avi@taux01.nsc.com (Avi Bloch) (01/28/91)
How do I read an entire file into a string other than the obvious way of reading it line by line and concatenating it. Please email since I don't usually have the time to read this newsgroup. Thanks Avi Bloch National Semiconductor (Israel) avi@taux01.nsc.com -- Avi Bloch National Semiconductor (Israel) 6 Maskit st. P.O.B. 3007, Herzlia 46104, Israel Tel: (972) 52-522263 avi@taux01.nsc.com
lixj@acf4.NYU.EDU (Xiaojian Li) (01/30/91)
> >How do I read an entire file into a string other than the obvious way of >reading it line by line and concatenating it. > >Please email since I don't usually have the time to read this newsgroup. > >Thanks > >Avi Bloch >National Semiconductor (Israel) >avi@taux01.nsc.com >-- > Avi Bloch I am new to perl(just two days old :-). Well, I have to say the manual is very hard, or impossible to read for non expert. Anyway, I do find by accident that the following might solve your problem: #!/bin/perl open(TMP, "$ARGV[0]"); @array = ( <TMP> ); # put the whole file into an array. I don't know if # this is correct, but it seems to work. Correct me if # I am wrong. chop(@array); print @array, "\n";
lixj@acf4.NYU.EDU (Xiaojian Li) (01/30/91)
From lixj@acf4.NYU.EDU Tue Jan 29 17:34:44 1991 Received: from ACF4.NYU.EDU by cmcl2.NYU.EDU (5.61/1.34) id AA00907; Tue, 29 Jan 91 17:34:44 -0500 Received: by acf4.NYU.EDU (5.61/1.34) id AA11782; Tue, 29 Jan 91 17:34:21 -0500 Date: Tue, 29 Jan 91 17:34:21 -0500 From: lixj@acf4.NYU.EDU (Xiaojian Li) Message-Id: <9101292234.AA11782@acf4.NYU.EDU> To: rnews@cmcl2.NYU.EDU Relay-Version: version nyu B notes v1.6.1 1/11/90; site acf4.NYU.EDU From: lixj@acf4.NYU.EDU (Xiaojian Li) Date: 29 Jan 91 17:34 EST Date-Received: 29 Jan 91 17:34 EST Subject: Re: Reading a file into a string Message-ID: <2540002@acf4.NYU.EDU> Path: acf4!lixj Newsgroups: comp.lang.perl Posting-Version: version nyu B notes v1.6.1 1/11/90; site acf4.NYU.EDU Organization: New York University References: <5231@taux01.nsc.com> > >How do I read an entire file into a string other than the obvious way of >reading it line by line and concatenating it. > >Please email since I don't usually have the time to read this newsgroup. > >Thanks > >Avi Bloch >National Semiconductor (Israel) >avi@taux01.nsc.com >-- > Avi Bloch I am new to perl(just two days old :-). Well, I have to say the manual is very hard, or impossible to read for non expert. Anyway, I do find by accident that the following might solve your problem: #!/bin/perl open(TMP, "$ARGV[0]"); @array = ( <TMP> ); # put the whole file into an array. I don't know if # this is correct, but it seems to work. Correct me if # I am wrong. chop(@array); print @array, "\n";
tchrist@convex.COM (Tom Christiansen) (01/30/91)
From the keyboard of avi@taux01.nsc.com (Avi Bloch): :How do I read an entire file into a string other than the obvious way of :reading it line by line and concatenating it. I answered in mail, but for the record: 1) $foo = `cat $file`; 2) open (FILE, $file); undef $/; $foo = <FILE>; close(FILE); The cat method costs a bit more, but both are more costly than iterating over the file a line at a time. --tom -- "Hey, did you hear Stallman has replaced /vmunix with /vmunix.el? Now he can finally have the whole O/S built-in to his editor like he always wanted!" --me (Tom Christiansen <tchrist@convex.com>)
tneff@bfmny0.BFM.COM (Tom Neff) (01/30/91)
Sorry to requote all 5, but Allah forbid someone gets this out of order :-) In article <11235@jpl-devvax.JPL.NASA.GOV> lwall@jpl-devvax.JPL.NASA.GOV (Larry Wall) writes: |In article <97130172@bfmny0.BFM.COM> tneff@bfmny0.BFM.COM (Tom Neff) writes: |: In article <1991Jan30.005821.13013@convex.com> tchrist@convex.COM (Tom Christiansen) writes: |: >1) $foo = `cat $file`; |: > |: >2) open (FILE, $file); |: > undef $/; |: > $foo = <FILE>; |: > close(FILE); |: |: 3) open (FILE, $file); $foo = ''; |: while (!eof(FILE)) { read(FILE, $x, 32767); $foo .= $x;}; |: |: You can use bigger numbers than 32767... | |You sure can, including the size of the file: | |4) open(FILE, $file); read(FILE, $foo, -s $file); | |or better in two ways, | |5) open(FILE, $file); sysread(FILE, $foo, -s FILE); | |I sincerely doubt that any other construct is going to beat that, |timewise. ... However, 4) and 5) don't work on pipes, named or unnamed. 1) works on named pipes but not unnamed. Only 2) and 3) work everywhere. -- Annex Canada now! We need the room, \) Tom Neff and who's going to stop us. (\ tneff@bfmny0.BFM.COM
tneff@bfmny0.BFM.COM (Tom Neff) (01/30/91)
In article <1991Jan30.005821.13013@convex.com> tchrist@convex.COM (Tom Christiansen) writes: >1) $foo = `cat $file`; > >2) open (FILE, $file); > undef $/; > $foo = <FILE>; > close(FILE); 3) open (FILE, $file); $foo = ''; while (!eof(FILE)) { read(FILE, $x, 32767); $foo .= $x;}; You can use bigger numbers than 32767...
