Menu
▾
▴
glestae-devel — Technical discussion of the engine
You can subscribe to this list here.
| 2009 |
Jan
|
Feb
|
Mar
|
Apr
|
May
|
Jun
|
Jul
|
Aug
(11) |
Sep
(4) |
Oct
|
Nov
(4) |
Dec
(15) |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 2010 |
Jan
(4) |
Feb
(25) |
Mar
(11) |
Apr
(1) |
May
(11) |
Jun
(13) |
Jul
(12) |
Aug
(4) |
Sep
(10) |
Oct
(7) |
Nov
(2) |
Dec
(26) |
| 2011 |
Jan
(18) |
Feb
(22) |
Mar
(10) |
Apr
(8) |
May
(13) |
Jun
(10) |
Jul
(1) |
Aug
(2) |
Sep
(3) |
Oct
(37) |
Nov
(19) |
Dec
(3) |
| 2012 |
Jan
|
Feb
|
Mar
|
Apr
|
May
|
Jun
(1) |
Jul
(1) |
Aug
|
Sep
|
Oct
|
Nov
|
Dec
(1) |
| 2014 |
Jan
|
Feb
|
Mar
|
Apr
|
May
|
Jun
|
Jul
|
Aug
(2) |
Sep
|
Oct
|
Nov
|
Dec
|
| 2015 |
Jan
|
Feb
|
Mar
|
Apr
|
May
(7) |
Jun
|
Jul
|
Aug
|
Sep
|
Oct
|
Nov
|
Dec
|
|
From: Nathan T. <hai...@us...> - 2015-05-22 11:54:48
|
Hello, I managed to get it running. I didn't even need to recompile lua or ogg/vorbis. VS2013 adds the C99 functions that were missing so I just needed to comment out the function aliases that VS 2013 supplies http://blogs.msdn.com/b/vcblog/archive/2013/07/19/c99-library-support-in-visual-studio-2013.aspx Also there's a problem with the compile events code. "cmd.exe" exited with code 1. I've just removed them completely from the project files for now and manually created version.cpp. I also needed to disable "Image Has Safe Exception Handlers" to (/SAFESEH:NO) in linker. Once I dealt with those issues it compiles and runs. Cheers, Nathan On 22 May 2015 at 07:34, Nathan Turner <hai...@us...> wrote: > Hi James, > > Yeah, I forgot to change the sending account because Gmail hides it until > clicking on the 'To' bar and defaults to the gmail one. > > I'm not a fan of Javascript. Fortunately, most of the time I only write > small snippets of it using jquery. TypeScript looks quite good, though, > from what I've seen of it. Hopefully I don't have to write much PHP in the > future. > > I'll give recompiling those a go. I'm on the network-2015 branch. I think > VS 2013 is the current one but there's 2015 coming soon. I'll try doing a > project conversion from VS 2008 from my laptop too, there might be some > flags missing that the newer one needs. One of the syntax errors seemed to > be specifically about log2 while the surrounding code with same format had > no issues. > > Cheers, > Nathan > > On 22 May 2015 at 06:39, James McCulloch <si...@gm...> wrote: > >> Hey Nathan, >> I assume this bounced because you sent from gmail and subscribe to the >> list through sf.net, added your gmail to the owners list, not sure if >> that will do... >> >> C# and Objective-C are perfectly sane, Javascript is weakly typed and >> can be a pain (depending on who wrote it!), but it's got nothing on PHP... >> It's commonly joked that the PHP language designers must be on drugs, after >> a while it seems less and less likely to be a joke, and certainly isn't >> very funny! >> >> I needed to rebuild Lua & Ogg/Vorbis for Win8.1 with VS 2012 Express for >> Desktop. Didn't know there was a newer one, is 2013 V12 or just V11.X ? >> >> I made some minor fixes for a missing Lua macro and a type ambiguity >> problem with int32, but got no major compile issues. Those fixes went into >> a branch named windows-fixage which has been merged to networking-2015 but >> not master. What branch are you on? >> >> Nothing obvious occurs to me as to why the compile errors, the enet lib >> is only configured for Debug and Release, so if trying RelWIthDebInfo it >> will want to link with the DLL runtime and you'll get a bunch of link >> errors. >> >> Have you got something funny in 'Include Directories' (the global one, >> not 'Additional...') ? Obviously you shouldn't be getting syntax errors in >> standard headers, something must be configured wrong. >> >> Not much help, sorry. >> >> Cheers, >> - James >> >> ---------- Forwarded message ---------- >> From: Nathan Turner <hai...@gm...> >> To: gle...@li... >> Cc: >> Date: Thu, 21 May 2015 22:32:16 +1000 >> Subject: Re: [Glestae-devel] Back in business & code now on github >> Hi, >> >> I normally use either C#, Objective-C or Javascript at work so C++ isn't >> as sane for me. It's been a while since I've done a lot of C++ coding, >> although I did a small amount when experimenting with Haxe's cpp target and >> bindings. >> >> Which libs did you need to rebuild? I can get the Visual Studio 2013 >> solution generated in CMake 3.2.2 after adding the GAE_WINDEPS and >> FREETYPE_LIB keys but then I get lots of syntax errors in stdlib.h and >> math.h, etc when trying to compile in VS Community 2013 on Windows 8.1. >> >> >> Cheers, >> Nathan >> >> >> >> >> On 19 May 2015 at 06:10, James McCulloch <si...@gm...> wrote: >> >>> Hi Guys, great to hear your both still about. >>> >>> As usual, no need to worry about how much time you can spare, it is what >>> it is. I think I’ll be sticking about long term this time, with spells of >>> limited activity no doubt, but for now, after a lot of PHP recently, I am >>> finding the sanity of C++ very pleasant! >>> >>> Re: Githhub, just seems like the place to be, nicer feeds and >>> whatnot... we can till use sf for file hosting if needs be. Organisations >>> are just like users really, but they can of course contain users, thus >>> there is no need for all the pull-request stuff you usually get (at least >>> for your trusted members! Others can fork and submit pull requests as >>> normal). >>> >>> I compiled with vc11 on my ultrabook and had to rebuild some libs, made >>> a small change because Lua5.3 seems to have cut down on the number a macros >>> it ships with, everything else was fairly smooth, but I didn’t build the >>> tools... >>> >>> Cheers, >>> - James >>> >>> *From:* Nathan Turner <hai...@us...> >>> *Sent:* Monday, May 18, 2015 9:32 PM >>> *To:* gle...