Some regex magic

monkie has asked for the wisdom of the Perl Monks concerning the following question:

Hi Monks, I'm expecting to pull from a win32 reg file of the uninstall keys, the Name, Version and Manufacturer of the software installed.

My perl:

#!/usr/bin/perl
use strict;

open (DATAFILE, "<", 'test.txt') or die "Couldn't open file for readin
+g: $!\n";
my   $temp;
$/ = ""; ## Paragraph mode

my $display_name;
my $version;

while( <DATAFILE> ) {

    m[
        (?= .*? \"DisplayName\"=\"([a-zA-Z0-9. -]*))
        (?= .*? \"DisplayVersion\"=\"([a-zA-Z0-9 .-]*))
        (?= .*? \"Publisher\"=\"([a-zA-Z0-9 .-]*))
    ]xsm;

    print "Name:$1\nVersion:$2\nPublisher:\n\n";

}

close(DATAFILE);
[download]

My Output:

Name:
Version:
Publisher:

Name:
Version:
Publisher:

Name:
Version:
Publisher:

Name:Adobe SVG Viewer 3.0
Version: 3.0
Publisher:

Name:
Version:
Publisher:

Name:
Version:
Publisher:

Name:Example Display Name
Version:v7.5.2
Publisher:

Name:
Version:
Publisher:
[download]

The data file:

Windows Registry Editor Version 5.00

[HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Windows\CurrentVersion\Uninstal
+l]
"DisplayName"=""
"UninstallString"="C:\\PROGRA~1\\CISCOS~1\\VPNCLI~1\\Profiles\\PROGRA~
+1\\CISCOS~1\\VPNCLI~1\\profiles\\UNWISE.EXE C:\\PROGRA~1\\CISCOS~1\\V
+PNCLI~1\\Profiles\\PROGRA~1\\CISCOS~1\\VPNCLI~1\\profiles\\INSTALL.LO
+G"

[HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Windows\CurrentVersion\Uninstal
+l\AddressBook]
@=""

[HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Windows\CurrentVersion\Uninstal
+l\Adobe SVG Viewer]
"DisplayName"="Adobe SVG Viewer 3.0"
"DisplayVersion"=" 3.0"
"DisplayIcon"="C:\\Program Files\\Common Files\\Adobe\\SVG Viewer 3.0\
+\Uninstall\\SetupRsrc.dll,-200"
"VersionMajor"=dword:00000003
"VersionMinor"=dword:00000000
"InstallLocation"="C:\\WINDOWS\\system32\\Adobe\\SVG Viewer 3.0"
"UninstallString"="C:\\Program Files\\Common Files\\Adobe\\SVG Viewer 
+3.0\\Uninstall\\Winstall.exe -u -fC:\\Program Files\\Common Files\\Ad
+obe\\SVG Viewer 3.0\\Uninstall\\Install.log"
"UninstallDir"="C:\\Program Files\\Common Files\\Adobe\\SVG Viewer 3.0
+\\Uninstall"
"Publisher"="Adobe Systems, Inc."
"RegCompany"="Adobe Systems Inc"
"URLInfoAbout"="http://www.adobe.com/"
"URLUpdateInfo"="http://www.adobe.com/svg/main.html"

[HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Windows\CurrentVersion\Uninstal
+l\Artemis 7.2.1 ODBC]
"NoRemove"=dword:00000001

[HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Windows\CurrentVersion\Uninstal
+l\Standalone]
"UninstallString"="C:\\WINDOWS\\IsUninst.exe -f\"C:\\Program Filesone.
+isu\""
"DisplayName"="This is-not SHOWING"

[HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Windows\CurrentVersion\Uninstal
+l\exampler]
"DisplayName"="Example Display Name"
"UninstallString"="c:\\applications\\uninst.exe"
"DisplayVersion"="v7.5.2"
"URLInfoAbout"="http://something.htm"
"Publisher"="NOT caught"

[HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Windows\CurrentVersion\Uninstal
+l\Branding]
"QuietUninstallString"="Rundll32 IedkCS32.dll,BrandCleanInstallStubs"
"RequiresIESysFile"="100.0"
[download]

As you can see from my results, I'm not catching everything, bearing in mind that the order of the 3 fields which I'm looking for may differ for each entry. Please can someone show me the folly of my ways?

