 no. as there are four elements are present in this array. So my array should look like this. @array1=(1 data1 2 data2 3 data3 4 data4); basically, i wanted to do indexing or to create hash. Since a hash is an option, to understand your question better I would like to ask whether the already available builtin array indexing $array[3] is enough. If not, why? -- fxn -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] http://learn.perl.org/
Re: One liner to change one line
On Aug 15, 2007, at 6:45 PM, Dennis G. Wicks wrote:
Greetings;
I have, conservatively, dozens of html files to change.
I can find them and pass the file name to perl and
do the usual s/// changes but there is one change I can't
figure out.
There is a line in each file that looks like
H1This-Is-The-Title/H1
of course, they are all different!
How can I change the hyphens to spaces in this line only?
Complicating the task is:
1. I don't know that there is only one such line per file.
I need to get them all.
2. I don't know that all H1 are upper case.
3. Not all of the H1 lines are the same record in the file.
I think this satisfies those constraints:
perl -0777 -pi.bak -we 's{(h1)(.*?)(/h1)}{$x = $2; $x =~
tr:-: :; $1$x$3}geis' *.html
We slurp the file with -0777 to be able to work across lines (I
undestand (3) that way). Then we capture stuff between H1s case
insensitive, tr/// what's in between. Note that $n-variables are read-
only, that's why we copy the capture into a regular variable.
Of course that assumes a simple regexp is enough, you need to judge
whether that's the case in your data set.
-- fxn
--
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
http://learn.perl.org/
Re: One liner to change one line
On Aug 15, 2007, at 7:04 PM, Xavier Noria wrote:
perl -0777 -pi.bak -we 's{(h1)(.*?)(/h1)}{$x = $2; $x =~
tr:-: :; $1$x$3}geis' *.html
A small improvement, groups are unnecessary because the elements are
guaranteed not to have hyphens (in general they could, for instance
in a class name, but in this case they don't):
perl -0777 -pi.bak -we 's{h1.*?/h1}{$x = $; $x =~ tr:-: :; $x}
geis' *.html
-- fxn
--
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
http://learn.perl.org/
Re: die and exit
On Aug 13, 2007, at 9:29 AM, Jeff Pang wrote:
Does die call exit when it's excuted?
If so,when I overwrote exit() function in my script,would die call
this customized
exit?
Do you want some code executed if the program dies? If that's the
case assign a coderef to $SIG{__DIE__}:
$SIG{__DIE__} = sub {
my $message = shift;
...
};
See %SIG in perlvar, and see the bottom of perldoc -f die.
-- fxn
--
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
http://learn.perl.org/
Re: die and exit
On Aug 13, 2007, at 1:57 PM, Jeff Pang wrote:
Yes that's fine.
I think when we say 'use Apache qw/exit/' in modperl scripts,it may
do the same things as you mentioned -- modify the typeglob directly
-- but I'm not so sure.
I don't know too much XS but looks like this code in src/modules/perl/
mod_perl.c does that (the comment says so as well):
/* *CORE::GLOBAL::exit = \Apache::exit */
if(gv_stashpv(CORE::GLOBAL, FALSE)) {
GV *exitgp = gv_fetchpv(CORE::GLOBAL::exit, TRUE, SVt_PVCV);
GvCV(exitgp) = perl_get_cv(Apache::exit, TRUE);
GvIMPORTED_CV_on(exitgp);
}
-- fxn
--
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
http://learn.perl.org/
Re: Problem with my code
El Aug 3, 2007, a las 1:45 PM, Mr. Shawn H. Corey escribió:
But the expression $hash{@cdr[2,3,6,7]} is the same as $hash{join
($,@cdr[2,3,6,7])} You have just replaced one special variable
with another and $ is a bad choice since it's default is a space
character, which may easily appear in one (or all) of @cdr[2,3,6,7]
A better choice is $hash{$cdr[2]}{$cdr[3]}{$cdr[6]}{$cdr[7]} but
you would need four for-loops to get to the value.
To be able to use arrays as keys there are a couple of modules out
there. One is mine, based on pack/unpack :
http://search.cpan.org/~fxn/Hash-MultiKey-0.06/MultiKey.pm
and there's another one that is implemented as nested hashes, but I
can't remember its name right now, does anybody know it?
-- fxn
--
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
http://learn.perl.org/
Re: Convert Scientific Notation to decimal equivalent
El Jul 19, 2007, a las 12:19 AM, Joseph L. Casale escribió:
Interesting,
I see from your regexp you use a \A and \z, from Perldoc this means:
\A Match only at beginning of string
\z Match only at end of string
I am not sure I understand this requirement?
^ and $ depend on flags, and $ allows an optional trailing newline.
When I want to match a complete string exactly, I tend to use \A and
\z because they convey that intention clearly.
If your code processes line by line and splits on whitespace, \A ...
\z is equivalent to ^ ... $.
In my case, I am checking an array of 3 scalars. Does this make sense:
next unless @data =~ /$RE {num}{real}/;
Does the regexp know to evaluate each element in the array
implicitly? Or do I need to tell it this?
Detecting whether something holds in an array is the job of grep:
my $numbers = grep /\A$RE{num}{real}\z/, @data;
next unless $numbers == @data;
-- fxn
--
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
http://learn.perl.org/
Re: Convert Scientific Notation to decimal equivalent
El Jul 18, 2007, a las 11:19 PM, Joseph L. Casale escribió:
How can I detect this, I have been running some code for a few days
to develop some files and ran into the situation where I am getting
the following data for input:
14.95313 14.45312 0
14.95313 1.570813E-015 0
14.95313 -14.45313 0
-14.95313 -28.90625 0
-14.95313 -14.45313 0
-14.95313 1.570813E-015 0
-14.95313 14.45312 0
14.95313 -28.90625 0
0 -28.90625 0
-14.95313 28.90625 0
0 28.90625 0
14.95313 28.90625 0
And my code is skipping some lines as it checks for any erroneous
data:
next if grep (/[^0-9.-]/, @data);
But that thinks the scientific notation is bad. I searched the net
and didn't find anything. How can I match this specific pattern and
convert it?
I am not sure I understand the problem to solve.
You need to filter out lines that contain something that's *not* a
number? If that's the case, is @data a split on whitespace for each
line? If that's the case in turn, have a look at
perldoc -q determine
or delegate the job to Regexp::Common:
$ perl -MRegexp::Common -wle 'print 1 if 1.570813E-015 =~ /\A$RE
{num}{real}\z/'
1
-- fxn
--
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
http://learn.perl.org/
Re: interpolation of function reference in a here doc
On Jul 2, 2007, at 2:46 PM, Gabriel Striewe wrote:
Dear List,
I wanted to interpolate a function reference in a here doc.