@li... >>> *Subject:* Re: [Glestae-devel] Back in business & code now on github >>> >>> Hello James and Frank. It's good to hear from you both. My username on >>> Github is enturn. >>> >>> Yeah, I never did get the multiplayer to work as well as I would like. I >>> have my old laptop and desktop here so I can help test over a network. I >>> actually have a decent computer for compiling now. Yay. >>> >>> I don't want to over promise but I can at least be an extra set of eyes. >>> I look forward to helping out. >>> >>> Cheers >>> Nathan >>> >>> On 18 May 2015 at 04:30, Frank Tetzel <te...@us...> >>> wrote: >>> >>>> Hi James, >>>> >>>> long time, no see. I haven't looked at the glest code in quite a while. >>>> We probably need to port some stuff to newer library versions, lua and >>>> wxwidgets most likely. >>>> >>>> So, we'll move to github? Fine by me. Megaglest is on github too, it >>>> seems. I guess, i can just push commits to the repo as i'm now part of >>>> the organization. I don't have any experience with organizations on >>>> github. I just try it out when i have something to contribute. >>>> >>>> Regards, >>>> Frank >>>> >>>> >>>> > Hi folks, >>>> > >>>> > I recently decided I needed a hobby, back to Glest it is. I’ve been >>>> > poking around in Nathan’s dedicated_server branch, after a while and >>>> > few attempted fixes for the TurnManager I decided it was never going >>>> > to work anyway (using ‘real-time’ turn management and command issue >>>> > was always going to lead to sync problems. So I branched from just >>>> > before the introduction of TurnManager (there being numerous good >>>> > fixes to that point) and cherry picked the rest of the fixes, the new >>>> > branch is networking-2015. >>>> > >>>> > Created a glestadv organisation on github, which is where said branch >>>> > can be found, I found and added Frank, but couldn’t find Mr >>>> > Hailstone. If you’re still out there Nathan, get me a github >>>> > username! Anyone else still out there wanting in, do likewise. >>>> > >>>> > Seems to be running ok while playing with two instances on one >>>> > machine, will try separate machines later, but it’ll be a little >>>> > while before I can test multi-platform. >>>> > >>>> > Cheers, >>>> > - James >>>> >>>> >>>> ------------------------------------------------------------------------------ >>>> One dashboard for servers and applications across Physical-Virtual-Cloud >>>> Widest out-of-the-box monitoring support with 50+ applications >>>> Performance metrics, stats and reports that give you Actionable Insights >>>> Deep dive visibility with transaction tracing using APM Insight. >>>> http://ad.doubleclick.net/ddm/clk/290420510;117567292;y >>>> _______________________________________________ >>>> Glestae-devel mailing list >>>> Gle...@li... >>>> https://lists.sourceforge.net/lists/listinfo/glestae-devel >>>> >>> >>> >>> ------------------------------ >>> >>> ------------------------------------------------------------------------------ >>> One dashboard for servers and applications across Physical-Virtual-Cloud >>> Widest out-of-the-box monitoring support with 50+ applications >>> Performance metrics, stats and reports that give you Actionable Insights >>> Deep dive visibility with transaction tracing using APM Insight. >>> http://ad.doubleclick.net/ddm/clk/290420510;117567292;y >>> >>> ------------------------------ >>> _______________________________________________ >>> Glestae-devel mailing list >>> Gle...@li... >>> https://lists.sourceforge.net/lists/listinfo/glestae-devel >>> >>> >> >> >> ------------------------------------------------------------------------------ >> One dashboard for servers and applications across Physical-Virtual-Cloud >> Widest out-of-the-box monitoring support with 50+ applications >> Performance metrics, stats and reports that give you Actionable Insights >> Deep dive visibility with transaction tracing using APM Insight. >> http://ad.doubleclick.net/ddm/clk/290420510;117567292;y >> _______________________________________________ >> Glestae-devel mailing list >> Gle...@li... >> https://lists.sourceforge.net/lists/listinfo/glestae-devel >> >> > |
|
From: Nathan T. <hai...@us...> - 2015-05-21 21:35:21
|
Hi James, Yeah, I forgot to change the sending account because Gmail hides it until clicking on the 'To' bar and defaults to the gmail one. I'm not a fan of Javascript. Fortunately, most of the time I only write small snippets of it using jquery. TypeScript looks quite good, though, from what I've seen of it. Hopefully I don't have to write much PHP in the future. I'll give recompiling those a go. I'm on the network-2015 branch. I think VS 2013 is the current one but there's 2015 coming soon. I'll try doing a project conversion from VS 2008 from my laptop too, there might be some flags missing that the newer one needs. One of the syntax errors seemed to be specifically about log2 while the surrounding code with same format had no issues. Cheers, Nathan On 22 May 2015 at 06:39, James McCulloch <si...@gm...> wrote: > Hey Nathan, > I assume this bounced because you sent from gmail and subscribe to the > list through sf.net, added your gmail to the owners list, not sure if > that will do... > > C# and Objective-C are perfectly sane, Javascript is weakly typed and can > be a pain (depending on who wrote it!), but it's got nothing on PHP... It's > commonly joked that the PHP language designers must be on drugs, after a > while it seems less and less likely to be a joke, and certainly isn't very > funny! > > I needed to rebuild Lua & Ogg/Vorbis for Win8.1 with VS 2012 Express for > Desktop. Didn't know there was a newer one, is 2013 V12 or just V11.X ? > > I made some minor fixes for a missing Lua macro and a type ambiguity > problem with int32, but got no major compile issues. Those fixes went into > a branch named windows-fixage which has been merged to networking-2015 but > not master. What branch are you on? > > Nothing obvious occurs to me as to why the compile errors, the enet lib > is only configured for Debug and Release, so if trying RelWIthDebInfo it > will want to link with the DLL runtime and you'll get a bunch of link > errors. > > Have you got something funny in 'Include Directories' (the global one, > not 'Additional...') ? Obviously you shouldn't be getting syntax errors in > standard headers, something must be configured wrong. > > Not much help, sorry. > > Cheers, > - James > > ---------- Forwarded message ---------- > From: Nathan Turner <hai...