Changeset 38740 for main/trunk/greenstone2

Timestamp:

2024-02-13T17:10:07+13:00 (4 months ago)

Author:

davidb

Message:

Handy to have a subroutine that can read in a text file (csv,tsv) raw as a string if we know it is definitely in UTF8

File:

• main/trunk/greenstone2/perllib/util.pm (modified) (4 diffs)

main/trunk/greenstone2/perllib/util.pm

@@ -27 +27 @@
 
 use strict;
+no strict 'subs';
 no strict 'refs'; # make an exception so we can use variables as filehandles
+
+
 use FileUtils;
 
@@ -35 +38 @@
 use File::Copy;
 use File::Basename;
+
+use gsprintf;
+use multiread;
+
 # Config for getting the perlpath in the recommended way, though it uses paths that are
 # hard-coded into the Config file that's generated upon configuring and compiling perl.
@@ -40 +47 @@
 # but if launched with plain "perl" (no full-path), that will be just what it returns.
 use Config;
+
 # New module for file related utility functions - intended as a
 # placeholder for an extension that allows a variety of different
@@ -2010 +2018 @@
 }
 
+sub read_utf8_textfile {
+    my ($input_filename) = @_;
+
+    my $text_file_content_raw = undef;
+    
+    if (!open(TEXT_FILE,"<$input_filename")) {
+        &gsprintf::gsprintf(STDERR, "util::read_utf8_file() {ReadTextFile.could_not_open_for_reading} ($!)\n", $input_filename);
+    die "\n";
+    }
+    
+    my $file_reader = new multiread();
+    $file_reader->set_handle('util::TEXT_FILE');
+    $file_reader->read_file(\$text_file_content_raw);
+
+    close(TEXT_FILE);
+
+    my $text_file_content = decode("utf8",$text_file_content_raw);
+
+    return $text_file_content;
+}
+
+    
 1;