The following works fine:
my $hello = sub {
return hello world!;
};
printf hello $s\n, $hello();
In Perl printf is rarely used because double-quote strings allow
interpolation of scalars and arrays, and because print accepts an
arbitrary number of arguments.
Interpolation does not understand function calls, though, so you
either use a multi-argument call like this:
print hello , $hello-(), \n;
or either use this hack:
print hello @{[ $hello-() ]}\n;
Here-documents have double-quote semantics unless you put single
quotes around the terminating string, so that last trick works in
here-documents as well.
-- fxn
--
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
http://learn.perl.org/
Re: Redirecting file handler to STDOUT
On Jun 22, 2007, at 11:57 AM, Ben Edwards wrote:
I am opening a log file:
open( LOGFILE, cronlog.txt );
This is being written to in lots of places.
I have been asked to change the program so if -m (manual) flag is
passed the stuff that goes to the log file is send to standard out
instead. Is it possible to change the above command to redirect
LOGFILE to STDOUT (i.e. make the two the same thing.
Given those requirements, a simple solution would be:
unless (m_flag_given) {
open my $fh, 'cronlog.txt' or die $!;
select $fh;
}
# the rest of the program uses regular prints
-- fxn
--
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
http://learn.perl.org/
Re: regexp problem
On Jun 14, 2007, at 12:04 PM, Jorge Almeida wrote: I'm missing something about Perl's regexp: 1 #!/usr/bin/perl -w 2 use strict; 3 my $s=STDIN; 4 $s=~s/\D*//; 5 $s=~s/\D*//; 6 print $s\n; When input is 'a123b', I get '123b', but I expected '123'. s/\D+//g; -- fxn -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] http://learn.perl.org/
Re: regexp problem
On Jun 14, 2007, at 12:10 PM, Martin Barth wrote: On Thu, 14 Jun 2007 11:04:51 +0100 (WEST) Jorge Almeida [EMAIL PROTECTED] wrote: I'm missing something about Perl's regexp: 1 #!/usr/bin/perl -w 2 use strict; 3 my $s=STDIN; 4 $s=~s/\D*//; 5 $s=~s/\D*//; 6 print $s\n; When input is 'a123b', I get '123b', but I expected '123'. I know I can substitute line 4 by '$s=~s/\D*//g;' and comment out line 5. It will work then, but that is not the point. I could also substitute line 5 by '$s=~s/\D+//;' and it would also work... the problem is * in your regex. i guess the \D* matches on zero non digits before 123. so nothing will be done. Not really. Although + is more concise because you want to delete non-empty sequences of non-digits and do not care about empty sequences, * will go ahead as much as possible and would work fine. Jorge was just missing /g. -- fxn -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] http://learn.perl.org/
Re: regexp problem
On Jun 14, 2007, at 12:49 PM, Jorge Almeida wrote: Martin was right (and I should have seen it from the start). The will go ahead as much as possible is true only in the sense that the greatest possible string of non-digits will be selected for deletion. With '$s=~s/\D*//;' in line 5 that string will be empty. In order to delete the trailing string, the '/g' is needed (but I knew that, as I said in the original post). Oh I am sorry, I replied too quick and didn't read that part. The use of two s/// in a row hinted to me that there was a naive use s/// there missing /g. Clearly that was not the case and the question was different. Sorry! -- fxn -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] http://learn.perl.org/
Re: Is there a perl equivalent to PHP variables $_POST and $_GET?
On Jun 10, 2007, at 9:18 PM, On Ali wrote: Is there a perl equivalent to PHP variables $_POST and $_GET? Perl is a general-purpose programming language and does not have web stuff builtin. To do web programming you need to pick some library/ framework like CGI.pm or Catalyst for example. -- fxn -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] http://learn.perl.org/
Re: Removing decimal points
On Jun 8, 2007, at 9:52 PM, ash wrote: I need to remove decimal points from numbers. For eg 1.23 or 1.77 would be just 1. Any suggestion is appreciated. Thank you. Use de int() function. -- fxn -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] http://learn.perl.org/
Re: Declaring constants in perl
On Jun 8, 2007, at 1:33 PM, Nath, Alok (STSD) wrote: Hi, What is the convention used to declare constants in perl ? In C we declare constant type in Capital letters , not sure how its in perl. Same convention: use constant FOO = 3; -- fxn -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] http://learn.perl.org/
Re: Algorithm or module needed
On Jun 1, 2007, at 4:29 PM, Beginner wrote: Hi, I am trying to copy some file from a removable HDD to a network drive. I want replicate the directory structure and omit un-needed files. The directory paths a 4-5 levels deep and I am trying to work out an efficient way to recreate the directory path. What I've been doing is using File::Find to collect the file and then in the coderef calling the following subroutine to check the path but i didn't go too far as my method seemswell not well formed. The path to a source file looks like: e:/Dir1/Day 1/AM/01/myfile.jpg I think you want File::Path's mkpath. File::Path is a standard module. -- fxn -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] http://learn.perl.org/
Re: Assign a delimiter variable
On May 15, 2007, at 6:42 PM, Mike Blezien wrote: Hello, this one has been driving me nuts and I'm sure it's something very simple I maybe overlooking. I need to assign a delimiter variable IE: Pipe or Comma delimiter: my $del = '|'; # use either a '|' or ',' my $dataline = 0|1|2|3|4|5|6|7|8|9; my @data = split(/$del/, $dataline); This does not work, it won't split the file line with the '|' delimiter, and get no errors. But if I do this: my $dataline = 0|1|2|3|4|5|6|7|8|9; my @data = split(/\|/, $dataline); Then it works prefectly, it splits the line as expected. What am I missing ?? The actual regexp is what you get _after_ interpolation. Since the pipe is a metacharacter it is being interpreted as such, as if you directly wrote split /|/, ... To prevent this there's quotemeta(), which is available in literals as \E: my @data = split(/\E$del/, $dataline); -- fxn -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] http://learn.perl.org/
Re: Assign a delimiter variable
On May 15, 2007, at 6:49 PM, Xavier Noria wrote: To prevent this there's quotemeta(), which is available in literals as \E: Oh sorry, I meant \Q there. -- fxn -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] http://learn.perl.org/
Re: call system command
On May 14, 2007, at 3:44 PM, Tatiana Lloret Iglesias wrote: Hi all, I have to execute this command from perl: my $status = system(d:\\blast\\bin\\blastall -p blastn -i $file -d $patDB -o $workdir\\blast_$blast_file_id.txt); but the problem is that $workdir contains spaces how can I make it work? Break that into a list of arguments: system(d:\\blast\\bin\\blastall, -p, blastn, -i, $file, ...); -- fxn -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] http://learn.perl.org/
Re: call system command
On May 14, 2007, at 3:52 PM, Tatiana Lloret Iglesias wrote: Thank a lot! Another related question,,, system command can be used also for linux? Sure. Of course it is unlikely that the arguments themselvels are portable in practice, I mean blastall won't probably be located at d: \\blast\\bin\\blastall on Linux, but you can certainly rely on having system(). -- fxn -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] http://learn.perl.org/
Re: cpan question
On May 14, 2007, at 3:50 PM, Robert Hicks wrote: I am working on a Perl 5.6.0 project and went to use cpan at the command line and behold it was not there. I download the 5.6.1 tar file and behold cpan the command line program wasn't in there either. Did it not come with 5.6.0/5.6.1? I don't see where I can get it either... Nope, cpan(1) was added in 5.8.1. You can use the traditional perl -MCPAN -e 'install Module' with that Perl. -- fxn -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] http://learn.perl.org/
Re: Detecting whether a script was called via CGI or the command line
On May 4, 2007, at 2:44 PM, Nigel Peck wrote:
Within the script I want to know where it was executed from; CGI or
command line/cron job. I'm currently checking @ARGV to do this i.e.
