jom

In the meantime I figured out in which installation the error occured. I upgraded 3.0.210 to 3.0.226. It works well now, running php 8.2.

back in office. sorry guys, I can't reconstruct where it was. All sites running on php 8.1., so I guess I rechanged php version after template change.

i'm on holidays until end of august, can't deliver more details until than 😐

samesame. downgrading to php 7.4 works. Using ProcessWire 3.0.210.

ok, this works. I begin to understand slightly more about the mechanisms ? - thanks!

I have a follow up question: How can I get access to the active page from within the __constructor? Something like this:

public function __construct(Template $tpl = null) {
    parent::__construct($tpl);
    $this->header = page()->title; //does not work
}

I guess, I have to pass the active page to the constructor? But how/where?

In my page template file programmliste.php I refer to the custom class like this:

$content = page()->header();

Including the active page here is too late, as the the custom class has already been created, I guess?

ok, beauty, this works. thank a lot, I appreciate very much!

Hi

I've started working with custom page classes. The basic example from Ryan's blog post works fine. Now: I'd like to set some object variables, which I can use in my methods. Usually this is done within the __constructor() method. My simplified example:

<?php namespace ProcessWire;

class ProgrammlistePage extends Page {
    public $header = '';
        
    public function __construct(){
        $this->header = 'some text';
        // parent::__construct();
    }

   public function header() {
        return $this->header;
    }
}

This results in an error in my nav function, related to Page.php (Trying to get property 'slashUrls' of non-object in /home/.../wire/core/Page.php on line 3245).

What am I doing wrong? I believe I should call the parent::__constructor(), but it didn't work either. I would appreciate a hint, where to look in the docs, because I feel I miss some basic understanding of the underlying mechanisms.

Thanks in advance!

Chäs & Co is a b2b cheese trader. They love good cheese and are proud of their excellent knowledge und connections to small cheese producers. They maintain a custom made database and used to copy/paste (!) the cheese infos to the former CMS (Contao). Since substituting the database with ProcessWire was no option, they export now the database to a csv file and import it to the new website (expanded version of module ImportPagesCSV).

From there, the pw magic begins. The staff works with the admin interface: They search and find products in notime using the build-in search field (using cheese names or id. Using ids only possible by adding hook, see this forum thread). They generate price lists easily based on a template, providing checkboxes and option buttons for some restrictions, choose i.e. the cheese origin and price audience (all implemented as page selectors). By saving the page a PDF list gets generated (using mpdf, a php pdf library).

The visitors can sort and filter the products using a form by shuffle.js. Product details are shown in a popup (fancybox). There's also the option for generating a nice pdf data sheet. No CSS framework being used.

Other modules used: ProcessJumpLinks, ProcessProMailer, ProcessWireUpgrade.

Visual design: Nicole Haller

beauty! Works perfectly well - thanks!

I got stuck in the forum search, looking for admin live search and integer values.

I cite here the code, just in case:

On 5/7/2020 at 11:56 PM, Robin S said:

 

$wire->addHookAfter('ProcessPageSearch::findReady', function(HookEvent $event) {
	$selector = $event->return;
	$q = $event->wire('input')->get('q');
	// If the admin search query is a number
	if(is_numeric($q)) {
		// Get the individual pieces of the selector
		$selector_pieces = explode(', ', $selector);
		// Modify the first piece so that it includes an ID clause in an OR-group
		$selector_pieces[0] = "({$selector_pieces[0]}), (id=$q)";
		// Replace the original selector
		$event->return = implode(', ', $selector_pieces);
	}
});

i'm at the same point as Ipa: need to add a number field to the admin live search. do I have to copy the module to the site/ folder and change it there?

just for documentation purposes: looks like Ryan added an option in $config:

$config->maxUrlSegmentLength = 128;

see also: 

I just noticed that the folders are deleted now. So maybe it's all fine now. Does anyone know how the deletion works? Does PW execute some cron jobs from time to time which runs over the temp directory?

Olsa, thanks for Somas post. This might be a solution, if wireTempPath won't work. I still hope to get to work. The question is actually quite simple: How can I get wireTempPath to delete it's folders?

Hi everyone

It seems that I don't fully understand the wireTempPath() function and I need some help.

I use wireTempPath() to create a new location in assets/cache/WireTempDir and than copy a pdf from the assets/files/page folder to the new folder. I want the file to be accessible only for a limited time, that's why I use wireTempPath.

The file seems to be copied to the right location, but gets deleted right afterwards, according to 

As mentioned in the topic above, 

$wireTempDir->setRemove(false);

prevents the file to be deleted. But I like the file to be automatically deleted after a few days. So, how can I do that?

