Jump to content


  • Posts

  • Joined

  • Last visited

  • Days Won


Everything posted by ryan

  1. This week we have some updates for the ProFields table field (FieldtypeTable). These updates are primarily focused on adding new tools for the editor to facilitate input and management of content in a table field. All the details can be found in the new blog post with an accompanying screencast video— https://processwire.com/blog/posts/table-field-with-actions-support/
  2. @TomPich I wouldn't recommend caching a field value like that, as I don't think you'll get any performance benefit since the amount of work for PW to do is the same either way there. What I'd recommend doing instead is caching the resulting output, which also means only loading the repeaterConsultants when it will be used to generate that cached markup. Whatever markup you generate, use that as your cache value, for example: echo $cache->get('consultants', 'hourly', function() use($pages) { $out = ''; $consultants = $pages->get(1027)->repeaterConsultants; foreach ($consultants as $consultant) { $out .= "<li>$consultant->consultantName</li>"; } return "<ul>$out</ul>" ; });
  3. @iank I had a look at this and it seems like it can happen if you omit homepage language segments, and are also using the PagePathHistory module. What happens is that PagePathHistory provides a shortcut for finding the page, leaving no language segments for PagePathHistory to detect the language. I don't think PagePathHistory supported multi-language URLs until 3.0.186, so maybe that's when the issue started. Though the entire URL-to-page mapping system has been rewritten since 3.0.165 via the new PagesPathFinder class, so who knows. I think maybe it's not been reported before because just about everyone is using language segments on the homepage, and I always recommend doing so, but didn't intend to require it. I've attempted a fix on the dev branch, it's just one line of code, so maybe it'll work just to modify your copy too. Do you find this fixes it there? https://github.com/processwire/processwire/commit/9e6b89cf9331fccb8f41ae0b1785e21c8c5d62f0
  4. @Robin S That's correct and good to point out. The term "move" refers to manual sorting, changing parent, or both.
  5. On the dev branch this week are a few issue fixes, but also some new features. The "Parent" field in the page editor (Settings tab) now is contextual to the page's template family settings. Now it will just show you the allowed parents, if you've configured them in the template. Previously it would show you a page tree to select from anywhere in the site. This saves you time, as well as the hassle of getting an error message after save, should you select a parent that isn't allowed by the page's template family settings. Next, the page tree "Move" actions (that you see when hovering a page in the tree) are now a little smarter. Previously, these Move actions would appear for any page that was either sortable or moveable. But now it also checks how many other allowed parents there are, per template family settings. If there aren't yet any other potential parent pages existing in the site, the page is no longer considered moveable, and thus won't show a Move action. This useful addition was added per Bernhard's request, as was the addition to a couple new classes used in the page tree to better differentiate between public vs non-public pages... something that I understand Bernhard's admin style may soon demonstrate. This week a question came up through email about how to make multi-language "required" fields require a value in all languages the page is active in. Currently multi-language required fields only require a value to be present in the default language. If you want to make them require all active languages, you can do so with a hook in /site/ready.php. I thought it was pretty useful so thought I'd share it here. Though maybe we'll have to add it as a feature. if($page->process == 'ProcessPageEdit') { $wire->addHookAfter('Inputfield(required=1,useLanguages=1)::processInput', function($event) { $inputfield = $event->object; $page = $inputfield->hasPage; if(!$page) return; foreach($event->wire()->languages as $language) { if($language->isDefault()) continue; if(!$page->get("status$language")) continue; // skip languages not active on page $value = $inputfield->get("value$language"); if(empty($value)) { $inputfield->error("Value required for language " . $language->get('title|name')); } } }); }
  6. Pruning is kind of an expensive operation as it involves finding existing log entries, exporting them to a new file, then deleting the old file, then renaming the new file to the old filename. I'd avoid doing it every time you save a log entry. I prefer to manually prune from the admin. The error messages seem to indicate that maybe the prune is colliding with the save. Likely pruning immediately after saving is happening in the same timestamp and not giving it a chance to consistently save a log entry. So at the least, I would disconnect your prune() operation from your save() operation, perhaps moving your prune operation into a LazyCron::everyDay hook.
  7. This week we have a new core version on the dev branch. Relative to the previous dev branch version, the new 3.0.237 version has 33 commits containing 20 issue fixes, 6 feature requests, and more. See the core updates section of ProcessWire Weekly 511, 512, 514 and 515 for more details. It's been about a month and a half since 3.0.236, which is a little longer than usual between version numbers, but that's largely because if the new Invoices site profile (see blog post). I'm off work tomorrow (Friday), so writing the usual weekly post a day early. As always, thanks for reading and I hope you have a great weekend!