if there is data in @ARGV then it was called from the command line,
but obviously this means there must be command line arguments and
I'm also not sure whether there could sometimes be arguments in
@ARGV when called through CGI?
Is there a better way to check?
The proper way would be
use constant RUNNING_AS_CGI = exists $ENV{GATEWAY_INTERFACE};
-- fxn
--
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
http://learn.perl.org/
Re: basename equivalent?
On Apr 26, 2007, at 2:34 AM, Nishi wrote: What is the equivalent of basename? Ie if I dont want to use basename package to get the filename off a path, what reg expr can i use in perl? File::Basename is a standard module, why you don't want to use it? -- fxn -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] http://learn.perl.org/
Re: Removing lines
On Apr 15, 2007, at 6:38 PM, Mário Gamito wrote: I have this site that auto generates an index.html file every 15 minutes (it's a blog aggregator). I need that every time the file is generated, all the contents between the lines h4 class=post-titlea href=http://domain.com/2006/08/bourne-shell.html;Bourne Shell/ a/h4 and pa href=http://domain.com/2006/08/bourne-shell.html;Sáb, 14 Abr 2007 12:31:07/a/p to be deleted (including these two ones). I've tried several ways and googled but i can't do the trick. Any help would be appreciated. Have a look at the range operator in scalar context, it is documented in perlop. The idea is while () print unless /matches initial h4 line/ .. /matches ending P line/; } -- fxn -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] http://learn.perl.org/
Re: Uninstalling perl module
On Apr 13, 2007, at 1:02 PM, Randal L. Schwartz wrote: No, because the CPAN.pm shell is an installer, not a packager. It does not maintain a list of which files belong to which installations, and will happily let two installations both write to the same file. This will be problematic when you want to remove one of those installations later. If you want package management, you'll have to use a distribution (like many Linux and BSD distributions) that have taken the extra time to package the CPAN distributions. That has its own downsides, of course. I am not familiar with module installations, let me ask a bit more to understand those limitations. Wouldn't the problem with file name collisions/rewrites potentially happen if packages were managed as in distros? Or do they have some sort of policy about unicity of files in packages? Besides that, if CPAN.pm cached the MANIFESTs of modules installed by it (which if it doesn't, I guess it would be a matter of programming), wouldn't that suffice to offer uninstalling? -- fxn -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] http://learn.perl.org/
Re: Is there any way to use C header files in perl program?
On Feb 20, 2007, at 7:49 AM, Dharshana Eswaran wrote: I need to use a C header file in Perl program in Unix. Is it possible to use the header files from perl in any way? I do not want to use any modules for the same. I do not want to enable the perl program to call a routine in C library through XS. I just want to use the values rather than any code. Can you explain what do you mean by use and values? -- fxn -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] http://learn.perl.org/
Re: DOS to Windows format tr/\n\r\t/ /
On Feb 11, 2007, at 1:36 AM, [EMAIL PROTECTED] wrote: I wonder if there is a format that preserves paragraph breaks while it kills line breaks? How do you define a paragraph break. Why does the subject mentions DOS to Windows. -- fxn -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] http://learn.perl.org/
Re: DOS to Windows format tr/\n\r\t/ /
On Feb 11, 2007, at 9:39 PM, [EMAIL PROTECTED] wrote: Well, here's my problem: I have found a way to wrap a text file down to a narrower column width. But it looses paragraph formatting. What I need is the same thing as converting a file from DOS (line break after each line) to Windows (line break after each paragraph) format, no? I am not sure I understand the problem, but do you know the Text::Wrap standard module in any case? -- fxn -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] http://learn.perl.org/
Re: Removing file extension
On Jan 23, 2007, at 9:55 AM, Dr.Ruud wrote: Use an anchor and a negated character set: s/\.[^.]*$// That solution is broken, think /foo/bar.baz/woo. Correct regexps and other approaches have already been posted. -- fxn -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] http://learn.perl.org/
Re: Removing file extension
On Jan 23, 2007, at 8:22 AM, Saravana Kumar wrote: Hi list, I am trying to remove the extension from the a list of filenames and manipulate the names further. Tried to doing this: $file=~ s/\..*//; The above works fine. I get the result 'filename' if the filename is filename.ext. There are some files whose names are like file.name.ext and the result i get for this is 'file' while the desired result is 'file.name'. Is there a way to fix this? Yes, you need to assert in the regexp dots are not allowed in the extension, for example something like this: $file =~ s/\.\w+$//; Note that no captures are needed. -- fxn -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] http://learn.perl.org/
Re: memory issues?
On Jan 19, 2007, at 5:53 PM, Bertrand Baesjou wrote:
Thank you very much, this is indeed the solution.
The explanation is that when you process lines this way
foreach my $line (FH) { ... }
the readline operator is evaluated in list context and, thus, the
file is slurped into a single list with all the lines. So, first the
file is fully put into memory, and then foreach starts.
In the line-oriented while loop
while (my $line = FH) { ... }
you are fetching line by line[*], and so memory is kept under control
(at least as much in control as the length of the lines in that file,
that wouldn't solve the problem if the file was GBs of data in a
single line, you see how it works).
-- fxn
[*] Under the hood there are actually some buffers, but that's the idea.
--
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
http://learn.perl.org/
Re: Cannot modify read-only value in while loop?
