EPTThanks man !
Sc2gears 14.3.3: Replay version 2.1 support - Page 146
| Forum Index > SC2 General |
|
NexCa
Germany954 Posts
Thanks man ! | ||
|
Dakota_Fanning
Hungary2335 Posts
On August 07 2012 03:31 JustPlay wrote: Not a whole lot changed internally. If you don't already, you should find pieces of code that reference addresses you use and then automate address finding in newer versions. It would take a few hours tops to implement if you already know the addresses for a given version. Then when a new version comes out SC2gears can update ten minutes later instead of ten days. edit: something like this, and then you can call it like: BYTE* youraddress=FindPattern((PBYTE)"\xFF\x15\x00\x00\x00\x00\x83\xF8\xFF\x74\x00\x85\xC0\x7E\x00", "xx????xxxx?xxx?"); edit2: And to be clear, this shouldn't be included in sc2gears. It should just be a tool that you use to make updating it a whole lot faster and easier. I'm not talking about the unit id constants and ability code ids. Several fields in the replay have changed (in bit-length as new bits are added), map format changed a little, battle.net profile pages changed a lot (not just the structure but also the information available) etc. You can't automate these with some address finder. And to be clear: Sc2gears does not read SC2 memory, Sc2gears doesn't require SC2 to be running or even being installed. | ||
|
peshay
Germany14 Posts
| ||
|
Dakota_Fanning
Hungary2335 Posts
On August 07 2012 22:32 peshay wrote: Will Donation help to fix it quicker? :D It never hurts...  STATUS UPDATE: The new Sc2gears is near to completion. Still left a few things to do before release. The new release will come with a new plugin API (3.0) which contains backward incompatible changes. So expect all plugins (not written by me) to stop working when the new release hits. Sorry. The good news is that developers will be able to update most plugins in just 1-2 minutes to work with the new release (depends on whether they use the changed incompatible interfaces). I had/have to rewrite many things. New features and (cool) improvements are also coming with the new release... | ||
|
theonlyepi
United States25 Posts
| ||
|
Taru
France88 Posts
STATUS UPDATE: The new Sc2gears is near to completion. Still left a few things to do before release. The new release will come with a new plugin API (3.0) which contains backward incompatible changes. So expect all plugins (not written by me) to stop working when the new release hits. Sorry. The good news is that developers will be able to update most plugins in just 1-2 minutes to work with the new release (depends on whether they use the changed incompatible interfaces). I had/have to rewrite many things. New features and (cool) improvements are also coming with the new release... This is awsome ! Thank you  | ||
|
coL.hendralisk
Zimbabwe1756 Posts
| ||
|
Mr. Nefarious
United States515 Posts
On August 07 2012 23:58 Dakota_Fanning wrote: It never hurts... STATUS UPDATE: The new Sc2gears is near to completion. Still left a few things to do before release. The new release will come with a new plugin API (3.0) which contains backward incompatible changes. So expect all plugins (not written by me) to stop working when the new release hits. Sorry. The good news is that developers will be able to update most plugins in just 1-2 minutes to work with the new release (depends on whether they use the changed incompatible interfaces). I had/have to rewrite many things. New features and (cool) improvements are also coming with the new release... Sickest baller ever. I always knew I loved SCII Gears but when it stopped working, I was ashen. My laddering life has just become wandering through the dark! I love this program. Can't wait to see 1.5, thanks for all your hard work! | ||
|
pali_
Germany48 Posts
I have one question regarding a feature / functionality. It is highly probable that I'm missing something, sorry if I do. Is there a way to calculate the effectivness of SCV production during a given period of time in the current version of SC2gears? When you're playing a macro game as a Terran player, you want constant SCV production for the first 12-13 minutes (gametime) of the game. I would really find it an interesting feature, it would give me the possibility to find out how many SCVs I missed. It would be something similiar like the chart showing larvae inject effectivness for zerg. The replay gives you information regarding the number of CCs/OCs and when they finished etc., plus the amount of SCVs at a certain point in the game (also the amount of SCVs lost). Given the building time of a SCV, it should be doable to calculate the effectivness of SCV production? | ||
|
graNite
Germany4434 Posts
These numbers are calculated by the game when it plays the replay. The replay only contains actions like build scv or attack the scv, not whether it is killed or not. | ||
|
pali_
Germany48 Posts
