| |
| |
| |
|
Page: 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35
Comments:
<0> Zaphar_work, what is a compiled regex?
<1> perldoc perlreref
<2> perlreref - Perl regular expressions quick reference. To access this perldoc please type, at a command line, 'perldoc perlreref'. You may also find it at http://perldoc.perl.org/perlreref.html
<0> something like: "foreach my $filename (@included) { # some regex magic here }"
<0> ?
<1> yes
<1> something like
<3> perldoc -f qr//
<2> qr. To access this perldoc please type, at a command line, 'perldoc -f qr'. You may also find it at http://perldoc.perl.org/functions/qr.html
<3> (qr - quote regex)
<0> hm, I don't know regex well enuff for this... just simple matching
<1> gr00ber: perfect time to learn :-)
<0> or does the regex engine know about wildcards?
<1> gr00ber: all you could wish to know are in the above links to documentation
<0> Zaphar_work, I guess :-) But my boss is almost sitting on my lap waiting for me to be done. Time to g00gle for shortcuts :-)
<0> Zaphar_work, do I need a double loop? I.e. inside the loop above, have another loop over @excluded that matches against $filename using some regex?
<0> or can regex work on items in an array directly?
<1> my $regex = gr/debug.*\.txt/; print "found a debug file" if $filename =~ $regex
<1> gr00ber: yes a double loop
<0> lemme see... g is global and r is ?
<1> actually that should be qr
<1> mistyped
<1> it creates a precompiled regex which you could then store in an array
<1> push @regex, qr/regex here/;
<0> will this work? my @ignorelist = (qr/debug.*\.txt/, ...);
<1> yep
<0> thanks
<3> gr00ber: It can never be global. Regex modifiers (such as the global modifier g) go at the end. Never at the start
<1> what he said :-)
<0> oh
<0> let's see: "debug.*\.txt" .... what does the dot after debug mean?
<1> read up on the meta characters when you create your regex's cause the wildcards and such don't work like a dos wildcard
<1> the dot stands for any single character
<0> ok
<0> like slash :)
<1> the star modifies the dot to mean 0 or more of any character
<3> AKA quantifier. :)
<0> ah
<1> and the slash escapes the next dot to mean an actual dot
<1> now go read :-)
<3> s/slash/back slash/;
<0> yup :)
<4> i got it to work guys, thanks Ani-_, Zaphar_work, and dkr. i used a relative path this time
<4> using a predefined base variable
<1> you'll quickly get addicted to regex's if you do much of this
<0> and \/ means an actual slash, right?
<1> you got it
<3> That's one way to write it. :)
<1> this being perl of course there's more than one way to write it :-)
<0> yeah, regex seems very powerful, but I never got around to learn'em. Time to roll up the sleeves.
<5> m#/slashes/are/fun# too
<3> If you have a regex with a lot of slashes then it is easier to change the delimiter. For example: m#a/b/c/d# would match the string: a/b/c/d (# can be whatever you want. But to do that you need the m)
<1> yeah for long regexes with a lot of slashes that is very helpful
<1> Spetsnazgg_: glad you got it working
<0> Ani-_, nice!
<3> If it's a really long regex then x-modifier can be handy aswell. (If you want to know what it does then look at the docs)
<6> don't use #, that can be a valid regex char, :)
<7> #perl topic for the evening "Regexen for fun and profit!"
<6> (if you use /x liek Ani-_ just mentioned ;)
<8> regexes are also useful in writing the actual code
<6> it's raining yet again and my convertible top is leaking, :(
<3> dkr, but how ofthen do you use /x? :)
<6> Ani-_: I just wrote one actually, and I even used # to help me figure out what part of the regex wasw rong, :)
<8> Sometimes you can prevent lot of typing with some good regexp
<6> Ani-_: it was for parsing an xml formatted logfile
<0> ok, so in this double loop I would like to move on in the outer loop. Is there an "next outer" thingy ? :-)
<3> dkr: I use m## more ofthen then m//x; (Mainly because I can't think of a better delimiter) :/
<6> yeah, use a label
<0> ah, like OUTER: for ...
<6> nod
<0> then next OUTER; ?
<3> gr00ber: perldoc -f redo
<2> redo. To access this perldoc please type, at a command line, 'perldoc -f redo'. You may also find it at http://perldoc.perl.org/functions/redo.html
<0> niiice
<0> hm
<3> redo has an example of it usage. It also applied to last and next
<0> right... worked nicely. Thanks.
<6> Ani-_: I like ~ as a delimiter since it doesn't mean anything to perl and rarely appears in strings. so my code has lots of q~ and qq~ in it, :)
<0> next task: get the timestamp of each file and put them in a database or something... maybe a text file. So the next time the script runs, compare against it's previous timestamp and skip building if unchanged.