My code so far (everything works, but the automatic removal of the tempDir folder):

//generate and show download link
$folder = time(); // timestamp as temporary folder
$maxAge = (int) $settings->options_downloadlink_valid_hours * 3600; //tempDir wants maxAge as seconds                
$options = array(
	'maxAge' => $maxAge
);                        

$wireTempDir = wireTempDir($folder, $options);
$wireTempDir->setRemove(false);
$src_file = $page->ebook_download->filename;

// Create a new directory in ProcessWire's cache dir
if(wire('files')->mkdir($wireTempDir, $recursive = true)) {
  if(wire('files')->copy($src_file, $wireTempDir)){                        
    //get subdirs from tempDir:
    $pos = strpos($wireTempDir, "WireTempDir");
    $subdir = substr($wireTempDir, $pos, 100);                

    $out .= "<p><a href='" . wire('pages')->get('template=passthrough')->httpUrl . "?file=" . $subdir . $page->ebook_download->basename . "' target='_blank'>$page->title</a></p>";  
  }
}     

I appreciate any ideas - thanks!

Oliver

Ok, thanks for the opinion!

Hi

I'm also playing with Ryans new module LOGIN/REGISTER/PROFILE (LoginRegister). I wonder if it's possible (and if it makes sense) to use it without the login/profile functionality. I need a registration form for the newsletter. The only fields I need are mail address and user name. I can't get rid of the required password field. I could hide it in the frontend of course and fill it with some generated data, but that's not very elegant. Actually I already have a working solution with SimpleForms, I just thought it might be more straight forward to use Ryans basic module. - What do you think?

In the meantime I got the mod_rewrite error log. Indeed, there is an error regarding my topic:

[Wed Oct 05 18:42:00.073367 2016] [core:error] [pid 85277:tid 34669321216] (63)File name too long: [client 12.34.56.78:54004] AH00127: Cannot map GET [here follows the repeated path]

It seems the paths gets in a loop - it is about 10 times repeated in one string. Now I'm lost - where to go from here? By the way in the first repetition the path is including the urlSegments, the following repetitions are without.

I'm on a shared hosting environment, not sure if I can get the mod_rewrite logs. I asked the hoster for it.

In the meantime I implemented a workaround. Since the urlSegments are not vital, I skip them, if a 404 is based on a too long page name. For this I wrote a function which hookes into pageNotFound:

public function init() { 
  $this->addHookBefore('ProcessPageView::pageNotFound', $this, 'displayItem');
}
        
public function displayItem($event){            
	// redirects to url without segments, if there is a long page name involved
            
	$pageWithLongTitle = '';            
	$file = parse_url($_SERVER['REQUEST_URI'], PHP_URL_PATH);            
	$urlParts = explode('/', $file);            
	$urlWithoutSegments='';
            
	foreach($urlParts as $urlPart){
		if($urlPart){
			$urlWithoutSegments .= $urlPart.'/';
			if (strlen($urlPart) > 110){
				$pageWithLongTitle = $urlPart;
				break;
			}
		}                
	}

	// if there is a long title in url and a page with this title exists, then redirect
	if($pageWithLongTitle && wire('pages')->get("title=$pageWithLongTitle")){
		wire('session')->redirect('/'.$urlWithoutSegments);
	}            
}

Thanks for the help!

Thanks for your input, BitPoet. I've tried with several computers and browsers on different locations. No difference.

The request URI matches to the logged URI. Not sure about the "it" parameter: I inserted 

echo wire('input')->get('it');

into index.php. No result. Did I get you wrong?

There weren't any third party modules present when the problem occured the first time (apart of  ImportPagesCS).

Where would you set the limit for the page name? Add a module to the page->save hook?

Yes, urlSegments are enabled, everything works fine with shorter urls. 

Hi everyone

Since this is my first post: thanks for this beautiful CMS/CMF und thanks for this high quality forum.

I'm struggling with following: A client imports a list of books from a csv file (using module ImportPagesCSV). The page name gets autogenerated by the core (it takes the value from the title field). If the title is very long, the page name gets truncated to 128 characters. This works fine, as long as I don't use urlSegments. Using them, PW returns a 404:

http://domain.tld/...long-title-hier-128-characters.../ : works
http://domain.tld/...long-title-hier-128-characters.../urlSegment1/urlSegment2/  : does not work

It works fine, when I shorten the page name manually to e.g. 100 characters. This is not a good practice, since the client imports the list every couple of weeks.

Is this a correct behaviour? Should I limit the page name via API to 100 characters? If yes, would I do this in the ImportPagesCSV-Module or better implement as a hook?

Thanks in advance for some light in the dark!

Oliver