gornycreative

Was not Tracy after all. My bad.

So I found another site built off the same template that had not had Tracy installed prior. After clearing the module cache and letting it rebuild, the site broke - so it must be another module that needs some rewarming from 161. Moving on meaning I need to start dropping other modules until the site works, since Tracy doesn't logically seem to be the issue.

Okay. Regarding the switch, there was no need to touch the htaccess at all. I had a single file that included a bunch of variable sets that were global (settings factory stuff) and a few variables set from the home page (gotten through pages() ) - none of the things I pulled required access to page() and those got moved to _init.php Everything else got moved to _main.php - which at the time was just a bunch of components in a layout that I was going to pick and choose and render in the main template file into regions - this was not supposed to be anything super crazy or elaborate. Then, when I turned off the prepend/append and switched to a plain jane HTML hello world scenario and still got a blank page, I got even more frustrated. I split a text file into three text files - pretty routine stuff. And boom. I do agree with you that it doesn't make sense for the HTML to still not render after uninstalling Tracy and putting the original templates back if Tracy was the problem. I will move on from this I guess. If somehow you come across a situation where you run into the same thing, please let me know.

The thing is, I had another site built using the exact same basic template setup as the original. All that I did to that site was upgrade it from 4.21.17 to 4.21.19. Nothing else on that site was touched - it went from loading the templates fine to a blank screen with . I'd rather not sacrifice another development site to prove the point without really understanding what changed between versions. Are you running this on 161 or 148? The strange thing is - I set the request logger on for the site - the HTML is getting passed in the response. I have no idea why the browser is not rendering the code. Chrome is up to date (which I know can be a problem sometimes).

It's really weird. And I'm afraid to experiment too much more on other working sites as I don't want to break anything further. I uninstalled 4.21.19 from the modules area, cleared all cache and then double checked the asset cache FileCompiler area. Turned off opcache in PHP. turned off as much caching as I could find in modules. Removed all traces of files. Purged CloudFlare cache and made sure dev mode was on. I've turned off prepend and append and even substituted a very basic HTML file as the output template with no look-ups. There are no PHP or LiteSpeed errors logged. There's nothing in the console. I decided after cleaning out the system I'd reinstall Tracy fresh. What I don't understand is that it is as if the PHP is working, but the HTML is just not getting output. With devtools open there not alot going on, but Tracy has an HTML DOM that it is sitting in, which is even stranger. I know you've pushed priority up on loading - perhaps it is possible there is a module that is failing to load immediately after? Maybe I can look at the modules table in adminer and sort by priority? It's absolutely nuts - when I look at source for the loaded page all I get is <!--tracy debugger bar --> and then the script tags. That's in an incognito window. Chrome puts it into a makeshift DOM but firefox has no idea what to do with it. There it is just a blank page.

So I am at a total loss. I was in the middle of working on a draft site and had recently updated Tracy to the latest version on one of them 4.21.19 from 4.21.17. I was in the process of revamping the site from delayed output to markup regions and then when I went to preview (which I hadn't done) the screen was blank with the exception of Tracy. No source, no nothing. No errors. Tracy reported the page object loaded, all the children, etc. I reverted to the old version of the templates. No difference. I compared it with another copy of the same site on a different domain that still had 4.21.17 installed. Front-end looked fine. I updated the second site to 4.21.19 - blank screen on the front end, no errors being shown, nothing. All else being equal, I have to assume there's something in the change from 4.21.17 to 19 - but I purged 19 from the site/modules, cleansed the copies from the FileCompiler cache directory - got rid of everything I could to remove it. Both site are still showing blank with nothing in the PHP logs and no displayed errors. Backend still works perfectly fine. I don't really know what to do next, but I'm pretty frustrated about the situation. If I had error codes to work with that would at least be something. If there are errors (which I could create) Tracy was reporting and displaying them. But I can't get the site templates to output anything. Not even replacing with generic echo. I'm running 3.0.161. Is there something in the page rendering hooks that got installed that I might need to manually remove? Any ideas on how I can get my sites back? At the moment I am probably going to pull down another site made with the same template and try doing a diff across the site/wire directories. I don't really know how to troubleshoot a situation where no errors appear, but no markup either.

I haven't been able to revert, and so I removed the plugins entirely and two sites of mine are now completely broken on the front end. Not happy about this at all.

This seems to be related to the new version of Tracy - I updated a second testing instance with all other files being equal and the same behavior occurred. I'm going to try to knock out the cached version and revert and see if the site returns.