/edit: hmm, after thinking about it for a little while, the "Train SCV" action should be sufficient? To keep it simple: Terran has one main building. The first 10 game minutes are 600 seconds, minus the time where the main is morphing (35sec), giving the player 565sec to build SCVs. Given the 17sec building time of a SCV, the Terran player should cast the "Train SCV" command 565 / 17 = 33 times. More mains make it more complicated, but it is still doable. The replay contains information about when a main started to build, e.g. "3:18 pali Build Command Center; target: x=29.5,y=79.5". Adding 100 sec + morphig time (35sec for OC / 50sec for planetary), you can calculate the time remaining to build SCVs during the first 10 game minutes. In this example, the CC (started at 3:18) would finish as OC at 5:33 (100sec build time + 35sec morph time), thus giving the player 267 seconds to build SCVs during the first 10 game minutes (600sec - 333sec = 267sec). So this main should use 267sec / 17sec = 15 "Train SCV" commands. So, in this case, 48 "Train SCV" commands should be casted by the player (33 for the main and 15 for the natural). Counting the "Train SCV" commands actually casted should be no problem. So for example, if 40 "Train SCV" commands were casted and the calculated maximum "Train SCV" commands were 48, the efficiency would be 40 / 48 = 83%. So by counting the "Train SCV" commands and using timestamps for "Build Command Center", this should actually be doable. | ||
|
DarkDoudou
France1 Post
Thank you for your work. you're awesome. | ||
|
Sartoe
1 Post
On August 09 2012 19:12 pali_ wrote: Ah ok, didn't know that. If this is true, then my proposed feature is not doable of course .. thanks! /edit: hmm, after thinking about it for a little while, the "Train SCV" action should be sufficient? To keep it simple: Terran has one main building. The first 10 game minutes are 600 seconds, minus the time where the main is morphing (35sec), giving the player 565sec to build SCVs. Given the 17sec building time of a SCV, the Terran player should cast the "Train SCV" command 565 / 17 = 33 times. More mains make it more complicated, but it is still doable. The replay contains information about when a main started to build, e.g. "3:18 pali Build Command Center; target: x=29.5,y=79.5". Adding 100 sec + morphig time (35sec for OC / 50sec for planetary), you can calculate the time remaining to build SCVs during the first 10 game minutes. In this example, the CC (started at 3:18) would finish as OC at 5:33 (100sec build time + 35sec morph time), thus giving the player 267 seconds to build SCVs during the first 10 game minutes (600sec - 333sec = 267sec). So this main should use 267sec / 17sec = 15 "Train SCV" commands. So, in this case, 48 "Train SCV" commands should be casted by the player (33 for the main and 15 for the natural). Counting the "Train SCV" commands actually casted should be no problem. So for example, if 40 "Train SCV" commands were casted and the calculated maximum "Train SCV" commands were 48, the efficiency would be 40 / 48 = 83%. So by counting the "Train SCV" commands and using timestamps for "Build Command Center", this should actually be doable. Sure that'd be sufficient. But let's say you're at the start of the game. You send your six scvs to mine, and then click on your CC and smash the "build scv" key twice. The replay data only shows that you hit the "build scv" key twice, but the replay does NOT show you how many scvs were actually being produced in this time-- in this case, one. Just looking at one of my replays and guestimating, I hit "build probe" about 500 times per macro oriented game. There's no way I built 500 probes, and there's not adequate information in the replay file to figure out which "probe creations" were legitimate. | ||
|
pali_
Germany48 Posts
I think the spam could be filtered out because you know that you can't actually build more than one scv per main within 17 seconds (SCV building time). The bigger issue here would be the scenario where you want to build a SCV, hit the build SCV button, have no SCV currently building but have no money to build one. I don't think you can extract the current mineral bank count from the replay? But you're probably right, there's probably no reliable way to calculate that. | ||
|
Dakota_Fanning
Hungary2335 Posts
On August 10 2012 04:35 pali_ wrote: I just checked a replay and you're right. Hm. I think the spam could be filtered out because you know that you can't actually build more than one scv per main within 17 seconds (SCV building time). The bigger issue here would be the scenario where you want to build a SCV, hit the build SCV button, have no SCV currently building but have no money to build one. I don't think you can extract the current mineral bank count from the replay? But you're probably right, there's probably no reliable way to calculate that. If you hit "Train SCV" twice, in general SC2 will only register 1 Train SCV because you don't have money for 2. SC2 will only register 2 SCVs if you "spam" those 2 SCVs really fast. The next release of Sc2gears will have the ability to detect and filter out most of these. | ||
|
pali_
Germany48 Posts
Really looking forward to the next version!! Do you have an approximate ETA? Like, this week, next week, in 2 weeks, etc.? Thanks for your effort! | ||
|
Dakota_Fanning
Hungary2335 Posts
On August 10 2012 05:39 pali_ wrote: Ah, ok, thanks for the response!! Really looking forward to the next version!! Do you have an approximate ETA? Like, this week, next week, in 2 weeks, etc.? Thanks for your effort! Approximate ETA is hours. | ||
|
Malpractice.248
United States734 Posts
Approximate love for you is huge. | ||
|
pali_
Germany48 Posts
| ||
|
felisconcolori
United States6168 Posts
This will be coming Soon (tm). But seriously, thank you for all your work. You deserve honorary mention in the halls of StarCraft 2 Fame. Can't wait to hear Smix's sweet voice saying "update available". I hope they don't mess too much more with things going on into HotS. | ||
| ||
It's become such an integral tool that I didn't realize how much I relied on it until it stopped working.