Comment on Some regex magic Select or Download Code

Replies are listed 'Best First'.
Re: Some regex magic by przemo (Scribe) on May 08, 2009 at 13:35 UTC
I'd choose another way and use one of myriad `Config::*` CPAN modules. For example, with Config::Tiny you can do something similar to (after removing everything before the first section in your data file): `use warnings; use strict; use Config::Tiny; my $cfg = Config::Tiny->read('test.txt') or die "Couldn't read file: $ +Config::Tiny::errstr"; for my $sect (values %$cfg) { my ($name, $version, $publisher) = map { $sect->{$_} } qw("Display +Name" "DisplayVersion" "Publisher"); print "Name: $name\nVersion: $version\nPublisher: $publisher\n"; + }` [download]	[reply] [d/l] [select]
Re: Some regex magic by johngg (Canon) on May 08, 2009 at 13:14 UTC
I would use three separate regular expressions rather than trying to use just one more complex one using look-aheads. Also, double-quotes are not metacharacters so need not be escaped. use strict; use warnings; open my $regFH, q{<}, \ <<'EOD' or die qq{open: << HEREDOC: $!\n}; Windows Registry Editor Version 5.00 [HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Windows\CurrentVersion\Uninstal +l] "DisplayName"="" "UninstallString"="C:\\PROGRA~1\\CISCOS~1\\VPNCLI~1\\Profiles\\PROGRA~ +1\\CISCOS~1\\VPNCLI~1\\profiles\\UNWISE.EXE C:\\PROGRA~1\\CISCOS~1\\V +PNCLI~1\\Profiles\\PROGRA~1\\CISCOS~1\\VPNCLI~1\\profiles\\INSTALL.LO +G" [HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Windows\CurrentVersion\Uninstal +l\AddressBook] @="" [HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Windows\CurrentVersion\Uninstal +l\Adobe SVG Viewer] "DisplayName"="Adobe SVG Viewer 3.0" "DisplayVersion"=" 3.0" "DisplayIcon"="C:\\Program Files\\Common Files\\Adobe\\SVG Viewer 3.0\ +\Uninstall\\SetupRsrc.dll,-200" "VersionMajor"=dword:00000003 "VersionMinor"=dword:00000000 "InstallLocation"="C:\\WINDOWS\\system32\\Adobe\\SVG Viewer 3.0" "UninstallString"="C:\\Program Files\\Common Files\\Adobe\\SVG Viewer +3.0\\Uninstall\\Winstall.exe -u -fC:\\Program Files\\Common Files\\Ad +obe\\SVG Viewer 3.0\\Uninstall\\Install.log" "UninstallDir"="C:\\Program Files\\Common Files\\Adobe\\SVG Viewer 3.0 +\\Uninstall" "Publisher"="Adobe Systems, Inc." "RegCompany"="Adobe Systems Inc" "URLInfoAbout"="http://www.adobe.com/" "URLUpdateInfo"="http://www.adobe.com/svg/main.html" [HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Windows\CurrentVersion\Uninstal +l\Artemis 7.2.1 ODBC] "NoRemove"=dword:00000001 [HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Windows\CurrentVersion\Uninstal +l\Standalone] "UninstallString"="C:\\WINDOWS\\IsUninst.exe -f\"C:\\Program Filesone. +isu\"" "DisplayName"="This is-not SHOWING" [HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Windows\CurrentVersion\Uninstal +l\exampler] "DisplayName"="Example Display Name" "UninstallString"="c:\\applications\\uninst.exe" "DisplayVersion"="v7.5.2" "URLInfoAbout"="http://something.htm" "Publisher"="NOT caught" [HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Windows\CurrentVersion\Uninstal +l\Branding] "QuietUninstallString"="Rundll32 IedkCS32.dll,BrandCleanInstallStubs" "RequiresIESysFile"="100.0" EOD { local $/ = q{}; while( <$regFH> ) { next unless m{^\[HKEY}; my $displayName = m{"DisplayName"="([^"]+)} ? do { my $s = $1; $s =~ s{^\s+}{}; $s =~ s{\s+$}{}; $s; } : q{Not found!}; my $displayVersion = m{"DisplayVersion"="([^"]+)} ? do { my $s = $1; $s =~ s{^\s+}{}; $s =~ s{\s+$}{}; $s; } : q{Not found!}; my $publisher = m{"Publisher"="([^"]+)} ? do { my $s = $1; $s =~ s{^\s+}{}; $s =~ s{\s+$}{}; $s; } : q{Not found!