So I'm running into a weird issue and I don't know what I am doing wrong exactly. I'm trying to convert a site from delayed output to markup regions. I've tried separating out some of the original parts between the three files: _init.php - I am using an init because I have a few config variables and other bits that are used to pull global links for css and other bits. home-page.php - which is where I plan on working through regions based on the home page content sections that are special. _main.php - here I haven't really changed much. It's got a namespace and does refer to some values from the page object, but nothing fancy. Generic scaffolding. There are echo commands in there, but based on the example Regular profile that seems to be fine. As soon as I made this separation, I loaded a view using the template and got a white screen - granted, I usually know what that means, but then I see Tracy has loaded! I go through the template info, it has loaded all of what you would expect. No php errors showing and so I am totally lost. I've even tried just running simple output in the template file - a print_r or an echo. Nothing. If I disable loading the init, I get php errors because the variables I'm trying to call are not defined, nothing shocking there. But if I include the files, I get blank output - no HTML at all. But I get Tracy at the bottom and I can access object properties/arrays etc within the page context from the console. The only thing I guess that is weird, and I will try digging into this, is that I don't actually have any markup regions defined with ids anywhere yet. Because I'm just starting the conversion. Is there a step I am missing someplace that is keeping the HTML markup in _main.php from showing up - just being blank?

Another change I had to make - the Awesomplete flyout sets a z-index of 1 - but ui-button in this theme also sets this to 1 and the hover layer to 0. I modified this in AdminThemeBoss/uikit/custom/theme/components/forms.less, line 122-135 button.ui-button { position: relative; z-index: 0; &:hover { z-index: 0; box-shadow: 0 5px 15px @theme-primary-color-rgba; +.ui-button { z-index: 0; } } } Changing the z-index values from 1, 1, 0 to 0, 0, -1 - so they'd hide behind the awesomeplete.

Looks like wireshell.pw has been hijacked by nefarious actors.

Am I right in thinking you could use this to install Duplicator and then install a Duplicator backup package? Would that be an alternative to building out a full blown recipe including all the modules, etc.? I have a number of installation images that I like to deploy for different use cases and most of them also include some default documentation. Modules, templates, fields, vendors scripts, etc. I suppose the advantage to a recipe would be always pulling the latest master branches for certain dependencies?

So this last topic just above me - where the person trying to use multiple dropdowns to send ajax filters to a page output container. I think this is what I am trying to figure out. I've got someone interested in moving from WordPress to Processwire but they really like the FacetWP plugin functionality and so I'm looking at techniques for replicating that process. If you've never seen the plugin: https://facetwp.com/ It seems very similar to what Ryan uses in the bike tours website. I do believe there is some caching involved in terms of result set combinations? - I haven't looked deeply into the FacetWP code, but my impression is that it runs something like this: Grab an initial catalog defined by a search selector(s) definition - this would be provided by the configuration - what pages to list in the results and then which page references fields to be able to filter on. Grab distinct page reference fields and selected values for those - package them into initiall object to output Push those template output items to a paginated search results template and a dynamically generated checkbox fieldset, or multiple selection box, etc navigation template. Live on the client, when an option is selected, send the new selector modifications back to the original search results template - adding the option to the original property selector At this point, there is an option branch that facet employs - you can either keep the full list of possible options and then grey out the options as unselectable where no valid combinations exist, or you can hide the options that are no longer valid - I don't believe there is any interaction between the typical wordpress search box and the facet filters - but I don't see why you couldn't have a selector title tie-in, for example - though precaching result combinations goes completely out the window there... Inject the updated options selectors template and paginated results. Does this sound about right? So I would want to call a navigation template/fieldset and the paginated result set/controls with ajax and then update both templates with ajax? I don't really know how the facetWP plugin degrades without javascript - I think it may just show the full list of checkboxes as a control that uses a plain form destination with getstring parameters passed. I suppose it would just go back to the search page results template! On the server side, could I just create a single search results output template that detect whether it should return a full page or just the page collection and pagination number depending on whether the request is ajax or not? I supposed I'd still need to have a separate filter generator template? Maybe it could all get loaded by the same search results/filter form template?

I just wanted to let you know I ran into an unusual error message with PW 161 revolving around selectors. I was getting a strange popup whenever page creation was operating - stating there was an invalid selector : | I uninstalled and then reinstalled the plugin and the problem went away. Not entirely sure why that would happen (I clear compiled files, cache etc) Just thought I would put a note here in case you were able to replicate the issue. I was upgrading from 160 to 161 and I know Ryan was doing a lot of selector work.

@kongondo quit making weird vuecat YT videos and get back to work. Or was that a new antificial intelligernce module you are building? ?

Just making a note here that Form Builder doesn't seem to see this field anymore as of the v3 - not sure if the last fix you mentioned above for v2 would work. I will give it a try.

I have found that the jquery date picker is by default set to a dark background with black text. By adding these lines in the blue.less files (etc.) @pw-datepicker-background: @theme-primary-color; @theme-button-secondary-background: @theme-muted-background; I am able to at least get the background to match the rest of the theme color variables but I have tried many things and cannot find the override that is keeping the text black. In the forms.less I have changed the color of the various ui-state element to the @theme-text-color-inverse variable which should make the text white but it is not working. Has anyone else had this problem or found a workaround yet?

That makes a lot more sense in terms of advantages - I will need to look into this some more! Thanks.

This looks great - I don't personally care for these types of editors because I'm an older user, but many clients have asked for this type of builder. I'd probably get used to it and enjoy it once I got into the habit.