On Jan 19, 2007, at 12:55 PM, Richard Jones wrote:
while ( my $ref = $sth-fetchrow_arrayref() ) { # $ref is arrayref
push @$ref, 1; # line xx
}
dies with 'Modification of read-only value attempted at .. line xx'
I can't see the difference here between the two $ref arrayrefs, but
Perl obviously can. Unfortunately 'use diagnostics' didn't help as
I think I must be missing something obvious.
Yes, that reference has a flag READONLY turned on, this is a dump of
an example I ran to figure this out (attached below):
SV = RV(0x1823abc) at 0x1856984
REFCNT = 1
FLAGS = (TEMP,ROK)
RV = 0x1856894
SV = PVAV(0x185848c) at 0x1856894
REFCNT = 2
FLAGS = (READONLY) -- look here
IV = 0
NV = 0
ARRAY = 0x4d80b0
FILL = 0
MAX = 3
ARYLEN = 0x0
FLAGS = (REAL)
Elt No. 0
SV = NV(0x1813c00) at 0x1856978
REFCNT = 1
FLAGS = (NOK,pNOK)
NV = 1
There are several lines in
http://search.cpan.org/src/TIMB/DBI-1.53/DBI.xs
with this code:
SvREADONLY_on(av);
So the answer is that you get a regular arrayref with that flag
turned on by hand from XS code. That's why it is read-only.
-- fxn
use DBI;
use strict;
use warnings;
use Devel::Peek;
my $dbh = DBI-connect(dbi:SQLite:dbname=foo.db);
$dbh-do(drop table if exists foos);
$dbh-do(create table foos (id integer));
$dbh-do(insert into foos (id) values (1));
my $sth = $dbh-prepare(select id from foos);
$sth-execute;
Dump $sth-fetchrow_arrayref;
--
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
http://learn.perl.org/
Re: comparing hashes
On Jan 13, 2007, at 6:29 PM, xavier mas wrote:
hello list,
I am trying to find if an element in one primary file (transformed
to array)
is included in two other different secondary files (transformed to
arrays,
too); the result is going to be printed as 1 or 0:
According to the code that's 1 or -1.
...
#creating arrays from its text files
@img_array=IMATGES; @dict_array=DICT; @in_array=IN;
#creating hashes from its arrays
foreach $in (@in_array) {chomp($in); $in_hash{$in}= 1;}
foreach $in (@dict_array) {chomp($in);$dict_hash{$in}= 1;}
foreach $in (@img_array) {chomp($in);$img_hash{$in}= 1;}
#searching primary element in secondary hashes
while (($key, $value) = each %in_hash) {
if (exists $dict_hash{$key}) {$dic_flag=1;}else {$dic_flag=-1}
if (exists $img_hash{$key}) {$img_flag=1;}else {$img_flag=-1;}
#printing result
print $img_flag, $dic_flag\n;
A bit of air would improve readability, the code is easy but dense
just because of lack of layout.
but it seems 'exists' function doesn't fly to do this -the element
isn't
always found into the secondary hashes. Any suggestion of why it
doesn't and
how to do it?
Besides de potential mismatch given by the fact that the img hash
has an img flag, but the dict hash has a dic flag (everything
seems correct in that snippet anyway), I see no problem. Could you
please send a minimal, self-contained code with minimal example files
that let us reproduce the issue?
-- fxn
PS: Please turn strict and warnings on.
--
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
http://learn.perl.org/
Re: comparing hashes
On Jan 13, 2007, at 7:43 PM, xavier mas wrote: Here's an example: in (file, array and hash) contains: woman, lion, ball img (file, array and hash) contains: ball, dog, cat, lion. dict (file, array and hash) contains: house, man, woman, kid, kitchen, lion Comparing in with dict ans img, I'll expect as a result (all previous code is between the while curly braces): -1, 1 1, 1 1, -1 but the result is, instead: -1, -1 -1, -1 -1, -1 that means never finds it. I hope this is enough data. It is better but not enough. The way you use hashes is apparently correct, so perhaps the problem comes from elsewhere. Please send self-contained, minimal code that reproduces the issue, not just an incomplete subset. Please turn strict and warnings on. -- fxn -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] http://learn.perl.org/
Re: Line ending with Gary^M on UNIX
This article explains how this works and has this and other tricks: http://www.onlamp.com/pub/a/onlamp/2006/08/17/understanding- newlines.html -- fxn PS: I am in fact the author but that is irrelevant. -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] http://learn.perl.org/ http://learn.perl.org/first-response
Re: calling system
On Oct 21, 2006, at 9:52 AM, xavier mas wrote:
Dear all,
Hi Xavier!
I'm trying to rename a bunch of files to the estination names
included in a
prepared files calling the system function mv inside a while loop
(that
reads the destination files one by one).
But when doing this, I get an error saying a destiantion operand is
missing in
line calling the system. Doesn't matter if I use double (for Perl
interpretaton) or single quotes (for system interpretation).
If doing same thing on system bash line it works. What could be the
reason for
that behavior? I need, anyway through Perl programm due to the big
ammount of
files I have to rename.
...
cont = 0;
while (FILE_IN) {
cont++
I bet the culprit are those barewords, write $cont instead. It is
good to enable strictness and warnings always, they catch lots of
potential bugs like that one.
$file_out = $_;
$file_in = $cont;
system mv $file_in $file_out;}
Shelling out has always its risks. Here we are assuming names do not
contain any space (which may be the case in your script). For the
same price you can robustely use the builtin rename or the move/mv
subroutines from the standard module File::Copy.