@gm...> > To: gle...@li... > Cc: > Date: Thu, 21 May 2015 22:32:16 +1000 > Subject: Re: [Glestae-devel] Back in business & code now on github > Hi, > > I normally use either C#, Objective-C or Javascript at work so C++ isn't > as sane for me. It's been a while since I've done a lot of C++ coding, > although I did a small amount when experimenting with Haxe's cpp target and > bindings. > > Which libs did you need to rebuild? I can get the Visual Studio 2013 > solution generated in CMake 3.2.2 after adding the GAE_WINDEPS and > FREETYPE_LIB keys but then I get lots of syntax errors in stdlib.h and > math.h, etc when trying to compile in VS Community 2013 on Windows 8.1. > > > Cheers, > Nathan > > > > > On 19 May 2015 at 06:10, James McCulloch <si...@gm...> wrote: > >> Hi Guys, great to hear your both still about. >> >> As usual, no need to worry about how much time you can spare, it is what >> it is. I think I’ll be sticking about long term this time, with spells of >> limited activity no doubt, but for now, after a lot of PHP recently, I am >> finding the sanity of C++ very pleasant! >> >> Re: Githhub, just seems like the place to be, nicer feeds and whatnot... >> we can till use sf for file hosting if needs be. Organisations are just >> like users really, but they can of course contain users, thus there is no >> need for all the pull-request stuff you usually get (at least for your >> trusted members! Others can fork and submit pull requests as normal). >> >> I compiled with vc11 on my ultrabook and had to rebuild some libs, made >> a small change because Lua5.3 seems to have cut down on the number a macros >> it ships with, everything else was fairly smooth, but I didn’t build the >> tools... >> >> Cheers, >> - James >> >> *From:* Nathan Turner <hai...@us...> >> *Sent:* Monday, May 18, 2015 9:32 PM >> *To:* gle...@li... >> *Subject:* Re: [Glestae-devel] Back in business & code now on github >> >> Hello James and Frank. It's good to hear from you both. My username on >> Github is enturn. >> >> Yeah, I never did get the multiplayer to work as well as I would like. I >> have my old laptop and desktop here so I can help test over a network. I >> actually have a decent computer for compiling now. Yay. >> >> I don't want to over promise but I can at least be an extra set of eyes. >> I look forward to helping out. >> >> Cheers >> Nathan >> >> On 18 May 2015 at 04:30, Frank Tetzel <te...@us...> >> wrote: >> >>> Hi James, >>> >>> long time, no see. I haven't looked at the glest code in quite a while. >>> We probably need to port some stuff to newer library versions, lua and >>> wxwidgets most likely. >>> >>> So, we'll move to github? Fine by me. Megaglest is on github too, it >>> seems. I guess, i can just push commits to the repo as i'm now part of >>> the organization. I don't have any experience with organizations on >>> github. I just try it out when i have something to contribute. >>> >>> Regards, >>> Frank >>> >>> >>> > Hi folks, >>> > >>> > I recently decided I needed a hobby, back to Glest it is. I’ve been >>> > poking around in Nathan’s dedicated_server branch, after a while and >>> > few attempted fixes for the TurnManager I decided it was never going >>> > to work anyway (using ‘real-time’ turn management and command issue >>> > was always going to lead to sync problems. So I branched from just >>> > before the introduction of TurnManager (there being numerous good >>> > fixes to that point) and cherry picked the rest of the fixes, the new >>> > branch is networking-2015. >>> > >>> > Created a glestadv organisation on github, which is where said branch >>> > can be found, I found and added Frank, but couldn’t find Mr >>> > Hailstone. If you’re still out there Nathan, get me a github >>> > username! Anyone else still out there wanting in, do likewise. >>> > >>> > Seems to be running ok while playing with two instances on one >>> > machine, will try separate machines later, but it’ll be a little >>> > while before I can test multi-platform. >>> > >>> > Cheers, >>> > - James >>> >>> >>> ------------------------------------------------------------------------------ >>> One dashboard for servers and applications across Physical-Virtual-Cloud >>> Widest out-of-the-box monitoring support with 50+ applications >>> Performance metrics, stats and reports that give you Actionable Insights >>> Deep dive visibility with transaction tracing using APM Insight. >>> http://ad.doubleclick.net/ddm/clk/290420510;117567292;y >>> _______________________________________________ >>> Glestae-devel mailing list >>> Gle...@li... >>> https://lists.sourceforge.net/lists/listinfo/glestae-devel >>> >> >> >> ------------------------------ >> >> ------------------------------------------------------------------------------ >> One dashboard for servers and applications across Physical-Virtual-Cloud >> Widest out-of-the-box monitoring support with 50+ applications >> Performance metrics, stats and reports that give you Actionable Insights >> Deep dive visibility with transaction tracing using APM Insight. >> http://ad.doubleclick.net/ddm/clk/290420510;117567292;y >> >> ------------------------------ >> _______________________________________________ >> Glestae-devel mailing list >> Gle...@li... >> https://lists.sourceforge.net/lists/listinfo/glestae-devel >> >> > > > ------------------------------------------------------------------------------ > One dashboard for servers and applications across Physical-Virtual-Cloud > Widest out-of-the-box monitoring support with 50+ applications > Performance metrics, stats and reports that give you Actionable Insights > Deep dive visibility with transaction tracing using APM Insight. > http://ad.doubleclick.net/ddm/clk/290420510;117567292;y > _______________________________________________ > Glestae-devel mailing list > Gle...@li... > https://lists.sourceforge.net/lists/listinfo/glestae-devel > > |