When this module is turned on, my CKEditor modals get buried under an overlay I can't obviously click through. The overlay gets set to a ridiculously high z-index. Normally the overlay sits at I think 10000 and then the modal table gets throw above that at like 10010-10060. When AOS is enabled - with no submodules activated AT ALL - there is a css override in aos.min.css and here's what it puts out in my install: .cke_dialog_background_cover { z-index: 103000 !important; } I'm not sure why that rule is in there, all that I know is that if I turn off AOS completely, it goes away. Trying to edit source code, trying to compare revisions - yeah those don't work. Am I the only one seeing this problem? It doesn't matter if the CKEditor autoload segment is turned on or not. I'm using WP 3.0.156 and AOS 2.0.20

This worked really nicely, but I have run into an interesting problem. I'm trying to use this in coordination with Settings Factory, and the issue I guess is that there is no real way for RockLESS to detect that the variables going into the CSS call have changed - so I can change the options all day long but the css won't regenrate because RockLess hasn't detected any changes in the less file it is calling - right? Would it make sense to hook the settings factory so that when the settings are saved the compiled file is deleted? Or would it be safer to just touch the less file the less file RockLess is trying to load? I'm curious if any others are trying to do something similar.

Thanks for those thoughts. I did consider using Hanna Code - I think for collecting the information that is a solid recommendation on its own. If all I was interested in was collecting footnotes, that would be ideal. The secret sauce of the plugin was what Jquery was used for - and while I agree that I'd rather not rely on javascript for the functionality, I don't know if the same effects could be achieved with pure CSS, but I will look into it. What the vanilla jquery placement function does is wicked simple: function placeNotes() { for (n = 1, refCount = $('a.side-matter-ref').length; n <= refCount; n++) { // Iterate through notes and position each var ref = '#ref-' + n; // Reference anchor var note = '#note-' + n; // Sidenote var refPosition = $(ref).position().top; // Position of reference anchor var notePosition = $(note).position().top; // Position of sidenote var noteOffset = refPosition - notePosition - noteAdjust; // Get current offset between reference and note, minus noteAdjust var finalOffset = (noteOffset < 0) ? 0 : noteOffset; // If offset is negative, set to 0 (prevents layout problems) $(note).css('marginTop', finalOffset); // Position note } } noteAdjust is a tuning parameter you set on the admin side that allows for you to adjust the height in case there is a difference between your article parent and your notes aside - but in my implementation I replaced this because it was simple enough to calculate so long as you could name the containers where both the article and the aside would sit. I also had to make some adjustments to prevent notes from stacking on each other by keeping tracks of an aggregate lastNoteBottom that made sure the note top would not blindly go next to the two refs on two consecutive lines and overlap each other - but I think perhaps setting the note to clear would resolve that issue. I'll take a look at your textformatter plugin. Thanks for the feedback!

So I have been digging more deeply into processwire working under the hood and I would like to try building a module. So there are two defunct modules I'd like to synthesize - or at least put the pieces together to make work for some of the more serious writing sites I've worked on. In WordPress, there was a plugin called Side Matter: https://wordpress.org/plugins/side-matter/ This plugin allowed a person to use a shortcode to indicate a reference [ref] - the original plugin would: process the shortcode contents (the note) look for nested [ref] shortcodes (and ignore them), process the shortcodes found in the note (if any), assign a note number or letter (depending on the type of numbering you wanted to use) replace the original shortcode with a linked superscript to an ref-1 anchor (optional) add the identifier and reference content to an aggregate div stack feed the content of that aggregate div stack into a second [ref_list] shortcode There was then a jquery script that would look at a stated 'parent' container and normally you would have an aside widget where you would put the [ref_list] shortcode or something like that. What you'd end up with was a height matched set of side notes next to the main article content. If you were crafty and used responsive frameworks like UIKit, the jquery code made a point of making sure two notes never overlapped and never were higher on the page that the original article reference. So in the case of having a responsive layout where the aside slipped beneath the article, the sidenotes would convert very conveniently into footnotes. I was looking at this module - which has been orphaned at WP for years, and I was also looking at the MarkupShortcodes by @Nico Knoll to manage the shortcode processing - at least as a starting point: https://github.com/nicoknoll/MarkupShortcodes Both of these codebases are pretty old, has this sortof plugin already been tried? Are there more recent resources/examples I should use to approach this properly? I think both of these plugins were released in PHP5 - I've been trying to get caught up on the architectural changes of PHP7 (it has been a while). Any thoughts or advice more specific to the kind of project I am looking at would be appreciated - would it be better to complete start from scratch using a more modern template? Thanks!

It looks like this module could allow for the introduction of field content prior to compiling - is that accurate? Is there anything special that needs to be done to prioritize it at runtime before ProCache? I'm looking to be able to push field values from SettingsFactory into the LESS file preproc.

Anyone curious, this module needs a namespace for v3. The module file should begin: <?php namespace Processwire; use DateTime;