Re: [SMARTY-DEV] Newlines
| From: | Monte Ohrt | Date: | Thu, 04 Apr 2002 17:39:34 +0000 |
| Subject: | Re: [SMARTY-DEV] Newlines | ||
| References: | 1 | Groups: | php.smarty.dev |
| Request: | Send a blank email to php-smarty-dev+get-23@lists.php.net to get a copy of this message | ||
Ah yes, the problem extends from the fact that PHP eats carriage returns immediately after a php close tag. We cannot simply remove this \n in Smarty, it will cause incorrect formatting.
I have attached a revised version of trimwhitespace, it is actually more efficient since one regex now takes care of both leading spaces and empty lines. This is accomplished with a negative match look behind, which PCRE thankfully supports ;-)
I'm also trying to figure out a way to selectively disable trimwhitespace (besides unregistering the function.) For instance, if your template is an e-mail message, you don't want to trim it! My first thought was to look for {* trimwhitespace off *} in the template, but in a post filter that would already be removed. We could change Smarty so that Smarty comments are kept in the source, so {* foo *} would be come <?php /* foo */ ?> in the compiled template. Any thoughts on this?
Actually I don't think this list supports attachments, so I'll paste the source here:
<?php
/*
* Smarty plugin
* -------------------------------------------------------------
* File: postfilter.trimwhitespace.php * Type: postfilter * Name: trimwhitespace* Version: 1.1
* Date: April 4, 2002* Purpose: trim leading white space and blank lines from
* template source after it gets interpreted, cleaning * up code and saving bandwidth. Does not affect * <PRE></PRE> and <SCRIPT></SCRIPT> blocks.* Install: Drop into the plugin directory. * Author: Monte Ohrt <monte@ispi.net> * Changes: 1.1 fixed bug with carriage returns after php tag * ------------------------------------------------------------- */ function smarty_postfilter_trimwhitespace($source, &$smarty) {
// Pull out the script blocks
preg_match_all("!<script[^>]+>.*?</script>!is", $source, $match);
$_script_blocks = $match[0];
$source = preg_replace("!<script[^>]+>.*?</script>!is",
'@@@SMARTY:TRIM:SCRIPT@@@', $source);
// Pull out the pre blocks
preg_match_all("!<pre>.*?</pre>!is", $source, $match);
$_pre_blocks = $match[0];
$source = preg_replace("!<pre>.*?</pre>!is",
'@@@SMARTY:TRIM:PRE@@@', $source);
// remove all leading spaces, tabs and carriage returns NOT
// preceeded by a php close tag.
$source = preg_replace('/((?<!\?>)\n)[\s]+/m', '\1', $source);
// replace script blocks
foreach($_script_blocks as $curr_block) {
$source = preg_replace("!@@@SMARTY:TRIM:SCRIPT@@@!",$curr_block,$source,1);
}
// replace pre blocks
foreach($_pre_blocks as $curr_block) {
$source = preg_replace("!@@@SMARTY:TRIM:PRE@@@!",$curr_block,$source,1);
}
return $source;
}
?>
Robert Amos wrote:
-- Monte Ohrt <monte@ispi.net> Director of Technology, ispi Inc.From an earlier email that Jimmy sent to the list:$this->_parse_vars_props($tag_command); return "<?php echo $tag_command[0]; ?>"; <<<<<<<< I REMOVEDthe \n at the end}there is a newline placed after the variable, which isnt normally affected in displaying the page, but combined with the trimwhitespace postfilter it removes the space from the start of the next line. -bok On 4/4/02 12:16 AM, Monte Ohrt wrote the following:Hmm, trimwhitespace should only trim space out from the beginning and ends of lines, from from between something like "{$var} text" Is that not the case? Jason Sweat wrote:I did some further testing because I encountered the same problem. It has to do with postfilter.trimwhitespace.php No postfilter, no problem. Or as another temporary solution if you want to keep the postfilter, hard code a non-breaking space in between the var and the text (i.e. {$var} text). I am not sufficiently adept with regexp to solve the problem, but at least it is better identified. Jason ----- Original Message ----- From: "Monte Ohrt" <monte@ispi.net> To: "Jimmy Harlindong" <jimmy@harlindong.com> Cc: <smarty-dev@lists.php.net> Sent: Tuesday, April 02, 2002 9:11 AM Subject: Re: [SMARTY-DEV] NewlinesI cannot reproduce this problem. Here is my simple test: index.tpl --------- begin TESTING {$var} test TESTING2 {$var} test done OUTPUT ------ begin TESTING foo test TESTING2 foo test done This is exactly the output I would expect. Something else must be affecting the output. Are you putting strip around this, or do you have pre/post filters that may affect it? strip would take carriage returns off of the beginning and ends of lines, but it should not affect "{$var} text" if they are together. Jimmy Harlindong wrote:When I have $var = "abc" and inside the template: {$var} text the output will be "abctext" How do I make it into "abc text" (with space there) On another case: get "abctext" only when I write {$var}text in the template? This seems to work: Smarty_compiler.class.php inside _compile_tag() function:-- Monte Ohrt <monte@ispi.net> Director of Technology, ispi Inc. -- Smarty Development Mailing List (http://smarty.php.net/) To unsubscribe, visit: http://www.php.net/unsub.php/* If the tag name matches a variable or section propertydefinition,we simply process it. */ if(preg_match('!^\$\w+(?>(\[(\d+|\w+(\.\w+)?)\])|((\.|->)\w+))*(?>\|@?\w+( :(?>' . $qstr_regexp . '|[^|]+))*)*$!', $tag_command) || // if a variablepreg_match('!^#(\w+)#(?>\|@?\w+(:(?>' . $qstr_regexp . '|[^|]+))*)*$!', $tag_command) || // or a configuration variable preg_match('!^%\w+\.\w+%(?>\|@?\w+(:(?>' . $qstr_regexp . '|[^|]+))*)*$!', $tag_command)) { // or a section property settype($tag_command, 'array'); $this->_parse_vars_props($tag_command); return "<?php echo $tag_command[0]; ?>"; <<<<<<<< I REMOVEDthe \n at the end}Any comments? Or perhaps there is a workaround to this problem?