|
From: James M. <si...@gm...> - 2015-05-21 20:39:33
|
Hey Nathan, I assume this bounced because you sent from gmail and subscribe to the list through sf.net, added your gmail to the owners list, not sure if that will do... C# and Objective-C are perfectly sane, Javascript is weakly typed and can be a pain (depending on who wrote it!), but it's got nothing on PHP... It's commonly joked that the PHP language designers must be on drugs, after a while it seems less and less likely to be a joke, and certainly isn't very funny! I needed to rebuild Lua & Ogg/Vorbis for Win8.1 with VS 2012 Express for Desktop. Didn't know there was a newer one, is 2013 V12 or just V11.X ? I made some minor fixes for a missing Lua macro and a type ambiguity problem with int32, but got no major compile issues. Those fixes went into a branch named windows-fixage which has been merged to networking-2015 but not master. What branch are you on? Nothing obvious occurs to me as to why the compile errors, the enet lib is only configured for Debug and Release, so if trying RelWIthDebInfo it will want to link with the DLL runtime and you'll get a bunch of link errors. Have you got something funny in 'Include Directories' (the global one, not 'Additional...') ? Obviously you shouldn't be getting syntax errors in standard headers, something must be configured wrong. Not much help, sorry. Cheers, - James ---------- Forwarded message ---------- From: Nathan Turner <hai...@gm...> To: gle...@li... Cc: Date: Thu, 21 May 2015 22:32:16 +1000 Subject: Re: [Glestae-devel] Back in business & code now on github Hi, I normally use either C#, Objective-C or Javascript at work so C++ isn't as sane for me. It's been a while since I've done a lot of C++ coding, although I did a small amount when experimenting with Haxe's cpp target and bindings. Which libs did you need to rebuild? I can get the Visual Studio 2013 solution generated in CMake 3.2.2 after adding the GAE_WINDEPS and FREETYPE_LIB keys but then I get lots of syntax errors in stdlib.h and math.h, etc when trying to compile in VS Community 2013 on Windows 8.1. Cheers, Nathan On 19 May 2015 at 06:10, James McCulloch <si...@gm...> wrote: > Hi Guys, great to hear your both still about. > > As usual, no need to worry about how much time you can spare, it is what > it is. I think I’ll be sticking about long term this time, with spells of > limited activity no doubt, but for now, after a lot of PHP recently, I am > finding the sanity of C++ very pleasant! > > Re: Githhub, just seems like the place to be, nicer feeds and whatnot... > we can till use sf for file hosting if needs be. Organisations are just > like users really, but they can of course contain users, thus there is no > need for all the pull-request stuff you usually get (at least for your > trusted members! Others can fork and submit pull requests as normal). > > I compiled with vc11 on my ultrabook and had to rebuild some libs, made a > small change because Lua5.3 seems to have cut down on the number a macros > it ships with, everything else was fairly smooth, but I didn’t build the > tools... > > Cheers, > - James > > *From:* Nathan Turner <hai...@us...> > *Sent:* Monday, May 18, 2015 9:32 PM > *To:* gle...@li... > *Subject:* Re: [Glestae-devel] Back in business & code now on github > > Hello James and Frank. It's good to hear from you both. My username on > Github is enturn. > > Yeah, I never did get the multiplayer to work as well as I would like. I > have my old laptop and desktop here so I can help test over a network. I > actually have a decent computer for compiling now. Yay. > > I don't want to over promise but I can at least be an extra set of eyes. I > look forward to helping out. > > Cheers > Nathan > > On 18 May 2015 at 04:30, Frank Tetzel <te...@us...> > wrote: > >> Hi James, >> >> long time, no see. I haven't looked at the glest code in quite a while. >> We probably need to port some stuff to newer library versions, lua and >> wxwidgets most likely. >> >> So, we'll move to github? Fine by me. Megaglest is on github too, it >> seems. I guess, i can just push commits to the repo as i'm now part of >> the organization. I don't have any experience with organizations on >> github. I just try it out when i have something to contribute. >> >> Regards, >> Frank >> >> >> > Hi folks, >> > >> > I recently decided I needed a hobby, back to Glest it is. I’ve been >> > poking around in Nathan’s dedicated_server branch, after a while and >> > few attempted fixes for the TurnManager I decided it was never going >> > to work anyway (using ‘real-time’ turn management and command issue >> > was always going to lead to sync problems. So I branched from just >> > before the introduction of TurnManager (there being numerous good >> > fixes to that point) and cherry picked the rest of the fixes, the new >> > branch is networking-2015. >> > >> > Created a glestadv organisation on github, which is where said branch >> > can be found, I