It is meant to take translated strings in a '.po' file and re-insert them into the source code. After that, a new '.pot' file would have to be regenerated from the code (the script doesn't do that).
I want to make this script run successfully against the Zikula core, so that I can re-insert all the reviewed strings (from the en_GB language on the translation portal) into the Zikula core, before the 1.2 release. After that, it can be used by the international communities and by 3rd-party devs to produce localised versions of Zikula and 3rd-party modules.
Problems:
1) First of all, it is failing to escape single and double quotation marks. This is causing syntax errors. For instance, when I try to run 'install.php' to install a site with the processed Zikula core, I'm getting:
Code
Parse error: syntax error, unexpected T_STRING in /var/www/zk120rc4/includes/pnAPI.php on line 1495
The gettext string in the file is:
Code
It is not escaped (backslashed), so it's a syntax error.
2) After the script has run, I'm finding that not all the strings have been replaced. For example, after processing a copy of the core, the file 'includes/LogUtil.class.php' at line 438 contains the string:
Code
"Unknown log destination [%s] ... "
But the reviewed string that *should* have replaced it is:
The script is not processing all strings properly.
3) This is not a bug, but it's something it should do: the script does not handle plural forms. To be a properly working tool for the community, it should be able to process plural-form gettext strings.
Any help troubleshooting these problems would be gratefully accepted.
Please reply here or contact me by mail (I might be willing to pay for the solution if you can make it work *perfectly*).
Thanks for any help or suggestions.
Here's the full code of the script.
Code
/**
* This is a little script to take a Zikula '.po' file
* and insert the language strings back into the
* source code files, so that the translated strings
* in the '.po' file become the gettext ID strings
* within the code files.
* It's a tool for language reviews and obtaining
* Zikula perfectly localised in the language of your
* Choice.
*
* You should run this tool on a local system. It uses
* too much memory to run it in a remote Web space.
* Put this script and a '.po' file in the root directory
* of a Zikula distribution.
* Then run this command from the command line:
*
* php replacepo.php
*
* It will examine every file in the tree structure of the
* distribution and perform the string replacement operations
* on each file. You'll need plenty of memory allocated
* for scripts (set to 128M below). Remember that you'll
* need write permissions for all files (on Linux, you
* might need to be logged in as root or, better still,
* prefix the above command with 'sudo' to simply run
* the script as root user.
*
* If you want to process a '.po' file other than 'zikula.po'
* then just edit the file name on line 46 below.
*
* Limitations: The script does not handle plural strings.
* You'll still have to do those by hand. This is a feature
* that should be added.
*
* Contact: David, commerce@traduction.biz
* License: GNU/GPL - http://www.gnu.org/copyleft/gpl.html
*/
// Set the script's memory requirement and then include the Zikula base API.
ini_set('memory_limit', '128M');
// Load the file we're going to process.
$poFile = file_get_contents ('zikula.po');
// Break the file down into an array of individual lines.
$poLines = explode ("\n", $poFile);
// Declare an array to hold the list of file names that contain a given language string.
$files = array();
// Declare an array of strings for the message IDs.
$msgIDs = array();
// Declare an array of strings for the message strings (the actual language strings).
$msgStrs = array();
// Declare a string to hold the counter keeping track of our position within the string we'll be parsing.
$count = 0;
// Start a loop of examining each string containing one line in the input '.po' file.
foreach ($poLines as $poLine) {
$tMsgID = '';
// Trim any white space off the beginning and end of the string.
$poLine = trim($poLine);
// If the line is one of the comment lines preceding each language string 'msgid' then
// that line gets broken into two fields: the comment tag and a file name.
// Each file name is a file contains that particular 'msgid'.
if (strpos ($poLine, '#:') === 0) {
$fields = explode (' ', $poLine);
// Just drop the comment tag. We don't need it.
unset ($fields[0]);
// Analyse the string containing the file name and remove the line number.
// We just want the file name (including the path).