lwall@jpl-devvax.JPL.NASA.GOV (Larry Wall) (01/31/91)
In article <97130172@bfmny0.BFM.COM> tneff@bfmny0.BFM.COM (Tom Neff) writes: : In article <1991Jan30.005821.13013@convex.com> tchrist@convex.COM (Tom Christiansen) writes: : >1) $foo = `cat $file`; : > : >2) open (FILE, $file); : > undef $/; : > $foo = <FILE>; : > close(FILE); : : 3) open (FILE, $file); $foo = ''; : while (!eof(FILE)) { read(FILE, $x, 32767); $foo .= $x;}; : : You can use bigger numbers than 32767... You sure can, including the size of the file: 4) open(FILE, $file); read(FILE, $foo, -s $file); or better in two ways, 5) open(FILE, $file); sysread(FILE, $foo, -s FILE); I sincerely doubt that any other construct is going to beat that, timewise. (Solutions using syscall(SYS_mmap...) not allowed. :-) Larry
rbj@uunet.UU.NET (Root Boy Jim) (01/31/91)
In article <97130173@bfmny0.BFM.COM> tneff@bfmny0.BFM.COM (Tom Neff) writes: >In article <11235@jpl-devvax.JPL.NASA.GOV> lwall@jpl-devvax.JPL.NASA.GOV (Larry Wall) writes: >|In article <97130172@bfmny0.BFM.COM> tneff@bfmny0.BFM.COM (Tom Neff) writes: >|: In article <1991Jan30.005821.13013@convex.com> tchrist@convex.COM (Tom Christiansen) writes: The question is how to slurp a string into a file. Er, vice versa. LW>|You sure can, including the size of the file: >| >|4) open(FILE, $file); read(FILE, $foo, -s $file); >| >|or better in two ways, >| >|5) open(FILE, $file); sysread(FILE, $foo, -s FILE); >| >|I sincerely doubt that any other construct is going to beat that, >|timewise. ... > TN>However, 4) and 5) don't work on pipes, named or unnamed. 1) works on >named pipes but not unnamed. Only 2) and 3) work everywhere. OK, so Larry was being too literal. Suppose we replace -s FILE with $MAXINT? Sysread will get whatever it can. However, this begs the question, how does $foo get allocated? By sending the third arg to malloc? In that case, perhaps one should set a limit on how much they slurp and die/complain on therwise. This would also be a good time to ask about args to ioctls. I have been precreating them out of paranoia: $sgtty = 'ioekfl'; ioctl(FH,$TIOCGETP,$sgtty); Is this necessary? In either case, why? (6) $"=''; @str = <FILE>; $str = "@str"; (7) @str = <FILE>; $str = join(//,"@str"); Horshoe mode .signature: perl -e 'print "Just another $0 hacker,"' -- Root Boy Jim Cottrell <rbj@uunet.uu.net> Close the gap of the dark year in between
lwall@jpl-devvax.JPL.NASA.GOV (Larry Wall) (01/31/91)
In article <120644@uunet.UU.NET> rbj@uunet.UU.NET (Root Boy Jim) writes:
: OK, so Larry was being too literal.
I'm like that sometimes. It started at an early age, when my mom told me
to hop in the bathtub.
: Suppose we replace -s FILE with $MAXINT?
: Sysread will get whatever it can. However, this begs the question,
: how does $foo get allocated? By sending the third arg to malloc?
Noises of affirmation.
: In that case, perhaps one should set a limit on how much they slurp
: and die/complain on therwise.