-- fxn
--
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
http://learn.perl.org/ http://learn.perl.org/first-response
Re: perl built in function for mean
On Sep 9, 2006, at 9:00 PM, chen li wrote: I want to calculate the mean of an array. I know how to let the job done by using a loop. But I just wonder if Perl has the short-cut/built-in function for this. There are modules, but nothing builtin. -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] http://learn.perl.org/ http://learn.perl.org/first-response
Re: perl built in function for mean
On Sep 9, 2006, at 9:48 PM, Randal L. Schwartz wrote: Xavier == Xavier Noria [EMAIL PROTECTED] writes: Xavier On Sep 9, 2006, at 9:00 PM, chen li wrote: I want to calculate the mean of an array. I know how to let the job done by using a loop. But I just wonder if Perl has the short-cut/built-in function for this. Xavier There are modules, but nothing builtin. List::Util is built-in as of 5.8, and back compatible to 5.5 I think. use List::Util qw(sum); my $average = sum(@input) / @input; I know, but that's not a builtin mean(). -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] http://learn.perl.org/ http://learn.perl.org/first-response
Re: Connecting to Oracle
On Aug 23, 2006, at 7:52 AM, anand kumar wrote: Can anyone please explain how to connect to oracle database server using perl script which is in separate linux server. Please note that currently I am using putty to connect to linux server for perl scripting and we have the Oracle database on a windows server. I have tried using the DBD::Oracle module from cpan.org but i could not install the module because the oracle drivers are accessible to Perl on the Linux server. AFAIK you need to install the Oracle client in the local machine. -- fxn -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] http://learn.perl.org/ http://learn.perl.org/first-response
Re: empty line in REs
On Aug 22, 2006, at 5:36 PM, chen li wrote: Just a quick question: what is the line code for an empty line using regular expression in Perl? Empty lines match /^$/, blank lines do not match /\S/. -- fxn -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] http://learn.perl.org/ http://learn.perl.org/first-response
Re: Help on chomp()
On Aug 21, 2006, at 8:51 PM, Mazhar wrote: use strict; use warnings; my $file_name=$ARGV[0]; open(FILE,$file_name) || die Not been Accessed ; No quotes needed around $file_name there. @host_array=FILE; That variable was not declared before, so that script does not run because it does not compile. Perhaps you are running a program which is not the one you are editing? -- fxn -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] http://learn.perl.org/ http://learn.perl.org/first-response
Re: Perl on Windows Server.
On Aug 16, 2006, at 2:33 PM, Krishnakumar K P wrote: Hi, There is a problem in coding perl in windows and using it in linux. When you code in windows the file will contain dos/windows new line pairs instead of the Linux/UNIX single New Line character which may cause error in linux server. That is unlikely, Perl does not care about the newline convention used in the source code, newlines of any kind are simply discarded. If you hard-code newlines in strings, here-docs, regexps, etc., you'd get a different literal, not necessarily get an error. Take into account anyway that the portable way to write a nelinwe in a string or regexp literal is \n. Any major version control system like Subversion handles newline conventions transparently for you. If you commit a source file edited in Windows 2000, and check it out in FreeBSD, each local copy will have the native conventions. That's why those systems distinguish between binary and text files. Yet another benefit of using a VCS. -- fxn -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] http://learn.perl.org/ http://learn.perl.org/first-response
Re: Perl on Windows Server.
On Aug 16, 2006, at 2:54 PM, Xavier Noria wrote: On Aug 16, 2006, at 2:33 PM, Krishnakumar K P wrote: Hi, There is a problem in coding perl in windows and using it in linux. When you code in windows the file will contain dos/windows new line pairs instead of the Linux/UNIX single New Line character which may cause error in linux server. That is unlikely, Perl does not care about the newline convention used in the source code, newlines of any kind are simply discarded. Sorry that's not exact. Let me add that newlines are the end-of-comment marker. You could get parsing errors if the file uses CRs and is being interpreted in a Unix machine for instance. If the source code uses either LF or CRLF there will be no problem as long as you are not running old Mac computers. Another gotcha has to do with shebangs ending in CRLF, I know no shell that understands them (that's not a Perl issue though). -- fxn -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] http://learn.perl.org/ http://learn.perl.org/first-response
Re: How to sort an array which contains a value several times ?
On Aug 15, 2006, at 2:02 PM, SFantar wrote:
Hello everyone
I would like to sort an array so that it does not contain any
values twice or three times.
Unfortunately, perldoc -q sort does not help me very much.
Here is an example of this array :
@array = ('smith ', 'john ', 'edward ', 'smith ', 'edward ', 'john ');
I would like to have : @array1 = ('smith ', 'john ', 'edward ');
Thanks in advance for your help and especially your explanations.
From the example looks like you want to filter out duplicates but
preserve the relative order of elements in the original array. A
possible way to accomplish that is:
feynman% cat ~/tmp/foo.pl
@array = ('smith ', 'john ', 'edward ', 'smith ', 'edward ', 'john ');
%seen = ();
@uniq = grep { $seen{$_} = !exists $seen{$_} } @array;
print @uniq\n;
feynman% perl ~/tmp/foo.pl
smith john edward
The code in the block for grep is a bit dense, it can be made more
explicit if needed.
-- fxn
--
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
http://learn.perl.org/ http://learn.perl.org/first-response
Re: How to sort an array which contains a value several times ?
On Aug 15, 2006, at 2:28 PM, SFantar wrote:
Thank for your quick reply.
Your solution should suit me.
Can you give me more details/ explanations about your code?
What does your code do exactly?
Sure, we are using a hash table as a way to deal with duplicates,
which is a standard idiom in Perl.
The idea is to use the elements you need to process as keys of a hash
table. Hash tables held just one entry per key, so if you reassign to
a key the entry in the hash table is overwritten, so to speak, not
repeated. That's the way hash tables work, and we leverage that
property.
For instance this is a way to count how many times each element of an
array is repeated:
my %counters = ();
$counters{$_} += 1 for @array;
After that code the keys of %counters are the distinct elements of
@array, and the hash associates the number of occurrences of each one
in @array. Do you understand that example?
In the example I sent before, we used a similar trick but using grep,
because we want to iterate over @array and filter duplicates out (you
need to know grep to understand this):
%seen = ();
@uniq = grep {
my $is_new = 0;
if not exists $seen{$_} {
$seen{$_} = 1;
$is_new = 1;
}
$is_new;
} @array;
That code is more explicit. We use a hash table called %seen to
register seen strings. The way to accomplish that is that we assign
an arbitrary value to $seen{john} the first time we encounter 'john',
and so subsequent calls to exists $seen{john} will return true. In
those cases $is_new remains 0 and grep filters that element out.
I wrote a condensed version of an analogous way to work that out.
-- fxn
--
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
http://learn.perl.org/ http://learn.perl.org/first-response
Re: How to sort an array which contains a value several times ?
On Aug 15, 2006, at 3:42 PM, Xavier Noria wrote:
%seen = ();
@uniq = grep {
my $is_new = 0;
if not exists $seen{$_} {
Oh sorry I wrote that off the top of my head and forgot the parens.
--
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
http://learn.perl.org/ http://learn.perl.org/first-response
Re: How to sort an array which contains a value several times ?
On Aug 15, 2006, at 5:49 PM, SFantar wrote: On the contrary, the code written above is not clear yet. What does your code do exactly when it comes to deal with smith, edward which appear twice in @array initially? What's what you don't understand, do you know grep? -- fxn -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] http://learn.perl.org/ http://learn.perl.org/first-response
Re: How to sort an array which contains a value several times ?
On Aug 15, 2006, at 6:19 PM, SFantar wrote:
I know the grep function.
Normally, grep needs a pattern to match with values found in @array.
