On Sep 10, 2007, at 01:31:54, BuildSmart wrote:
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On Sep 10, 2007, at 01:01:19, Christian Stocker wrote:On 10.9.2007 6:53 Uhr, BuildSmart wrote:On Sep 10, 2007, at 24:41:47, Christian Stocker wrote:On 10.9.2007 3:53 Uhr, BuildSmart wrote:I was asked to look into the pdoru patch and extension by a client, thisis where I noticed that a similar patch is already applied to the rfc1867.c file(http://cvs.php.net/viewvc.cgi/php-src/main/rfc1867.c? r1=1.173.2.1&r2=1.173.2.1.2.1&pathrev=PHP_5_2&view=patch),is this patch compatible with the pdoru Upload Progress Meter extensionor do I need to write something from scratch?http://pecl.php.net/package/uploadprogressOk so after examination of this extension it looks like it's basicallyusing the same code with the exception of no support for memcachefromhttp://pdoru.from.ro/upload-progress-meter/upload-progress-meter- v4.1/upload_progress_meter/upload_progress_meter.cstatic int mmcache_loaded(void) { return 0; } static void * callback_mmcache( void *pointer, int read_bytes, int total_bytes, int what_happened ) { return NULL; } Doesn't look like much support for m(e)mcache :)Yes but that isn't the code I'm using, he started it and never completed it, the basic code is there but he never completed the memcache required functions, I've filled in the missing memcache code in the version I have, if it's not compatible with his basic extension then I either need to code from scratch or modify the pecl extension to support it and this is what I'm tyring to determine.and this is a requirement but my question was, is the provided patchcompatible with the pdoru extension (the pdoru extension works properl)?I don't know exactly, but I don't think, that it's 100% compatible, there was a reason for the PECL extension...Due to your remark I built it using my version of the pdoru code and ran a few simple tests, you are correct it's not 100% compatible, it has arbitrary issues validating the identifier and thus believes the identifier is invalid which in turn cancels the upload, only 1 out of 7 succeeded.I tested in file mode only and saw no need to test the memcache code due to the high failure rate.In light of this I think my best bet is to take the pecl extension code and add my memcache routines to satisfy my client.Now since the code is somewhat similar in design with some minor changes, mainly flow-control, I'm wondering how stable the extension is but since it's in pecl one might conclude that it's fairly stable/dependable or is this a poor assumption?Thanks for pointing me in the right direction, I'll run some cursory tests with the pecl extension before going any further.chregu- -- Dale
hmm, didn't this open up a big can of worms while I wasn't paying attention!!!.
I installed the turk memcache (mmcache) software a while back and while I hadn't tested the pdoru extension using it (or anything else for that matter) I thought it was working properly but it seems that it's not really compatible because a lot of issues are popping up quickly where things like phpgacl and phpbb2 become almost non- useable and simply disabling the software gets things running properly again.
Perhaps I need to rethink the customers needs and go with the file convention if they want the feature otherwise it doesn't make logical sense to enable an additional feature that negatively affects everyone else.
I took a quick look at Eaccelerator as a possible solution but haven't installed it because it comes with it's own set of compatibility issues, is there any other caching mechanism that is easy to implement and more compatible with the newer version of PHP or is this one of those things that should be used as implemented since no real speed benefits are achieved using such additional software that isn't already a consideration in the newer PHP distributions?
-- Dale
PGP.sig
Description: This is a digitally signed message part