  8. That sucks, but I'm guessing they might make exceptions for some projects, so I'll have to ask them. Still, not nearly as much of an issue as what CKEditor did. But if it stays GPL and they don't make an exception for any projects, then most likely we couldn't include TinyMCE 7 with the core. In that case, we'd develop it was a non-core module, and folks would have to install it as a 3rd party module (in /site/modules/). ProcessWire's core is completely separate from what people develop or what they might add-on their site, so they don't have to share the same license. PW is built so that modules are independent of ProcessWire in the same way WP and PW are separate applications that can run on the same webserver, or a website is independent of the webserver that's delivering it, or a browser is independent of the HTML it renders or JS it executes. https://processwire.com/about/license/3rd-party-files/
  9. This week there is a new version of the Site Profile Exporter module released (ProcessExportProfile). This is the module that was used to export last week's Invoices Application Site Profile. While this module has been around for a decade now, this latest version makes some nice improvements, which I'll cover below. @bernhard pointed out to me that he's using the module to make a new site profile, but he found it cumbersome to enter all the profile information every time he wanted to export the profile. So this new version now lets you update an existing profile with 1-click, making it much easier to re-export a profile. This version also adds a preview of what your profile will look like in the ProcessWire installer (the part where the user would select it for installation). Lastly, this version has several other small improvements and fixes as well. If you've ever thought of building a site profile, this module now makes it that much easier. This week I've also been starting to focus on the next long term client project, which is kind of a different and unique one that I look forward to. That's in part because I expect it will also involve a lot of improvements to the ProcessWire core and ProFields modules as part of it. Some of my favorite ProcessWire improvements have accompanied projects like this. There just isn't any substitute for real-life, large-scale projects when it comes to improving and optimizing the core and modules. Next week will be a shortened week here, so I'll likely post the usual weekly update Thursday rather than Friday. Thanks for reading and have a great weekend!
  10. @joe_g It looks to me like you've got an autoload module that's injecting scripts into your admin (a cookie content module or SEO type module?). Such a module should only be injecting scripts into the front-end, and not into your admin. So I would identify what module is doing that and stop using it, or update it to only load on the front-end. It you can identify what module it is, let me know and I can suggest what to change about it. Likely it's just a change to 1 or 2 lines.
  11. @joe_g In addition to what Bernhard mentioned, I would figure out the location of the prepareInjection.js file at the top of your JS console. That looks like a React file, and I'd wonder why it's loading here. Perhaps it's coming from some module? You should be able to click on it to identify the location, or view your Network tab and find it. You might see it coming from some site/modules/ModuleName/ directory? If so, try temporarily disabling that module to see if that's the issue.
  12. The invoices application site profile that I uploaded last week now has a companion blog post: https://processwire.com/blog/posts/invoices-site-profile/ Thanks for reading and have a great weekend!
  13. @fruid The cache time should be fine. I'm guessing maybe it just didn't work on the first use for whatever reason, and the non-working data was cached.
  14. @fruid Make sure you are using the newest version (version 3). Go to Modules > Site > ServiceCurrencyConversion. At the bottom is a checkbox to "Refresh rate data now". Check the box and save. Now make sure that you see HKD in the table shown on the configuration screen. If you don't, edit your /site/modules/ServiceCurrencyConversion/currencies.txt file and make sure it's in there, adding it if it isn't, and refreshing again. If you don't see any data loaded, I'm guessing that your server is blocking the outbound http request to the exchange rates API. You could modify the module WireHttp call to force it to use CURL or fopen, depending on what your server supports. Let me know if you need help.
