Development Discussion Board

I've been delving into the configuration on 3 different shared hosting setups to get a little better understanding of their performance characteristics. As I mentioned above one of the key architectural drivers is that they need to enforce UID based access control to ensure that individual accounts ...

OK, I have been pondering this approach further. The new unifile ACP module is simply another ACP module which anyone is free to try as a performance enhancement (the same way acm_memcache, etc.), and there's no reason in principle why, once reviewed, it can't be included as an option in the 3.1 ...

nn- , perhaps you might want to look over the next version of this ACM cache module. I've called it acm_unifile.php. Note that I've moved it to the new 3.1 coding standard. It also has proper lock control. I did a bench test where I cranked the reads and writes to the cache by a factor of over ...

This post follows up and documents a PM dialogue that nn- and I had for the benefit of the wider phpBB developer / maintainer community. This may seem like a 101 lesson and if so then I apologise, but in my experience at best 1 in 10 application developers are aware of the issues and techniques ...

The inline documentation tags clearly have a syntax. Which documentation tool does phpBB use/proposing to use? It's just that if we want to encourage inline documentation, then it would be nice to know which syntax to conform to -- e.g. phpDocumentor , Doxygen , or ???

[edit] Having had a scan, it ...

Yes, I agree that compression costs, but so does opening and reading (or writing files). The former involves 1 file read rather than N (typically 12 or so), as I said this earlier in the topic: Re the CPU overhead issues of compressing, I discussed this in depth in one of the articles that I ...

It would probably need to be called acm_singlefile because as far as I can see the new autoloader naming convention reserves underscore for directory hierarchy.

I assumed that my description "beta code to show the sort of approach adopted here" would imply that I was submitting it for general ...

nn-, yes, yes and yes, and I am aware of all this. However, the flock and *nix move trick only ensure write integrity and do not address the coherency issue that you raised in your previous post which was about transactional integrity of the <read, (process), write> sequence, and problems caused by ...

The trigger for this post was a point made by nn- on another thread in regard of a new ACM module that I am proposing: I think there is one problem. On a moderately busy board the removal of data from cache may not work correctly since concurrent processes can put the old data back. If this affects ...

Thanks and I see that you've already moved a lot of this into the coding standards. I am not trying to cause work here. It's just that if I were going to rewrite / refactor / replace the templating engine, then I would regard such a document as a key control document and statement of the invariant ...