All the values found are returned in an array.
Is that right?
Not quite.
The elements of the array are passed to grep's block, one at a time,
and the block is just expected to return some value to be interpreted
as boolean. Whether the value is related to the element or not it
does not matter, grep just takes it an applies its logic: If the
value returned by the block is true the element passes, otherwise the
element is filtered out.
So the first time 'smith' is seen the code creates an entry for it in
%seen and sets $is_new to 1, which is the value returned, and true.
Therefore, the first time we see 'smith' it passes to @uniq. Next
time 'smith' is seen in @array $is_new ends up being 0 because exists
$seen{smith} is true. Now the returned value--- $is_new--- it is
false, so that second 'smith' does not pass. In subsequent 'smith's
the same reasoning applies, none of them pass.
Note that %seen is shared among invocations of the block because it
was declared outside.
-- fxn
--
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
http://learn.perl.org/ http://learn.perl.org/first-response
Re: How to split a file on the CR carriage Return and/or replace the \r with \n ??
On Jul 28, 2006, at 9:50 PM, Joshua Colson wrote: On Fri, 2006-07-28 at 21:28 +0200, Dr.Ruud wrote: Robert Citek schreef: Try a different value for the INPUT_RECORD_SEPARATOR ($/): $/ = \n ; (spaces added for clarity) Is that different from the default? (see perlvar) Using ActiveState on Windows it is. That's wrong. $/ is eq \n everywhere, and \n has length 1 everywhere. \n is eq \012 in all systems except MacOS pre-X where it is eq \015. In particular \n is *not* CRLF on Windows. My article about newlines is moving forward at O'Reilly, I hope I can soon provide a pointer to it. Misconceptions about newlines arise too often, which was my main motivation to write it. -- fxn -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] http://learn.perl.org/ http://learn.perl.org/first-response
Re: How to split a file on the CR carriage Return and/or replace the \r with \n ??
On Jul 26, 2006, at 21:59, Steve Pittman wrote:
I am using activestate on a windows box ...the files I am parsing are
Unix files...I tried this so far...
open ( IN, $_[0] )||die Can't open DAT source file: $tempFile
$!\n;
while (IN){s/\r/\n/g;@lines = IN;}#
close (IN);
foreach $line (@lines)
{
If the file has Unix conventions a line-oriented loop on Windows
works out of the box:
while (my $line = $fh_of_unix_text_file) {
chomp $line;
# ...
}
The reason why this works is that no pair CRLF ever gets into $line
on Windows working in text mode. That's because when reading the I/O
layer substitutes them on the fly with simple LFs. The readline
operator returns $/-separated chunks, and $/ is \n by default,
which is precisely eq to LF on Windows. As you know, LF is the line
separator on Unix.
If the file actually has CR conventions (as said in the subject)
instead of Unix conventions (as said in the body), then just set $/
accordingly and use a regular line-oriented loop as well:
{
local $/ = \015; # tell readline to use CR
while (my $line = $fh_with_cr_convention) {
chomp $line;
# ...
}
}
-- fxn
--
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
http://learn.perl.org/ http://learn.perl.org/first-response
Re: split function
On Jul 19, 2006, at 9:57, Sayed, Irfan ((Irfan)) wrote: I need to split following string cs_backup_restore_cmvobsvr1mum the output which i am looking for is cs_backup_restore and _cmvobsvr1mum Which is the criteria, everything up to the last underscore? -- fxn -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] http://learn.perl.org/ http://learn.perl.org/first-response
Re: split function
On Jul 19, 2006, at 10:05, Sayed, Irfan ((Irfan)) wrote: I think criteria shud be _ but I need output in following manner That criteria is ambiguous becasue there are several _s and you need to deal with them differently, that is, ignoring some and splitting on some. Can you be more specific? -- fxn -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] http://learn.perl.org/ http://learn.perl.org/first-response
Re: doubt..
On Jul 19, 2006, at 13:59, Ankam, Ramesh Babu wrote: Can any one please tell me what this pattern meanss|.*/|| . Assuming there are no newlines in the string, that s/// means remove everything up to, and including, the last slash. It's a typical regexp for getting the basename of a filename. If that's the usage please delegate to File::Basename instead. -- fxn -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] http://learn.perl.org/ http://learn.perl.org/first-response
Re: Next
On Jul 6, 2006, at 6:26, Geetha Weerasooriya wrote: Dear all, When I was reading a Perl code I found the following line. Can u please explain what it means? !defined($rt_nearest) or $dh$dist or next; It means next unless !defined($rt_nearest) or $dh $dist; or, equivalently, next if defined($rt_nearest) and $dh = $dist; The trick in the original code is: the or operator returns as soon as some of the operands evaluates to true, if any. So if !defined($rt_nearest) holds nothing to its right is evaluated. Otherwise we go for $dh $dist If it holds we are done, otherwise we evaluate the following operand next which has the secondary effect of jumping to the next iteration somewhere. Some people dislike using boolean operators to control flow that way, some people don't. -- fxn -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] http://learn.perl.org/ http://learn.perl.org/first-response
Re: Need help with error from certain ASCII characters in a CSV file
On Jul 6, 2006, at 10:41, Roman Daszczyszak wrote: Does anyone have a suggestion for how I can handle this, or even where I can look to solve this issue? Is there another possibility for where the error is occurring that I am not seeing? I would create the filehandle specifying the character encoding in the open call (as documented in perldoc -f open). I didn't test this, just an idea. -- fxn -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] http://learn.perl.org/ http://learn.perl.org/first-response
Re: newlines on win32, old mac, and unix
On Jun 19, 2006, at 22:45, Anthony Ettinger wrote:
# order matters
$raw_text =~ s/\015\012/\n/g;
$raw_text =~ s/\012/\n/g unless \n eq \012;
$raw_text =~ s/\015/\n/g unless \n eq \015;
Does it make any difference if I use s/\cM\cJ/cJ/ vs. s/\015\012/\n/
g ?
The regexp is OK, the replacement string is not, because \cJ is not
necessarily eq \n. The latter is portable, the former is not.
Since the newline convention is not necessarily the one in the
runtime platform you cannot write a line-oriented script. If files
are too big to slurp then you'd work on chunks, but need to check by
hand whether a CRLF has been cut in the middle.
I'm reading each line in a while loop, so it should work fine on a
large file?
The while loops over lines ***as long as they are encoded using the
conventions of the runtime platform***. The diamond operator uses $/
as separator, which in turn is \n by default. Since the purpose of
your script is to deal with *any* newline convention, in general a
while loop like
while (my $line = $fh) { ... }
looks suspicious. The variable should be called $chunk_of_text,
instead of $line. You don't know whether you'll get a line.
