Jump to content
Latest News: (loading..)


  • Content count

  • Joined

  • Last visited

  • Days Won


BrockleyJohn last won the day on March 11

BrockleyJohn had the most liked content!


Profile Information

  1. @wHiTeHaT Henry, I'm not saying I'm against classes now (yet!) and I share your concern over database hits. I do take the general point. Modular implementation needs global thinking.
  2. For Final, my focus is stuff that is easy and further enables addons without changing core code (although if it's Final, what actually matters on this is ease of installation) ie. - modularity as already proposed - extra getContents etc Plus the techy stuff that's largely covered. Is BS4 too big - or pointless without a considerable rewrite to use it? I haven't looked at Henry's class stuff yet. On the whole classes are a good thing and lend themselves to adapting but it sounds more like opening a can rather than closing the lid on one. I'd like more time to think that through rather than try and do it in a rush.
  3. @burt good then we're on the same page () I'll get cracking on it...
  4. For the approach to index_products: I can't see why the standard modules wouldn't be analogous to nested, ie. cm_ip_title cm_ip_description cm_ip_product_listing A debate could be had around how much to put in the modules and how much to leave in index. My first thought was to leave the select-building in index and basically modularise the output but I've come around to more or less the opposite - title module gets the category/manufacturer and outputs title, description module outputs description listing module does all the list query building and sort stuff and inclusion of product listing module This leaves it open to a store-owner to add extra content modules anywhere below the header in the page without changing core code.
  5. 1c. on it... I think Rainer's modules cover the last 7 bullets but they need bringing up to date with Edge - they're fairly gold.
  6. A canonical link is used to give a single name to a page that may be accessed via different paths or with different parameters, eg. products can be in more than one category so there are different paths to the same product. The canonical link gives a single name for that page so it doesn't look like duplicate content to search engines. Categories can only be in one category, and the views of them don't have paging, so those aspects of index.php don't need a canonical link.
  7. @raiwa It lives on... https://github.com/BrockleyJohn/Responsive-osCommerce/tree/Modular-Product-Info
  8. I have a couple of clients live on variants of raiwa's modular product_info which you thought was too granular. Would you like that rolling up into something more suitable - or are you just thinking along the lines of an extra hook or two? On the subject of which, if they don't come under Everything Else: extra hooks aren't something needing a lot of testing; I'd really like to get index_products in there too if I can - I could do that alongside finishing things in index.php if you're amenable. There are probably other hooks it would be good to get in there before it's set in stone, shopping_cart springs to mind but there may be others.
  9. TLS1.2 again

    @douglaswalker yes and no. I also think I missed something in your posts before. If you get a fail and a pass from the Paypal app, that means your server supports TLS1.2 but doesn't use it by default so it must be forced. There is now a setting in the app to force TLS1.2 which presumably you have set. Your independent code does not force TLS1.2 so fails. If you add the option curl_setopt($ch, CURLOPT_SSL_CIPHER_LIST, 'TLSv1'); before your independent test curl_exec it should succeed. What I have learnt further is: the urls in Paypal sandbox and tlstest are not identical despite what pp say. I ran them on https://www.ssllabs.com/ssltest/ and found that the cipher suites available are different and their priorities are different - with more secure ciphers available on tlstest and with higher priority than the insecure ones, while there are fewer secure on sandbox and there are insecure ones with higher priority. However, I think this is more likely to lead to successful calls against tlstest that fail against sandbox than the other way round. an empty call to tlstest.paypal.com only tests if TLS1.2 is available by default or on demand. It's not a sufficient test of the whole secure handshake. Specifically, it has no data so it only does the first bit of the handshake where they agree on what protocol they are speaking. It doesn't get as far as agreeing the encryption and understanding what each other are saying. I tried adding extra data to the call to tlstest but that alone doesn't make it fail (perhaps the other end needs to try and read it?). so the tests in the app go as far as possible with the tools that Paypal have offered to check whether the app can function in that environment but they don't prove it can - only the real calls can do that Hosting support for the site I've been working on eventually responded that the server does not have ciphers available for TLS1.2 so we're working to move it. This stuff is all to do with how the server is set up, what is available and in what order of priority so it's not at all surprising that your code gets different results on different servers.
  10. PayPal Standard not completing once returned to site

    Use the paypal app - you can still use express or standard in the app Go through the settings carefully in the documentation and make sure you get them right. Particularly (in standard) the setting of the return url and PDT on.
  11. [Addon} Modular Front Page

    @ce7 Hi Lyn, I think that's your admin index.php not your catalog one.
  12. That used to work. Enforcement may differ depending on your country but in in several countries google will reject products that don't have a valid GTIN and not show shopping ads for them. This is being rolled out worldwide. The only exceptions I'm aware of are things you make yourself, where you can get a global exemption. Spare parts are quoted as an example in adwords help of products where a valid gtin may not exist but in practice they get disapproved despite this. Anyway, you can change the new code in the execute function of the header_tags module to pull different data and put it into the gtin array and see if it goes. I doubt very much if product reviews will work on anything other than a real GTIN - think about it - how could they? These are global reviews on Google shopping not reviews on your site.
  13. OK folks - version 1.1 is now in the apps area. I've added an option to send products for review and a check on whether your database has the products_gtin field in it. One changed file, which is now self-updating once you've copied it in so no messing about un/re-installing.
  14. so many addons, so little time...
  15. @Roaddoctor I'll put in an option to add the products for review. I'm not sure it's such a great idea using them, though. It might distract people from writing the merchant review which is what you really want.