// Then we add the file name to an array of file names that contain the 'msgid'.
foreach ($fields as $field) {
$fields2 = explode (':', $field);
$file = $fields2[0];
$files[$file] = $file;
}
// Else if the line delimits the start of a plural-case string ID or
// one of the plural cases of the string, then just ignore it.
// This script does not yet handle plural cases.
} elseif (strpos ($poLine, 'msgid_plural') === 0 || strpos($poLine, 'msgstr[') === 0) {
continue;
// Else if the line begins with an 'msgid' then take the double-quote-bounded
// string that follows immediately after, break it down into an array of characters.
} elseif (strpos ($poLine, 'msgid') === 0) {
$msg = substr (trim($poLine), 6);
$chars = str_split ($msg);
unset ($chars[count($chars)-1]);
unset ($chars[0]);
$msg = addslashes(implode('', $chars));
$_msgID1 = '"' . $msg . '"';
$_msgID2 = "'" . $msg . "'";
$msgID1 = '__(' . $_msgID1 . ')';
$msgID2 = '__(' . $_msgID2 . ')';
$msgID3 = 'gt text=' . $_msgID1;
$msgID4 = 'gt text=' . $_msgID2;
// Else if the line begins with an 'msgstr' then it's the string associated with the 'msgid'.
} elseif (strpos ($poLine, 'msgstr') === 0) {
if (!$count++) {
continue;
}
$msg = substr (trim($poLine), 7);
$chars = str_split ($msg);
unset ($chars[count($chars)-1]);
unset ($chars[0]);
$msg = addslashes(implode('', $chars));
$_msgString1 = '"' . $msg . '"';
$_msgString2 = "'" . $msg . "'";
$msgString1 = '__(' . $_msgString1 . ')';
$msgString2 = '__(' . $_msgString2 . ')';
$msgString3 = 'gt text=' . $_msgString1;
$msgString4 = 'gt text=' . $_msgString2;
$msgIDs[] = $msgID1;
$msgStrs[] = $msgString1;
$msgIDs[] = $msgID2;
$msgStrs[] = $msgString2;
$msgIDs[] = $msgID3;
$msgStrs[] = $msgString3;
$msgIDs[] = $msgID4;
$msgStrs[] = $msgString4;
}
}
//foreach ($msgIDs as $k=>$v) {
//print "$v\n";
//print "$msgStrs[$k]\n\n";
//} exit();
foreach ($files as $file) {
print "Processing $file ...\n";
$fData = file_get_contents ($file);
$newFileData = str_replace ($msgIDs, $msgStrs, $fData);
file_put_contents ($file, $newFileData);
}
* This is a little script to take a Zikula '.po' file
* and insert the language strings back into the
* source code files, so that the translated strings
* in the '.po' file become the gettext ID strings
* within the code files.
* It's a tool for language reviews and obtaining
* Zikula perfectly localised in the language of your
* Choice.
*
* You should run this tool on a local system. It uses
* too much memory to run it in a remote Web space.
* Put this script and a '.po' file in the root directory
* of a Zikula distribution.
* Then run this command from the command line:
*
* php replacepo.php
*
* It will examine every file in the tree structure of the
* distribution and perform the string replacement operations
* on each file. You'll need plenty of memory allocated
* for scripts (set to 128M below). Remember that you'll
* need write permissions for all files (on Linux, you
* might need to be logged in as root or, better still,
* prefix the above command with 'sudo' to simply run
* the script as root user.
*
* If you want to process a '.po' file other than 'zikula.po'
* then just edit the file name on line 46 below.
*
* Limitations: The script does not handle plural strings.
* You'll still have to do those by hand. This is a feature
* that should be added.
*
* Contact: David, commerce@traduction.biz
* License: GNU/GPL - http://www.gnu.org/copyleft/gpl.html
*/
// Set the script's memory requirement and then include the Zikula base API.
ini_set('memory_limit', '128M');
// Load the file we're going to process.
$poFile = file_get_contents ('zikula.po');
// Break the file down into an array of individual lines.