Suspicious, may signal the programmer does not fully understand
what's going on.
For instance, TextWrangler is known to use old-Mac conventions by
default (last time I checked). If you read a file like that with that
while in either Unix or Windows you'll slurp the entire file in a
single iteration. That is, $line will contain the whole file.
In general, to be robust to newline conventions you need to to some
munging by hand before using regular, portable line-oriented idioms.
-- fxn
--
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
http://learn.perl.org/ http://learn.perl.org/first-response
Re: Help understand documentation
On Jun 13, 2006, at 6:14, Vijay Kumar Adhikari wrote:
I understand. My question is that Net::Ping does not use that info at
all. If a host is alive, it will be listed in the output no matter
what port you specify. Is this a bug?
The module uses the port in several subroutines, like this
$saddr = sockaddr_in($self-{port_num}, $ip);
but I think there is an errata in the SYNOPSIS indeed, it should read
$p-{port_num} = (getservbyname(http, tcp))[2];
-- fxn
--
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
http://learn.perl.org/ http://learn.perl.org/first-response
Re: newlines on win32, old mac, and unix
On Jun 13, 2006, at 20:26, Anthony Ettinger wrote:
I have to write a simple function which strips out the various
newlines on text files, and replaces them with the standard unix
newline \n
In Perl \n depends on the system, it is eq \012 everywhere except
in MacOS pre-X, where it is \015. The standard Unix newline is \012.
In text-mode, I open the file, and do the following:
while (defined(my $line = INFILE))
{
my $outline;
if ($line =~ m/\cM\cJ/)
{
print dos\n;
($outline = $line) =~ s/\cM\cJ/\cJ/;
#win32
} elsif ($line =~ m/\cM(?!\cJ)/) {
print mac\n;
($outline = $line) =~ s/\cM/\cJ/g; #mac
} else {
print other\n;
$outline = $line; #default
}
print OUTFILE $outline;
}
It works fine on unix when I run the unit tests on old mac files, win,
and unix files and do a hexdump -C on themhowever, when I run it
on win32 perl 5.6.1, it is not doing any replacement. Teh lines remain
unchanged.
My understanding is that \n is a reference (depending on which OS your
perl is running on) to CR (mac), CRLF (dos), and LF (unix) in
text-mode STDIO.
That is a common misconception. The string \n has length 1 always,
everywhere. It is not CRLF on Windows.
To explain this properly I'd need to reproduce here an article I've
written for Perl.com, not yet published. But to address the problem
is enough to say that in text mode there is an I/O layer (PerlIO)
that does some magic back and forth between \n and the native newline
convention. That's the way portability is accomplished, inherited
from C.
To be able to deal with any newline convention the way you want in a
portable way you disable that magic enabling binmode on the
filehandle. The easiest solution is to slurp the text and s/// it
like this (written inline):
binmode $in_fh;
my $raw_text = do { local $/; $in_fh };
# order matters
$raw_text =~ s/\015\012/\n/g;
$raw_text =~ s/\012/\n/g unless \n eq \012;
$raw_text =~ s/\015/\n/g unless \n eq \015;
# $raw_text is normalized here
Since the newline convention is not necessarily the one in the
runtime platform you cannot write a line-oriented script. If files
are too big to slurp then you'd work on chunks, but need to check by
hand whether a CRLF has been cut in the middle.
-- fxn
--
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
http://learn.perl.org/ http://learn.perl.org/first-response
Re: plugins
On Jun 11, 2006, at 11:16, OROSZI Balázs wrote: Chad Perrin wrote: On Sat, Jun 10, 2006 at 10:53:30PM +0200, OROSZI Balázs wrote: Makefiles are for installing modules, not for using them. If you use the CPAN module management that is provided with common Perl distributions, you shouldn't have to deal with makefiles even for installation, unless you wrote the module yourself and are packaging it for CPAN. It seems this is REALLY not what I want :) Now I'm currently using do $file and it is working fine. Maybe one day I'll do it better, but for now it is ok for me. You seem to have some misconception about modules. That do-based approach is very similar to the suggested solution indeed, see perldoc -q require. -- fxn -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] http://learn.perl.org/ http://learn.perl.org/first-response
Re: Help understand documentation
On Jun 11, 2006, at 14:20, Vijay Kumar Adhikari wrote:
This is from http://perldoc.perl.org/Net/Ping.html
# Like tcp protocol, but with many hosts
$p = Net::Ping-new(syn);
$p-{port_num} = getservbyname(http, tcp);
foreach $host (@host_array) {
$p-ping($host);
}
while (($host,$rtt,$ip) = $p-ack) {
print HOST: $host [$ip] ACKed in $rtt seconds.\n;
}
Can some body tell me what the following line does?
$p-{port_num} = getservbyname(http, tcp);
The object in $p is a blessed hashref. That's Perl object-oriented
jargon, if you are not familiar with it think $p is a regular
hashref. In any case,
$p-{port_num} = $value;
assigns $value to the key port_num of the hash %$p.
That usage is a bit unusual, normally the underlying implementation
of an object is not public for clients, and access to its
attributes in the case of hashrefs is normally encapsulated in
method calls.
Anyway, that's what that code means.
-- fxn
--
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
http://learn.perl.org/ http://learn.perl.org/first-response
Re: Newline error
On Jun 6, 2006, at 12:59, joseph wrote: All, Just like to ask for correction on what's wrong with my script it gives spit out this error when i run it. Unsuccessful open on filename containing newline at disksize.pl line 8. Can't open file No such file or directory But it runs without this error whenever i feed it up when an existing file output by df-h. ex:## open(FL,path/toactual/file) or die blabalha; Does this mean it can't trap the output of `df-h`? Dealing with system commands (ls, ps, df, etc.) that way is usually much tricky than it looks. Normally one goes to CPAN and checks whether someone has done it right already. In that case one installs the module and writes an easy and robust script in 5 minutes. For this particular problem delegate to Sys::Filesystem. -- fxn -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] http://learn.perl.org/ http://learn.perl.org/first-response
Re: Newline error
On Jun 6, 2006, at 14:16, John W. Krahn wrote:
Mr. Shawn H. Corey wrote:
$output_file =~ s/\r//g;
# chomp only removes linefeed characters \n
# BTW, there is no such thing as a newline;
# it is either a linefeed: \n ASCII LF 0x0A
# or a carriage return: \r ASCII CR 0x0D
\n is inherited from the C programming language and is the
newline escape
sequence. On Unix and Unix-like systems the newline is equivalent
to the
ASCII line feed character but on other systems it could be one or more
different characters.