found and added Frank, but couldn’t find Mr >> > Hailstone. If you’re still out there Nathan, get me a github >> > username! Anyone else still out there wanting in, do likewise. >> > >> > Seems to be running ok while playing with two instances on one >> > machine, will try separate machines later, but it’ll be a little >> > while before I can test multi-platform. >> > >> > Cheers, >> > - James >> >> >> ------------------------------------------------------------------------------ >> One dashboard for servers and applications across Physical-Virtual-Cloud >> Widest out-of-the-box monitoring support with 50+ applications >> Performance metrics, stats and reports that give you Actionable Insights >> Deep dive visibility with transaction tracing using APM Insight. >> http://ad.doubleclick.net/ddm/clk/290420510;117567292;y >> _______________________________________________ >> Glestae-devel mailing list >> Gle...@li... >> https://lists.sourceforge.net/lists/listinfo/glestae-devel >> > > > ------------------------------ > > ------------------------------------------------------------------------------ > One dashboard for servers and applications across Physical-Virtual-Cloud > Widest out-of-the-box monitoring support with 50+ applications > Performance metrics, stats and reports that give you Actionable Insights > Deep dive visibility with transaction tracing using APM Insight. > http://ad.doubleclick.net/ddm/clk/290420510;117567292;y > > ------------------------------ > _______________________________________________ > Glestae-devel mailing list > Gle...@li... > https://lists.sourceforge.net/lists/listinfo/glestae-devel > > |
|
From: James M. <si...@gm...> - 2015-05-18 20:10:57
|
Hi Guys, great to hear your both still about. As usual, no need to worry about how much time you can spare, it is what it is. I think I’ll be sticking about long term this time, with spells of limited activity no doubt, but for now, after a lot of PHP recently, I am finding the sanity of C++ very pleasant! Re: Githhub, just seems like the place to be, nicer feeds and whatnot... we can till use sf for file hosting if needs be. Organisations are just like users really, but they can of course contain users, thus there is no need for all the pull-request stuff you usually get (at least for your trusted members! Others can fork and submit pull requests as normal). I compiled with vc11 on my ultrabook and had to rebuild some libs, made a small change because Lua5.3 seems to have cut down on the number a macros it ships with, everything else was fairly smooth, but I didn’t build the tools... Cheers, - James From: Nathan Turner Sent: Monday, May 18, 2015 9:32 PM To: gle...@li... Subject: Re: [Glestae-devel] Back in business & code now on github Hello James and Frank. It's good to hear from you both. My username on Github is enturn. Yeah, I never did get the multiplayer to work as well as I would like. I have my old laptop and desktop here so I can help test over a network. I actually have a decent computer for compiling now. Yay. I don't want to over promise but I can at least be an extra set of eyes. I look forward to helping out. Cheers Nathan On 18 May 2015 at 04:30, Frank Tetzel <te...@us...> wrote: Hi James, long time, no see. I haven't looked at the glest code in quite a while. We probably need to port some stuff to newer library versions, lua and wxwidgets most likely. So, we'll move to github? Fine by me. Megaglest is on github too, it seems. I guess, i can just push commits to the repo as i'm now part of the organization. I don't have any experience with organizations on github. I just try it out when i have something to contribute. Regards, Frank > Hi folks, > > I recently decided I needed a hobby, back to Glest it is. I’ve been > poking around in Nathan’s dedicated_server branch, after a while and > few attempted fixes for the TurnManager I decided it was never going > to work anyway (using ‘real-time’ turn management and command issue > was always going to lead to sync problems. So I branched from just > before the introduction of TurnManager (there being numerous good > fixes to that point) and cherry picked the rest of the fixes, the new > branch is networking-2015. > > Created a glestadv organisation on github, which is where said branch > can be found, I found and added Frank, but couldn’t find Mr > Hailstone. If you’re still out there Nathan, get me a github > username! Anyone else still out there wanting in, do likewise. > > Seems to be running ok while playing with two instances on one > machine, will try separate machines later, but it’ll be a little > while before I can test multi-platform. > > Cheers, > - James ------------------------------------------------------------------------------ One dashboard for servers and applications across Physical-Virtual-Cloud Widest out-of-the-box monitoring support with 50+ applications Performance metrics, stats and reports that give you Actionable Insights Deep dive visibility with transaction tracing using APM Insight. http://ad.doubleclick.net/ddm/clk/290420510;117567292;y _______________________________________________ Glestae-devel mailing list Gle...@li... https://lists.sourceforge.net/lists/listinfo/glestae-devel -------------------------------------------------------------------------------- ------------------------------------------------------------------------------ One dashboard for servers and applications across Physical-Virtual-Cloud Widest out-of-the-box monitoring support with 50+ applications Performance metrics, stats and reports that give you Actionable Insights Deep dive visibility with transaction tracing using APM Insight. http://ad.doubleclick.net/ddm/clk/290420510;117567292;y -------------------------------------------------------------------------------- _______________________________________________ Glestae-devel mailing list Gle...@li... https://lists.sourceforge.net/lists/listinfo/glestae-devel |
