tsdtsdtsd

The problem with uberspace is that it clearly depends on which virual machine you will get assigned to. Some have great performance, some are rather slow. It's a great service, but usually I had problems with sites runninng highres image stuff. For example, the initial resizing of images seem to take ages sometimes because of CPU workload. It just depends on your individual project.

Hi, I don't have great knowledge of the processwire core, but I think that the following lines in wire/core/Pageimage.php are responsible: $basename = basename($this->basename(), "." . $this->ext()); ... $basename .= '.' . $width . 'x' . $height . $crop . $suffixStr . "." . $this->ext(); $filenameFinal = $this->pagefiles->path() . $basename; $exists = file_exists($filenameFinal); if(!$exists || $options['forceNew']) { // Generate and save ... That means, that indeed the image options are not a dependency for the image file generation. One can argue about that, but I think that they should be.

Think you are right here Martijn. The usage notes are missleading though: But well, thats fine for me

Hi, the HannaCode snippets are currently stored in site/assets/cache/... I was wondering if this is the right place for them. Usually people tend to skip cache/* folders in backups and repositories (.gitignore etc). This could lead to serious data loss if one isn't aware of the fact where the snippets are stored. What do you think? Cheers

Not sure if I understodd you correctly. You are already iterating over project_tags and spit out one html tag for every element in project_tags, right? Usualy you would store the color information somewhere to keep it dynamic, but if you want to keep it simple, create a mapping table. Something like $colorMap = array( 1025 => '#ffff00', 1029 => '#6d8d24' ); forach($project->project_tags as $tag) { echo '<span data-color="' . $colorMap[$tag] . '"></span>'; }

Hi Guru, one thing that comes to my mind is caching: If you are caching your pages, you can't just simply put your counting logic into a template as PHP. You need a request via javascript or a counting pixel. I understand your 'overkill' argument here, but I guess the extra data you could mine wouldn't hurt anyway I would still prefer tools like Piwik or Google Analytics. No need to reinvent the wheel here. Those tools can get you a lot more reliable data than you could implement on your own in a reasonable time. I installed Piwik on a second host to split the load btw. If you still want to keep it simple and do it on your own, I would just use some fields that are present on any template. Maybe create a module for that and move the fields to a 'Stats' tab or something like that. But that surely depends on what kind of data you want to keep track track of. That method would wok with simple request counting, but not if you want to track devices and stuff for example. I don't think that $log is the right place for that kind of content, but that's just me. The $log API logs into files, which could get nasty with heavy load in terms of IO, response blocking and maybe race conditions. Also you would need to analyse/calculate big files, which isn't fun... Cheers

Hi, take a look into the API documentation, there is something that exactly fits your needs: $page->rootParent Cheers

Thanks Martijn! Just played around with it a little myself. As a great marketing-guy (who I am not), let me clarify it this way *cough*: processwire encourages the user to use a sophisticated url structure (in favour of e.g. security, SEO) through forced non-cachability of oldschool parameterized pages. Does this make any sense? Sounds good to me Anyways, to be a little more technical again: What I do want to get cached? Static or dynamic content, which can be influenced by 'some kind of' variable. This is possible through "url segments" configuration, I don't need to use classic GET for that job -> check What I do not want to get cached? Content that is specific for a session/user. Content for a single session can still be dynamic by GET/POST-parameters, but we don't wan that to be cached - and it can't anyway -> check So my initial proposal isn't necessary at all and this thread can be closed.

Hey mates, when setting up template caching, we are able to define "Cache disabling GET variables". While this is very useful, I would like to propose the ability to define "Cache forking GET variables". The given variables (and their combinations) would then create seperate cache files for different values. Basicly, whenever we use GET/POST parameters in our application, we are not able to serve cached content, while there are lots of situations where it would make sense. I think that this is a mayor caching mechanism in general. The performance impact should be minimal, but I smell a little security problem: I/O impact could get interresting, if someone bombards your site with requests with random values for known variables. I have no clue how the big caching engines out there manage this problem. Any ideas? Please discuss Cheers BTW: Does someone know how processwire handles the caching with url segments and page numbers? Never tried...

The site is currently hosted on an utterly lazy shared webspace. Guess that alone makes a huge difference Also, caching is currently completely deactivated - and I have no senseful reason for that, mark it as "testing" I'll activate it in the near future

Thank you @processing and @gemini I have to do something with the header i think, it's too simple, I don't like it anymore. Guess I need to develop some logo designing skills

Had a lot of fun on the site, great job! You could optimize here and there (move JS to bottom blabla..) but nevertheless a great site. Bug found: I'm on a low-res laptop atm (1366x768). Texts on the cases detail pages are cut and I can't scroll down to read the rest. Not much though, I miss 2-3 lines of text.