Perhaps one should, as long as they aren't Perl. Any limit Perl might set
would be too small for some people and too large for others.
: This would also be a good time to ask about args to ioctls. I have
: been precreating them out of paranoia:
: $sgtty = 'ioekfl'; ioctl(FH,$TIOCGETP,$sgtty);
: Is this necessary? In either case, why?
Never necessary on a BSD-derived system, and seldom necessary on other systems.
$TIOCGETP encodes the required length of $sgtty on BSD systems, so it
preallocates that much for you. On non-BSD systems it preallocates 256
bytes, so you only have to preallocate it if you're expecting more info
back than that. It also adds a sentinel byte at the end so it can die
gracefully when you blow it.
fcntl works the same way.
syscall does nothing for you by way of pre-extension.
: (6) $"=''; @str = <FILE>; $str = "@str";
: (7) @str = <FILE>; $str = join(//,"@str");
:
: Horshoe mode .signature: perl -e 'print "Just another $0 hacker,"'
That may not do what you expect.
You might be interested to know that 4.0 will allow you to assign to $0 to
modify what ps sees. (Suggested by Tom Christiansen.)
die "shucks" if $] < 4.000;
$0 = 'Just another Perl hacker,';
print grep(s/.*Just/Not just/, `ps`);
Larrytchrist@convex.COM (Tom Christiansen) (01/31/91)
From the keyboard of rbj@uunet.UU.NET (Root Boy Jim):
:This would also be a good time to ask about args to ioctls. I have
:been precreating them out of paranoia:
:$sgtty = 'ioekfl'; ioctl(FH,$TIOCGETP,$sgtty);
:Is this necessary? In either case, why?
I believe ioctl() works like syscall(), in that you need to pre-extend the
string so the kernel has a place to write what it needs. This is
different from sysread, who knows how much it'll need.
BTW, I like to do stuff like this:
$sgttyb_t = 'c4 s';
sub sgttyb {
wantarray ? unpack($sgttyb_t, @_) : pack($sgttyb_t, @_);
}
You could then say to init:
$sgttyb = &sgttyb();
But don't forget the parens. This time they matter. You don't
want the previous guy's old @_.
:(6) $"=''; @str = <FILE>; $str = "@str";
It seems a tad wasteful to my efficiency-maniacal self to allocate all
those array members just to cat them together. At least undef it or let
it go out of local scope.
:(7) @str = <FILE>; $str = join(//,"@str");
If you didn't set $" to null, you'll get the "wrong" answer there,
i.e. extra spaces by default. Skip the quotes if you want.
:Horshoe mode .signature: perl -e 'print "Just another $0 hacker,"'
"Just another /tmp/perl-e029116 hacker,"? Well, it IS different.
--tom
--
"Hey, did you hear Stallman has replaced /vmunix with /vmunix.el? Now
he can finally have the whole O/S built-in to his editor like he
always wanted!" --me (Tom Christiansen <tchrist@convex.com>)lwall@jpl-devvax.JPL.NASA.GOV (Larry Wall) (01/31/91)
In article <1991Jan31.012957.25993@convex.com> tchrist@convex.COM (Tom Christiansen) writes:
: From the keyboard of rbj@uunet.UU.NET (Root Boy Jim):
: :(7) @str = <FILE>; $str = join(//,"@str");
:
: If you didn't set $" to null, you'll get the "wrong" answer there,
: i.e. extra spaces by default. Skip the quotes if you want.
If you skip the quotes you get the wrong answer too. The value of // is
likely to be 1.
Larryrbj@uunet.UU.NET (Root Boy Jim) (01/31/91)
In article <11242@jpl-devvax.JPL.NASA.GOV> lwall@jpl-devvax.JPL.NASA.GOV (Larry Wall) writes: >In article <120644@uunet.UU.NET> rbj@uunet.UU.NET (Root Boy Jim) writes: >: In that case, perhaps one should set a limit on how much they slurp >: and die/complain on therwise. > >Perhaps one should, as long as they aren't Perl. Any limit Perl might set >would be too small for some people and too large for others. Good point. I meant that if one is really going to slurp a file, one should set one's own limit, or iterate, etc. >: Horshoe mode .signature: perl -e 'print "Just another $0 hacker,"' > >That may not do what you expect. It's a joke. Close only counts in horseshoes, and my solution is only close. Here's another one. Lottery mode .signature: print pack('c',32+int rand 95) while ++$z % 26 Altho I think it'll take a random number generator which produces more random bits and an infinite number of camels to print the magic slogan. However, I did manage to produce some famous initials towards the end of the sixth iteration, along with my own, at the very end. -- Root Boy Jim Cottrell <rbj@uunet.uu.net> Close the gap of the dark year in between