|
From: Nathan T. <hai...@us...> - 2015-05-18 11:33:13
|
Hello James and Frank. It's good to hear from you both. My username on Github is enturn. Yeah, I never did get the multiplayer to work as well as I would like. I have my old laptop and desktop here so I can help test over a network. I actually have a decent computer for compiling now. Yay. I don't want to over promise but I can at least be an extra set of eyes. I look forward to helping out. Cheers Nathan On 18 May 2015 at 04:30, Frank Tetzel <te...@us...> wrote: > Hi James, > > long time, no see. I haven't looked at the glest code in quite a while. > We probably need to port some stuff to newer library versions, lua and > wxwidgets most likely. > > So, we'll move to github? Fine by me. Megaglest is on github too, it > seems. I guess, i can just push commits to the repo as i'm now part of > the organization. I don't have any experience with organizations on > github. I just try it out when i have something to contribute. > > Regards, > Frank > > > > Hi folks, > > > > I recently decided I needed a hobby, back to Glest it is. I’ve been > > poking around in Nathan’s dedicated_server branch, after a while and > > few attempted fixes for the TurnManager I decided it was never going > > to work anyway (using ‘real-time’ turn management and command issue > > was always going to lead to sync problems. So I branched from just > > before the introduction of TurnManager (there being numerous good > > fixes to that point) and cherry picked the rest of the fixes, the new > > branch is networking-2015. > > > > Created a glestadv organisation on github, which is where said branch > > can be found, I found and added Frank, but couldn’t find Mr > > Hailstone. If you’re still out there Nathan, get me a github > > username! Anyone else still out there wanting in, do likewise. > > > > Seems to be running ok while playing with two instances on one > > machine, will try separate machines later, but it’ll be a little > > while before I can test multi-platform. > > > > Cheers, > > - James > > > ------------------------------------------------------------------------------ > One dashboard for servers and applications across Physical-Virtual-Cloud > Widest out-of-the-box monitoring support with 50+ applications > Performance metrics, stats and reports that give you Actionable Insights > Deep dive visibility with transaction tracing using APM Insight. > http://ad.doubleclick.net/ddm/clk/290420510;117567292;y > _______________________________________________ > Glestae-devel mailing list > Gle...@li... > https://lists.sourceforge.net/lists/listinfo/glestae-devel > |
|
From: Frank T. <te...@us...> - 2015-05-17 18:48:23
|
Hi James, long time, no see. I haven't looked at the glest code in quite a while. We probably need to port some stuff to newer library versions, lua and wxwidgets most likely. So, we'll move to github? Fine by me. Megaglest is on github too, it seems. I guess, i can just push commits to the repo as i'm now part of the organization. I don't have any experience with organizations on github. I just try it out when i have something to contribute. Regards, Frank > Hi folks, > > I recently decided I needed a hobby, back to Glest it is. I’ve been > poking around in Nathan’s dedicated_server branch, after a while and > few attempted fixes for the TurnManager I decided it was never going > to work anyway (using ‘real-time’ turn management and command issue > was always going to lead to sync problems. So I branched from just > before the introduction of TurnManager (there being numerous good > fixes to that point) and cherry picked the rest of the fixes, the new > branch is networking-2015. > > Created a glestadv organisation on github, which is where said branch > can be found, I found and added Frank, but couldn’t find Mr > Hailstone. If you’re still out there Nathan, get me a github > username! Anyone else still out there wanting in, do likewise. > > Seems to be running ok while playing with two instances on one > machine, will try separate machines later, but it’ll be a little > while before I can test multi-platform. > > Cheers, > - James |
|
From: James M. <si...@gm...> - 2015-05-16 04:33:02
|
Hi folks, I recently decided I needed a hobby, back to Glest it is. I’ve been poking around in Nathan’s dedicated_server branch, after a while and few attempted fixes for the TurnManager I decided it was never going to work anyway (using ‘real-time’ turn management and command issue was always going to lead to sync problems. So I branched from just before the introduction of TurnManager (there being numerous good fixes to that point) and cherry picked the rest of the fixes, the new branch is networking-2015. Created a glestadv organisation on github, which is where said branch can be found, I found and added Frank, but couldn’t find Mr Hailstone. If you’re still out there Nathan, get me a github username! Anyone else still out there wanting in, do likewise. Seems to be running ok while playing with two instances on one machine, will try separate machines later, but it’ll be a little while before I can test multi-platform. Cheers, - James |
|
From: Tom R. <to...@me...> - 2014-08-19 22:44:30
|
gha...@ya... schrieb: > hi all, > i need help please > i downloaded the gae, and completed all the required steps to compile it on my win 32 machine with visual studio 2008, but when i run the glestadv solution , the build gives an error: " LINK : fetal error LNK1181 : cannot open input file 'lua_script.cpp.lib' ? > > what should i do? > i searched my pc, and only found a file named lua_script.c++, it doesn't build a lib, now what? > i need to compile the game urgently, so your help will be appreciated > thanks in advance Are you using Git or some previous version? Out of interest, what makes you need to compile it urgently? Tom |