  15. @teppo Not very straightforward, but here's a function that would tell you whether or not a translation exists for the given phrase in the given file, in the current language. Maybe we need a dedicated core function for it? function isTranslated($text, $textdomain) { $translator = wire()->user->language->translator(); $translations = $translator->getTranslations($textdomain); $hash = $translator->getTextHash($text); return isset($translations[$hash]) ? $translations[$hash]['text'] : false; } Example $text1 = 'About Us'; $text2 = isTranslated($text1, '/site/templates/about-us.php'); if($text2 === false) { echo "Not translated: '$text1'"; } else { echo "Translation of '$text1' is '$text2'"; }
  16. This week the new ProcessWire Invoices site profile has been released on GitHub here: Invoice Application Site Profile. This particular profile is much broader in scope than the others I've developed for ProcessWire, so will benefit from more descriptive information about what it includes, how to use it and modify it, and how you might build further from it. That info will all be coming next week in a new blog post. But feel free to download and install the site profile sooner if you'd like. If you are already familiar with ProcessWire, then perhaps most of it will be self explanatory. While this site profile doesn't cover everything that you might do with an invoicing application, it does cover everything that I've needed over the last year of using it with my clients. Though admittedly, I don't have a lot of clients, nor do I send a lot of invoices. Given that, when my existing invoicing service raised the monthly rate from $3/month to $20/month (a year or so ago), that's what motivated me to build this site profile. And it does everything my previous invoicing service did, and in fact does it better. While much of it was built several months ago, major improvements have been made to it over the last couple of weeks, preparing for it for release as a ProcessWire site profile. My hope is that you'll find this site profile easy to work with, and easy to build out further where needed. For instance, I imagine some may want to add in the ability to pay an invoice. It would be relatively simple to add in FormBuilder with its Stripe Processor plugin, or perhaps some other payment solution. But all my clients pay by check, whether physically or digitally, so I've not needed to add payment ability to the application yet. In any case, I hope that you find this site profile useful, and please let me know if you run into any issues with it or have suggestions for future upgrades to it. Thanks for reading and have a great weekend!
  17. @hellomoto It looks to me like you can drop the use of DatabaseQuerySelectFulltext. I don't see any text columns or fulltext indexed columns in your query, and I don't think it's doing anything at present. That DatabaseQuerySelectFulltext class is primarily for doing partial matches on text columns, most often using MySQL's fulltext indexes. Your schema doesn't have any columns that would typically be used with this class, and I don't see any indexes on the columns. Currently I'm wondering if you even need a getMatchQuery() method? Your schema is pretty simple, and it seems like the core Fieldtype::getMatchQuery() should handle it fine. If not, can you give me an example of a selector that is not working? (remove or comment out your getMatchQuery method completely to test). The only thing I can spot in your schema that confuses me is the CHAR column with no length specified. I've only ever seen CHAR(n), where n is the fixed length of the CHAR column. Would "CHAR" be the same as CHAR(0), which either holds a 0-length string or a null? If so, then this may be a reason to have your own getMatchQuery, as PW consider blank string and NULL to be the same when querying the DB. So a simpler solution might be to just change it to a TINYINT, CHAR(1) or ENUM, so that you can be explicit about the ON or OFF (1 or 0, etc.) value, rather than trying to distinguish between two empty values (blank vs. null). Then it should work with the core getMatchQuery. The sf1 and sf2 columns may need to be indexed for best performance though.
  18. There are a few commits on the dev branch this week, but nothing particularly notable. I had to do some client work this week but also ventured back into the Invoice site profile, which I mentioned quite awhile ago, but hadn't yet released. I'm hoping to finish that up and release it as another site profile option for ProcessWire very soon. Perhaps as soon as next week. I had originally planned on building out that Invoice site profile quite a bit more, but having used it for my own invoices for many months (maybe a year?), I think it's better to keep it simple. Otherwise I'd be making assumptions about what others might need. Even in its relatively simple state, it suits my own needs well. And I think if it gets more complex, then people are less likely to explore and modify it, making it less useful as a site profile. The site profile is also simple enough right now that it doesn't need to be a Pro module or need Pro-module support. Since we don't have a lot of site profiles to choose from, I'd rather keep it free. It is admittedly more of a mini application than a website, which is why I think it might bring some more diversity to the available site profiles. Chances are it won't replace whatever invoice service you might be using, but I think it's still pretty useful, whether using it to create invoices, or just using it to explore more of ProcessWire. More next week. Have a great weekend!
  19. @hellomoto It looks like you are building a custom Fieldtype? If so, can you paste in the contents of your getDatabaseSchema() method? Matching an empty value can depend on the schema, though usually is handled by an OR condition that matches an empty string or the non-presence of a row. Matching the non presence of a row is done by performing a LEFT JOIN on the field_* table along with a WHERE condition that matches "IS NULL field_table_name.pages_id". It can be a little tricky, so luckily PW handles this situation for you internally, and doesn't require your Fieldtype to handle it unless you tell PW that you want it to… To describe further, ProcessWire matches empty values automatically, and so may not even call your getMatchQuery() for most empty-value matching selectors. But if PW isn't correctly matching it for you, then you may need to implement your own "empty" match logic. PW asks your Fieldtype::isEmptyValue() method (if implemented) if you would like to handle that situation (details). If it sends a Selector object to your isEmptyValue() method and your method returns true, then you are telling PW you want to handle matching the empty value: public function isEmptyValue(Field $field, $value) { if($value instanceof Selector) { // tells PW your getMatchQuery will match empty values return true; } return parent::isEmptyValue($field, $value); } So if you have a method in your Fieldtype like above, then your getMatchQuery() would have to have logic to match an empty value. I don't think your current getMatchQuery() would correctly match an empty value, so I'd need to see the schema to identify what would be needed.