Very well done mate! And nice to see a font loader There are some console.logs, maybe you want to remove them.

Yeah sure, you are absolutely right. I'm just a cynical guy

Hey Soma, that is great! Perfect layout for a newspaper site, responsive, very fast. good job man!

All that backend modules you created look awesome and like a real bunch of work, nice job! but ugly as hell

Hi, i just launched my litte blog orkork.de It's a german blog about nerdy topics like IT, development, photography, games, movies and more. Responsive design, built on top of bootstrap. Different teaser types and widgets are configurable in the backend, for every category node. I use require.js for the internal js stuff and use gulp as a building tool for less compilation and css/js minification. unveil.js manages lazy loading of images, including retina versions. Though, it's not heavily optimized yet, but you have to start somewhere I would love to get some feedback, if you have any! Cheers

Hadn't thought about that -.- Sorry for the inconvenience caused... and thank you guys

Hi, I'm in need of a Fieldtype which I couldn't find in the modules database and as I'm new to module development I wanted to ask for your opinions. What i need is a Filedtype like "Page" (the PageArray multiselect to be precise), but instead of choosing pages from the tree I need to be able to choose multiple users, which for example I will iterate over later on and compare with the current $user. My question is - before I dive into module development and have a hard time figuring it out - is my fieldtype even possible? As far as I can think of, I would need to retrieve a list of users from the database directly and work with it. Also, if you have any awesome hints for a mod-dev-noob (like links to good articles, except ryans guide) it would be appreciated but that's not that important as there are lots of modules I can learn from anyway. Thanks, and have a good night

Hi, sorry for the not so awesome topic headline. I would like to know how well processwire performs on a huge site with high traffic. I'm speaking of sites with 50-100k+ articles and dozens of millions page impressions per month. I have the feeling that processwire could perform not that bad if the hardware is well scaled and with proper caching. But i think the backend user could get usability problems with that amount of articles and pages. So, whats the biggest known site made with processwire? Any numbers available? Also I would like to know if anybody has experiance with multi-server setups. Lets say one DB-Server, one for the backend, several frontend servers and a nginx loadbalancer. As long as you don't use sessions features it should be doable with processwire i think. Thanks

Looks promissing, cool little helper. I'll give it a try after work

Thank you mate

Finally I managed to rework my personal portfolio page with processwire and bootstrap. Nothing special, just some references and contact info. Maybe a blog will follow some day, atm I don't have the time to write blog entries even if I like to. Fully responsive, so you can check it out on your mobile too. Any feedback appreciated www.about-orkan.de (ah it's in german only, so you probably will get a error-500-face ^^)

Thanks for your answer ryan. Why I use Redirect/RedirectMatch instead of Rewrite is because it uses mod_alias, which does less logic and is therefor a tiny tiny tiny tiny tiny bit faster than mod_rewrite But it doesn't matter for this project, it's not a high traffic site anyway. I'm not sure, but this quote from the apache docs is maybe the explanation: I used relative path for the destination, so i guess my redirect from /contact to /legal redirects to http://domain.com/legal/contact which is a 404. Doesn't explain why i get the extra ?it parameter but well... enough explanation for me

Thanks arjen. I prefer the rewrites directly on apache because it's faster / uses less resources. A new version from PW is not a problem for me, because i don't touch the original. While deploying to live server, a shell script automatically prepends my additions to it. -- OFF-TOPIC -- Maybe someone is interested, i'll explain it in detail: - I store my projects in private git repositories on bitbucket.org - I have environment variables on my local machine, so I can differentiate in config.php which config is needed on runtime (like database, debug mode...). That allows me to have one version of the project which auomatically runs on live and development server - when i want to do a live or preview update, i just start a shell script which: 1. creates a fresh clone of the repository in a subdirectory like "live-2012-06-28" 2. if needed, it sets chown and chmod for caching directories 3. it compresses my javascript and css files with yui-compressor (my environment switch will print out minified in live and normal scripts in development) 4. it checks for htaccess-extra files, which then get prepended or appended to the original .htaccess (like redirects, expires) 5. the docroot is not a folder, it's a symlink. final step is to set the symlink to the new updated folder And thats it. I can make changes in my code without touching the PW core, push everything to git, and run the update script to deploy everything. If something goes wrong, I only have to reset the symlink to the previous version folder. A second script does nearly the same, but instead deploying live it depoys to a preview subdomain. So i can check my changes on the live server environment first, before destroying the live