|
From: <gha...@ya...> - 2014-08-19 19:28:01
|
hi all, i need help please i downloaded the gae, and completed all the required steps to compile it on my win 32 machine with visual studio 2008, but when i run the glestadv solution , the build gives an error: " LINK : fetal error LNK1181 : cannot open input file 'lua_script.cpp.lib' ? what should i do? i searched my pc, and only found a file named lua_script.c++, it doesn't build a lib, now what? i need to compile the game urgently, so your help will be appreciated thanks in advance Sent from Yahoo Mail on Android |
|
From: Austin P. <aus...@gm...> - 2012-12-07 03:09:34
|
Crash Version: Advanced Engine 0.3.93+git_d404c6b 0.3-1000-gd404c6b Built: Thu Dec 6 21:27:09 EST 2012 Time: Thu Dec 6 21:59:37 2012 Description: SIGABRT: Address: 0x3e800002e60 Backtrace: ./source/game/glestadv(_ZN6Shared8Platform24PlatformExceptionHandler7handlerEiP7siginfoPv+0x33c) [0x9c4e4e] /lib/libc.so.6(+0x33af0) [0x7f80ac931af0] /lib/libc.so.6(gsignal+0x35) [0x7f80ac931a75] /lib/libc.so.6(abort+0x180) [0x7f80ac9355c0] /lib/libc.so.6(__assert_fail+0xf1) [0x7f80ac92a941] ./source/game/glestadv() [0x6a3669] ./source/game/glestadv(_ZNK5Glest3Sim3Map7getTileERKN6Shared4Math4Vec2IiEE+0x2b) [0x6a3695] ./source/game/glestadv(_ZN5Glest8Graphics8Renderer25renderObjectsForSelectionEv+0x104) [0x77001e] ./source/game/glestadv(_ZN5Glest8Graphics8Renderer15computeSelectedERSt6vectorIPNS_8Entities4UnitESaIS5_EERPKNS3_9MapObjectERKN6Shared4Math4Vec2IiEESI_+0x255) [0x76b9c3] ./source/game/glestadv(_ZN5Glest3Gui13UserInterface9mouseMoveEii+0xce) [0x781bac] ./source/game/glestadv(_ZN5Glest3Gui9GameState9mouseMoveEiiRKN6Shared8Platform10MouseStateE+0x332) [0x731fd2] ./source/game/glestadv(_ZN5Glest4Main7Program9mouseMoveEN6Shared4Math4Vec2IiEE+0x58) [0x7b7fac] ./source/game/glestadv(_ZN5Glest7Widgets12WidgetWindow14eventMouseMoveEiiRKN6Shared8Platform10MouseStateE+0x10b) [0x903aa1] ./source/game/glestadv(_ZN6Shared8Platform6Window11handleEventEv+0x268) [0x9c3f4e] ./source/game/glestadv(_ZN5Glest4Main7Program4loopEv+0x4dc) [0x7b7d1c] ./source/game/glestadv(_ZN5Glest4Main9glestMainEiPPc+0x353) [0x7b3574] ./source/game/glestadv(main+0x72) [0x7b390e] /lib/libc.so.6(__libc_start_main+0xfd) [0x7f80ac91cc4d] ./source/game/glestadv() [0x68e149] ======================= |
|
From: Frank T. <te...@us...> - 2012-07-10 13:24:05
|
Hi, i just got an email from sourceforge. They discontinue their hosted apps service at the end of the year[1]. So, we have to migrate trac to some other place. We haven't used the other apps (CodeStriker, IdeaTorrent) much and i don't think we need them. Most of their migration documentation is still work in progress, the trac one too[2]. It looks like they suggest moving trac to the sourceforge webspace. Seems reasonable. We schould probably wait till they finish their migration docs before taking any action. Just a heads up so we don't forget it. Regards, Frank. [1] http://sourceforge.net/p/forge/community-docs/Hosted%20Apps%20Retirement/ [2] http://sourceforge.net/p/forge/community-docs/Migrating%20Trac%20from%20Hosted%20Apps/ |
|
From: Frank T. <te...@us...> - 2012-06-29 23:45:03
|
Hi all, i just played around with xml schemas the other day and got into hacking mood. So i generated a schema for glest's unit xmls and merged some documentation from the wiki into it. I just wanted to test auto-completion and documentation of xml files. Auto-completion works quite well in Eclipse (with web tools), Netbeans and Visual Studio. In Eclipse and VS one also gets nice documentation when hovering over a tag. Just open archmage.xml in the attachment and see for yourself (only the rood node was changed). VS shows the documentation twice. That's on purpose one could say as i wanted to get it to work in Netbeans too. One other interesting thing i tried was generating a xml reference out of the schema. "Similar" to the one discussed here: http://glest.org/glest_board/index.php?topic=6323.0 Just open unit.xsd in firefox and have a look. The transformation is done by xs3p.xsl which is clearly aimed at developers. I'm not convinced that this is better than the wiki. The only benefit: It's auto-generated from the source. If we force validation against the schema, the developer adding a new tag has to add it to the schema as well. And hopefully adding documentation too. This was all just a test i wanted to share with you. I don't know how much work it is to make a good xml schema which is meant for validation. This one was just infered from the xml data we currently have in git master. Of course all these xml files are valid but many wrong ones probably too (just look at the cluttered command definition). At least some decent auto-completion can be done with it. The used shell script is also included in the attachment (needs trang, tidy, xmlstarlet, wget). Regards, Frank. |
|
From: Frank T. <te...@us...> - 2011-12-30 14:11:17
|