To be more precise, \n in Perl is eq \012 everywhere except in
Mac OS pre-X, where it is eq \015. In CRLF platforms like Win32
\n is transparently converted to CRLF on writing and back on
reading by PerlIO in text mode. Thus, in a regular line-oriented
script like
while (my $line = FH) {
# work with $line
}
$line ends with \n but does not contain a pair CRLF (assuming
native conventions in the input). On the other direction, the string
foo\n has length 4 in all systems. When you print that string into
a file in text mode on Windows the bytes on disk have an extra
\015, but that's transparent to the programmer. That's the point of
using \n as logical/portable newline in Perl.
I have written an article about newlines in Perl not yet published.
All those fine details are explained there.
-- fxn
--
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
http://learn.perl.org/ http://learn.perl.org/first-response
Re: monitoring file like tail -f in perl
On May 29, 2006, at 18:19, Ken Foskey wrote: I want to monitor a daemon's logs that updates randomly through the day and night. Is there a way to open the log file and read the new records as they come in the same way as tail -f does? Sure, with File::Tail for example. -- fxn -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] http://learn.perl.org/ http://learn.perl.org/first-response
Re: suppressing selected warning messages
On May 24, 2006, at 18:48, Gavin Bowlby wrote:
I can suppress all warning checks by doing a:
no warnings;
if ($a == 7) {
use warnings;
but I would like all other normal Perl warnings to be used for the
statement:
if ($a == 7)
other than the check for the LHS of the equality check being a numeric
value.
Is this possible?
Warnings have a hierarchy documented in perllexwarn. The category
each warning belongs to is documented in perldiag. In that case the
most you can do is to disable the numeric category, which is the
one that warning belongs according to perldiag:
{
no warnings qw(numeric);
# issues use of uninitialized ... but not the numeric one
print undef if foo == 0;
}
-- fxn
--
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
http://learn.perl.org/ http://learn.perl.org/first-response
Re: Converting a string to a filehandle
On May 22, 2006, at 4:44, Wijaya Edward wrote: Is there a way to do it? Module for it? Suppose I have this large string. my $string = 'foo bar qux woo etc etc'; I would like to convert that string as if it is stored inside a file, and bypassing the file creation step. For example, my $filehandle = do_sth_to_convert($string); Yes, since 5.8.0 that's built-in: open my $filehandle, '', \$string; -- fxn -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] http://learn.perl.org/ http://learn.perl.org/first-response
Re: Extracting a bit of a scalar variable
On Apr 21, 2006, at 20:59, Praveena Vittal wrote: Is there any way to print a particular bit of a digit after converting to binary form. That can be easily done with substr y sprintf, this would take the second bit of 7 written in base 2: % perl -wle 'print substr sprintf(%b, 7), -2, 1' 1 That's quite readable, but if efficiency is an issue then pack/unpack might be better, it had to be measured in any case. -- fxn -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] http://learn.perl.org/ http://learn.perl.org/first-response
Re: regular expressions
On Apr 21, 2006, at 16:10, Bowen, Bruce wrote: In perldoc under this topic s is listed as Treat string as a single line and m as Treat string as multiples lines. If I have text that has varying spaces at the begging of each line, and I use $string =~ s/^\s+//; It will remove the spaces from in from of the first line but not any other lines. That is clear to me. However, it does not clear all of the leading spaces from all of the lines if I use $string =~ m/^\s+//; Modifiers go to the end: $string =~ s/^\s+//m; -- fxn -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] http://learn.perl.org/ http://learn.perl.org/first-response
Re: regex help
On Apr 19, 2006, at 9:13, Murphy, Ged (Bolton) wrote:
Hi all.
I have a log containing strings as follows :
21259 auditO 72398 Mar 09 00:18 dll/ldr/elf.c
The format is the same throughout with the exception of the 'O', as it
doesn't always appear.
I need to match when the 'O' appears and when it does, I need to save
the file path, i.e. 'dll/ldr/elf.c'
Here is a snippet from my code containing my regex
if (/\sO\s.+([\w\/]+)$/) {
print found $1\n;
Based on yours, the regexp would be for instance:
\sO\s.+\s(\S+)$
Note that assumes no other field can be O except in those cases,
you can add restrictions as needed if that does not hold.
You could base that on split as well, it may be more readable:
my @fields = split;
print found $fields[-1]\n if $fields[2] eq O;
-- fxn
--
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
http://learn.perl.org/ http://learn.perl.org/first-response
Re: split function in perl
On Apr 17, 2006, at 10:30, Irfan J Sayed wrote: Hi, I have a following line stored in one variable $test. deliver.Admin_Irfan_Project.20060413.212355 . I need to split this line into the words and store the output in array. words should like this. deliver admin irfan project 20060413 212355 I am using following code to split this line $test =~ s/^\s+//; @array = split(/\W/, $test); Problem is that _ belongs to \w or, equivalently, it does not belong to \W. For that particular example it is enough to include _ in the regexp: % perl -wle 'print for split /[\W_]/, deliver.Admin_Irfan_Project. 20060413.212355' deliver Admin Irfan Project 20060413 212355 -- fxn -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] http://learn.perl.org/ http://learn.perl.org/first-response
Re: Combine similar patterns
On Mar 29, 2006, at 10:28, anand kumar wrote: Sorry for not posting the question clearly, Please find the attachment of the sample file. The matter enclosed in f1…./ f1=c, f1=c…../f1=c….f1=c……/f1 are all the foot notes that are spanning on various pages, now what I am trying to do is that to combine all the related footnotes that are spanned in various pages and place them at on place. Here f1 indicates the starting of the footnotes, f1=c indicates that the footnote continues from previous page, the /f1=c indicates that the foot notes continues in next page and /f1 indicates the end of the footnotes Looks like the code I sent before is a good starting point. The only new relevant stuff I see are the fn tags, but you could adapt the script easily to deal with that correctly and to do what's needed with empty lines. If you understand the flip-flop operator you've got the script. -- fxn -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] http://learn.perl.org/ http://learn.perl.org/first-response
Re: [SPAM DETECT] Re: Combine similar patterns
On Mar 29, 2006, at 12:01, anand kumar wrote: Yes, of course the code u have sent is very helpful and thanks for that. But the problem here is that in a single file there may be various number of foot notes i.e. f1,f2f3.so on. For all these the conventions are the same as told before. I am having problem in combining the relevant ones. Please help in this matter I think we can go back to the simplification. Would you please send an analogous simplified example that contains sections f1, f2, f3, following the same structure as in the original data? -- fxn -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] http://learn.perl.org/ http://learn.perl.org/first-response