$poLines = explode ("\n", $poFile);
// Declare an array to hold the list of file names that contain a given language string.
$files = array();
// Declare an array of strings for the message IDs.
$msgIDs = array();
// Declare an array of strings for the message strings (the actual language strings).
$msgStrs = array();
// Declare a string to hold the counter keeping track of our position within the string we'll be parsing.
$count = 0;
// Start a loop of examining each string containing one line in the input '.po' file.
foreach ($poLines as $poLine) {
$tMsgID = '';
// Trim any white space off the beginning and end of the string.
$poLine = trim($poLine);
// If the line is one of the comment lines preceding each language string 'msgid' then
// that line gets broken into two fields: the comment tag and a file name.
// Each file name is a file contains that particular 'msgid'.
if (strpos ($poLine, '#:') === 0) {
$fields = explode (' ', $poLine);
// Just drop the comment tag. We don't need it.
unset ($fields[0]);
// Analyse the string containing the file name and remove the line number.
// We just want the file name (including the path).
// Then we add the file name to an array of file names that contain the 'msgid'.
foreach ($fields as $field) {
$fields2 = explode (':', $field);
$file = $fields2[0];
$files[$file] = $file;
}
// Else if the line delimits the start of a plural-case string ID or
// one of the plural cases of the string, then just ignore it.
// This script does not yet handle plural cases.
} elseif (strpos ($poLine, 'msgid_plural') === 0 || strpos($poLine, 'msgstr[') === 0) {
continue;
// Else if the line begins with an 'msgid' then take the double-quote-bounded
// string that follows immediately after, break it down into an array of characters.
} elseif (strpos ($poLine, 'msgid') === 0) {
$msg = substr (trim($poLine), 6);
$chars = str_split ($msg);
unset ($chars[count($chars)-1]);
unset ($chars[0]);
$msg = addslashes(implode('', $chars));
$_msgID1 = '"' . $msg . '"';
$_msgID2 = "'" . $msg . "'";
$msgID1 = '__(' . $_msgID1 . ')';
$msgID2 = '__(' . $_msgID2 . ')';
$msgID3 = 'gt text=' . $_msgID1;
$msgID4 = 'gt text=' . $_msgID2;
// Else if the line begins with an 'msgstr' then it's the string associated with the 'msgid'.
} elseif (strpos ($poLine, 'msgstr') === 0) {
if (!$count++) {
continue;
}
$msg = substr (trim($poLine), 7);
$chars = str_split ($msg);
unset ($chars[count($chars)-1]);
unset ($chars[0]);
$msg = addslashes(implode('', $chars));
$_msgString1 = '"' . $msg . '"';
$_msgString2 = "'" . $msg . "'";
$msgString1 = '__(' . $_msgString1 . ')';
$msgString2 = '__(' . $_msgString2 . ')';
$msgString3 = 'gt text=' . $_msgString1;
$msgString4 = 'gt text=' . $_msgString2;
$msgIDs[] = $msgID1;
$msgStrs[] = $msgString1;
$msgIDs[] = $msgID2;
$msgStrs[] = $msgString2;
$msgIDs[] = $msgID3;
$msgStrs[] = $msgString3;
$msgIDs[] = $msgID4;
$msgStrs[] = $msgString4;
}
}
//foreach ($msgIDs as $k=>$v) {
//print "$v\n";
//print "$msgStrs[$k]\n\n";
//} exit();
foreach ($files as $file) {
print "Processing $file ...\n";
$fData = file_get_contents ($file);
$newFileData = str_replace ($msgIDs, $msgStrs, $fData);
file_put_contents ($file, $newFileData);
}
You'll find it's actually much faster that way, and you don't have to worry about the file size -- it will work on whatever file you give it, AND can than work without problems on the server side if needed.
), and have since been reviewing the generated distribution with all the reviewed lang strings incorporated. I need to do a final review of the result and deliver it to Drak within the next 24 hours or so.