  20. @diogo PAGEGRID also looks amazing by the way. I don't have much experience with page builders, but it seems like PAGEGRID is really well thought out and does everything you could want.
  21. @diogo @jploch Wow that is awesome! It's completely different than anything I've seen before. It's really fun to scroll through too. I'm curious about the development side, how do you take over the scroll behavior in that way? If I view the source, there are two completely separate <html> documents in the output, how is that possible? 🙂
  22. On the dev branch this week we have a good collection of issue fixes and feature requests. The dev branch commit log has all the details. One feature added this week which I think could come in very handy is #520 (via @Jonathan Lahijani) which adds the ability to hide individual images in an images field. When an image is hidden, you can see and work with it in the admin, but it gets removed from the field value on the front-end of the site at runtime, effectively hiding it. I know I'll use this a lot, particularly on photo galleries where I may want to remove an image or two from appearing in the grid of photos, but don't necessarily want to delete them. Images can be hidden (or unhidden) from the Actions select of each image, where you'll see a "Hide" option (or an "Unhide" option if the image is already hidden). Hidden images are also dimmed out when viewing the images field in the admin. On the API side, you can hide or unhide images and files using $image->hidden(true) to hide, $image->hidden(false) to unhide, and $image->hidden() to get a true or false as to whether or not the image is hidden. Though this will only be useful on unformatted field values, since hidden images are automatically removed from formatted field values. The same can be used with regular file fields, but we don't currently have a UI/interface for hiding or unhiding items from regular (non-image) file fields. Likely we'll add one soon, but I figured it's likely to get more use with image fields than file fields, so figured we'd start there. More next week. Thanks for reading and have a great weekend!
  23. This week continued the addition of new feature requests and PRs, and the version number has been bumped to 3.0.236. In fact, the addition of feature requests is the theme of version 3.0.236. My favorite addition for this week is probably feature request #474 which adds support for opening and collapsing family groups of repeater items together as one. To enable, see the field "Details" tab setting in "Repeater depths/indents" > "Open/close items as a family?". The same commit also made it possible to disable the automatic scrolling to new repeater items when you click the "Add" button, should that suit your need. Both of these updates also apply to Repeater Matrix as well, but don't require a Repeater Matrix upgrade. More updates can be found in the dev branch commit log, and ProcessWire Weekly has great coverage of what's new with version 3.0.236 in issues #507, #508 and #509. Thanks for reading and have a great weekend!
  24. I've always enjoyed PHP, but never really enjoyed Javascript until I started using jQuery. I continue to enjoy using jQuery much more than vanilla Javascript. Using jQuery has not always been about filling in a gap of Javascript so much as it has been about preferring and even enjoying the interface. It’s fun to use, and Vanilla JS not as much, at least to me. PW has some of its API inspired by jQuery, and together they have always felt just right. When it comes to open source stuff, I like to focus on tools that I enjoy using, as that's what keeps me interested and keeps me going. If we were to take jQuery out of the admin, it would be a huge amount of work, and then leave something that would be less interesting to maintain. So I’m not so enthusiastic about taking jQuery out of the admin. Where I would be enthusiastic about it is with front-end modules that might currently be using on jQuery and don’t necessarily need to. Take FormBuilder and LoginRegisterPro as examples (though there are many more). Perhaps those modules don’t need to require jQuery unless one of the Inputfield modules in use requires it. That way, modules like that aren’t introducing jQuery in an environment where it might not otherwise be in use. And maybe there are some Inputfield modules that currently use jQuery and don't need to. Since Inputfield modules can be either admin or front-end, it makes sense to use vanilla JS when possible with those. So yes, I'm all for reducing or removing the use of jQuery in spots, but not so interested in removing it from the admin.
  25. @ryangorley Nice work, great sites! Would you consider adding them to the sites directory? https://processwire.com/sites/submit/
  • Create New...