SharkX 1.7 beta 1

[taz]

Hi my friends  ,

After gathering more knowledge regarding slot control (via zBOOM), it's time to fix a long lasting problem   with my mods - Slot control (mainly slot focus).

When issue is resolved   I'll change slot control code within AcKroNiC to the outcome (current code within AcKroNiC is partial functioning merge of zz and WiZ code, which is not really that good merge).


Please do not post this beta on official emule project site - since
new slot focus code might need some changes ...

Change log :

Code: [Select]

SharkX v1.7B1 - 03/11/09

based on AcKroNiC v5.4B1 (emule 0.49c)

change : Slot control code - mainly slot focus (taz)

Complete change log from SharkX v1.6 (i.e. showing changes in AcKroNiC
v5.4B1 code as well) :

Code: [Select]

SharkX v1.7B1 - 03/11/09

fix : memory leak @CSearchList::ProcessSearchAnswer (gureedo)
update : ECR + Emulate others
change : transparency for FakeAnalyzer (10'x Eulero)
fix : Don't Reset Connection Settings (Stulle)
change : Slot control code - mainly slot focus (taz)

For beta testers :

please note that "true" slot focus should kick in only when upload
limit is above 12 KBytes/sec.

I'm looking for slot focus behavioral changes with and without USS
(upload speed sense) and with and without safe connect (under extended
settings).
The most important thing to try to figure is whether the "drops" in
upload speed still lasts for a while (more than 1 minute) or whether
issue is solved - this won't show on clipped stats but rather on
upload graph at stats.

 

[Raid]

lets test

[taz]

Vast majority of UL drops (now even when occuring, which is quite rare, are shorter according to Israelli beta testers - however here our UL is quite limited  ) are related to [XL build ...] clients - an alike issue is also reported on official site by MorphXT users / testers.

Via a simple twick of CA - most of those [XL build ...] clients are being banned due to successive failing DL sessions, and I'm now on initial testings of a fix (if proven useful should also be applied to official since since is missing code ...) which should cause some of the very few [XL build ...] that escaped CA trigger to be caught.

Next beta (2) will probably be out next week - however since the actuall UL code is not altered collected data by beta 1 will remain valid.