<0> design suggestions before I hack on ? :)
<6> someone brought up the point the other day that if you use m{} then yuo can have ensted { and } inside
<3> dkr: $x =~ m~ ~g; looks a bit strange... :/
<6> Ani-_: yeah, I just use it for quotes, not matches
<6> oh yeah, ~ does mean something to perl, forgot, :)
<1> it looks like a lot of characters swimming around in the code :-)
<3> What character doesn't? :)
<9> you can even use ';' as a delimeter
<10> http://www.flickr.com/photos/clearhop/165029704/in/pool-wm2006/
<10> what's with the Keep me English
<11> "rbellamy" at 67.169.146.106 pasted "SVK mirror to mirror copy error" (6 lines, 407B) at http://sial.org/pbot/18036
<3> eval: "mxm" =~ mmxm ? "Match" : "No match";
<2> Ani-_: No match
<1> gr00ber: why are you making a build system anyway? why not just use one of the freely available ones?
<3> deparse: "mxm" =~ mmxm ? "Match" : "No match";
<2> (('mxm' =~ /mmxm/) ? 'Match' : 'No match');
<3> Bah.
<3> deparse: "x" =~ msxs ? "Match" : "No match";
<2> (('x' =~ /msxs/) ? 'Match' : 'No match');
<0> Zaphar_work, it's highly specialized
<3> eval: "x" =~ m mxm ? "Match" : "No match";
<2> Ani-_: Match
<0> I usually use GNU make
<12> I have a ~/.cpan/CPAN/MyConfig.pm and a /etc/perl/CPAN/Config.pm. They're identical and work fine on another system. But when I try to install a module, I get "we have to reconfigure CPAN.pm due to the following uninitialized parameters..." followed by a list of parameters covered in my config, followed by a MyConfig.pm initialized line, then am I ready for manual configuration. How can I get cpan to use my config?
<0> maybe a text file with "filename, timestamp since epoch" pairs on each line will do?
<9> eval: quote for you
<2> pravus:
<1> gr00ber: how many files are you talking?
<0> Zaphar_work, about 2000
<13> How can I flush STDIN? I'm trying to create a script that accepts piped input (eg. echo "test" | myscript.pl ) and then asks the user for further input. The problem I have is that the piped input is being used when I want the further user input. I'm currently using Term::ReadKey to get the input, is there perhaps a better way of getting the user input on the console?
<1> hrmmm not too bad but might be easier to just store in a sqlite3 db
<3> gr00ber: you do know that one can change the timestamp of a file?
<14> You can't flush STDIN
<14> The thing that sends data to you, has to flush its STDOUT
<0> Ani-_, yes?
<3> gr00ber: yes. man touch
<1> gr00ber: you might want to use a hash of the file instead to check if it changed
<0> oh, you mean not the modified date but something else
<0> Zaphar_work, yeah, been thinking about that
<3> I do mean the modified date.
<13> LeoNerd: OK. Thanks. Is there an alternative way of doing what I want to do (accept piped input and then ask for further user input)?
<0> Ani-_, well... that's the point: only build files who's modification date has changed
<15> If a user enters a relative filesystem path, can someone reccomend how I can change this to complete path?
<3> man touch: -m Change the modification time of file
<0> Ani-_, i don't get your drift
<14> Sure.. You can open /dev/tty, and if your process is connected to a terminal, you'll get that.
<3> gr00ber: and my little point is that one can change the file and modify it's timestamp.
<3> To make it look to your script that it didn't change
<0> Ani-_, ah, but I don't care about that, since I control all the files
<1> he means that a timestamp isn't a reliable indicator
<0> they are only changed by my editor
<0> Zaphar_work, fair
<0> but not an issue
<0> it would be an issue in a production quality system, but this is a specialized, short-lived script for prototype system
<1> ahh
<3> I never said it was an issue. I only asked if you knew that you could.
<0> Ani-_, ok :)
<1> Ani-_: is looking out for ya :-)
<1> he's got your back
<1> if it's just a quick and dirty script then a text file is probably just fine
<1> if it has to be used more robustly then sqlite would be my preference for storing the file and timestamps
<0> Zaphar_work, ok... but I've used sqlite in C++ and it worked out nicely
<0> are the Perl bindings easy to use?
<0> and install?
<1> install DBD::sqlite in cpan and that's all you have to do
<1> DBI is pretty simple to use
<1> perldoc DBI
<0> that way I could store even the regexp's... you see: it's not THAT shortlived - we're making prototypes of a huge system with multiple languages, etc... so it will live at least 6 months or so. So it should **** to much :)
<0> DBI == database interface? :)
Return to
#perl
or
Go to some related
logs:
xorg-server app autostart
nukepuppy
automated id3-tagging linux -windows
merge strings python
#oe
programs i digitalbit
#css
#xine
Player for encrypted wmv
djpirate
|
|