}; print qq{Name: $displayName\n}, qq{Version: $displayVersion\n}, qq{Publisher: $publisher\n\n}; } } close $regFH or die qq{close: << HEREDOC: $!\n}; [download] The output. `Name: Not found! Version: Not found! Publisher: Not found! Name: Not found! Version: Not found! Publisher: Not found! Name: Adobe SVG Viewer 3.0 Version: 3.0 Publisher: Adobe Systems, Inc. Name: Not found! Version: Not found! Publisher: Not found! Name: This is-not SHOWING Version: Not found! Publisher: Not found! Name: Example Display Name Version: v7.5.2 Publisher: NOT caught Name: Not found! Version: Not found! Publisher: Not found!` [download] I hope this is helpful. Cheers, JohnGG Update: Fixed typo, changed mot to not	[reply] [d/l] [select]
Re: Some regex magic by philipbailey (Curate) on May 08, 2009 at 14:04 UTC
Another suggestion, orthogonal to those given so far, is to obtain your data directly from the Windows registry. There are various modules to do this, although I have had most luck with Win32::TieRegistry.	[reply]
Re: Some regex magic by johngg (Canon) on May 09, 2009 at 22:00 UTC
Although I advocated using three separate regular expressions in my first reply I wondered whether the OP's use of three look-aheads with captures in a single regular expression could work. The look-aheads had to cope with the fact that none, all or any combination in between of the three relevant items might exist. Trying various combinations of non-greedy and zero-or-one quantifiers inside each look-ahead resulted in a lot of head scratching but no useful output. Finally, it occurred to me to try putting the zero-or-one quantifiers on the look-aheads themselves, thus making each look-ahead optional. That did the trick. Read more... (6 kB) The output. Read more... (1003 Bytes) I hope this is of interest. Cheers, JohnGG Update: Fixed typo, `s{scatching}{scratching}`	[reply] [d/l] [select]
Re^2: Some regex magic by monkie (Novice) on May 11, 2009 at 08:46 UTC
Thank you so much - I decided to go with your first suggestion which is working very well - doing exactly what it's meant to do!	[reply]
Re: Some regex magic by bichonfrise74 (Vicar) on May 08, 2009 at 22:53 UTC
Try this... #!/usr/bin/perl use strict; $/ = "\n\n"; while( <DATA> ) { my ($name) = $_ =~ /DisplayName\"\=\"(.)\"/; my ($version) = $_ =~ /VersionMajor\"\=(.)/; my ($publisher) = $_ =~ /Publisher\"\=\"(.*)\"/; $name = defined($name) ? $name : "n/a"; $version = defined($version) ? $version : "n/a"; $publisher = defined($publisher) ? $publisher : "n/a"; print "\nName: $name\nVersion: $version\nPublisher: $publisher\n"; } __DATA__ [HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Windows\CurrentVersion\Uninstal +l] "DisplayName"="" "UninstallString"="C:\\PROGRA~1\\CISCOS~1\\VPNCLI~1\\Profiles\\PROGRA~ +1\\CISCOS~1\\VPNCLI~1\\profiles\\UNWISE.EXE C:\\PROGRA~1\\CISCOS~1\\V +PNCLI~1\\Profiles\\PROGRA~1\\CISCOS~1\\VPNCLI~1\\profiles\\INSTALL.LO +G" [HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Windows\CurrentVersion\Uninstal +l\Adobe SVG Viewer] "DisplayName"="Adobe SVG Viewer 3.0" "DisplayVersion"=" 3.0" "DisplayIcon"="C:\\Program Files\\Common Files\\Adobe\\SVG Viewer 3.0\ +\Uninstall\\SetupRsrc.dll,-200" "VersionMajor"=dword:00000003 "VersionMinor"=dword:00000000 "InstallLocation"="C:\\WINDOWS\\system32\\Adobe\\SVG Viewer 3.0" "UninstallString"="C:\\Program Files\\Common Files\\Adobe\\SVG Viewer3 +.0\\Uninstall\\Winstall.exe -u -fC:\\Program Files\\Common Files\\Ado +be\\SVG Viewer 3.0\\Uninstall\\Install.log" "UninstallDir"="C:\\Program Files\\Common Files\\Adobe\\SVG Viewer 3.0 +\\Uninstall" "Publisher"="Adobe Systems, Inc." "RegCompany"="Adobe Systems Inc" "URLInfoAbout"="http://www.adobe.com/" "URLUpdateInfo"="http://www.adobe.com/svg/main.html" [download]	[reply] [d/l]