Hi all, i came across this site the other day: http://pootle.unknown-horizons.org/ It's an online translation tool for the game unknown horizon. They use the program pootle[1]. There are similar online tools like transifex[2] and Rosetta which is the integrated translation tool of launchpad. I always thought that translating glest is quite a pain. There's no tool support. You have to do everything by hand: extracting strings from the source code and XML files, merge them into all the language files, ... I had a look at GNU gettext to get some standard tools. I pushed a new branch translate_tools where i added translation support to the map editor and g3d viewer. wxWidgets uses gettext for i18n. So i started with that one to test gettext. It automatically extracts new and changed strings from source files and merges them with existing translations. It even does fuzzy merges: If the string change is small it still uses the old translation but marks it as fuzzy, so the translator should have a look at it. It uses the language of your system (LANG environment variable on linux). I only tested it on linux but i hope it also works on windows when you install the gettext utilities[3]. To edit the translation in the po subdirectory use a po-editor like lokalize or poedit. A normal text editor does the job too. To add a new translation get in the po/program directory and run: $ msginit -l es -o es.po -i program.pot (example for spanish, replace es with your locale code) There's only a german translation yet, and naturally english in the source files. If we want to use gettext for the game too we should get away from the current keyword system. The translator should translate an english phrase or word not some obscure keyword. The only problem are unit names. The english "screen name" of a unit isn't saved in the XML file but in the english translation. We should get rid of an english translation and have the english strings directly in the code. Extracting strings from XML was a bit tricky too but i have a workaround using XSLT with xmlstarlet. So, what do you think about using standard translation tools like gettext and changing the keyword system? One big disadvantage of gettext: You have to "compile" translations into a binary format. If the XML stuff uses gettext too, we would need to help the modders with some kind of toolchain. Unicode is probably another problem we have to face. [1] http://translate.sourceforge.net/wiki/pootle/index [2] https://www.transifex.net/ [3] http://gnuwin32.sourceforge.net/packages/gettext.htm Regards, Frank. |
|
From: Nathan T. <hai...@us...> - 2011-12-19 08:35:39
|
Hello,
I think I found the cause of one sync error:
======== client ==============
> 616257: Frame: 3096 :: NetworkInterface::postCommandUpdate(): UnitID: 2,
> FactionIndex: 0, CurrSkillId: 70
> 616258: Frame: 3096 :: NetworkInterface::postCommandUpdate(): UnitID: 3,
> FactionIndex: 0, CurrSkillId: 70
> 616258: Frame: 3096 :: NetworkInterface::postCommandUpdate(): UnitID: 6,
> FactionIndex: 0, CurrSkillId: 0
> 616258: Frame: 3096 :: NetworkInterface::postCommandUpdate(): UnitID: 12,
> FactionIndex: 1, CurrSkillId: 70
> 616258: Frame: 3096 :: ClientInterface::checkCommandUpdate(): Sync Error:
> unit id: 12 unit type: worker, skill class: STOP
>
> ======= server ============
> 599145: Frame: 3096 :: NetworkInterface::postCommandUpdate(): UnitID: 2,
> FactionIndex: 0, CurrSkillId: 70
> 599145: Frame: 3096 :: NetworkInterface::postCommandUpdate(): UnitID: 3,
> FactionIndex: 0, CurrSkillId: 70
> 599145: Frame: 3096 :: NetworkInterface::postCommandUpdate(): UnitID: 6,
> FactionIndex: 0, CurrSkillId: 0
>
There's an extra postCommandUpdate called with the client which means
Unit::update would need to have returned true for the client and false for
server. So units are dying on the server before the client can execute the
commands that causes the death?
Also occasionally there will be a corrupted MoveSkillUpdate or
ProjectileUpdate on the client when changing window focus with left click
or something. I'm not entirely sure what causes it. Sometimes all the
values will be zero and others it will be more like below.
======== client =============
> 57465: Frame: 864 :: KeyFrame::getMoveUpdate(): Pos Offset:(0, -2) Frame
> Offset: -471
> 57465: Frame: 864 :: ClientInterface::updateMove(): Ids don't match.
> UnitIdLocal: 12 UnitIdRemote: 3312
>
> ======== server ==============
> 37826: Frame: 864 :: KeyFrame::addUpdate(MoveSkillUpdate): Pos Offset:(1,
> 0) Frame Offset: 22 UnitId: 12
> 37826: Frame: 864 :: NetworkInterface::postCommandUpdate():
> unit->getNextPos(): (45, 12)
>
Both errors appear with the auto commands commented out.
Nathan
On 26 November 2011 20:39, Nathan Turner
<hai...@us...>wrote:
> On 2 November 2011 06:33, James McCulloch <si...@gm...> wrote:
>
>> the problem
>> appears to be with auto-commands being generated at different times...
>> this shouldn't be possible of course, but that's what's happening.
>>
>
> Could it be the order things are done?
>
> In Program::loop:
> simulationInterface->asNetworkInterface()->update(); // client sends
> messages to server
>
> In SimulationInterface::updateWorld:
> world->processFrame();
> In Unit::doUpdateCommand:
> getCurrCommand()->getType()->update(this); // issues auto commands by
> updating unit commands (stop, move, repair, etc)
> g_simInterface.updateSkillCycle(this); // syncs moves
> frameProccessed(); // process commands from server (update keyframe)
>
> The auto commands would be using the client's unit positions rather than
> the server's synced ones.
>
> Also the map height is float instead of fixed. I'm not sure if that could
> cause issues setting the next pos with the pathfinder.
>
> Nathan
>
>
> On 24 November 2011 11:13, Nathan Turner <hai...@us...
> > wrote:
>
>> Hello,
>>
>> I said previously the network design didn't really fit how ENet works so
>> I've done a redesign which I think is a lot cleaner. It's split into host
>> and peer (session) with client and server specific host functions in child
>> classes. ClientInterface and ServerInterface handles the connect and
>> disconnect events while messages are managed by sessions.
>>
>> I ran a game with idle server player, idle networked player and an AI.
>> When the AI defeated one of the idle players it had a sync error.
>>
>> That's a good point about the structs Daniel. I imagine it would be the
>> same with classes.
>>
>> Nathan
>>
>>
>> On 2 November 2011 13:33, Daniel Santos <dan...@at...> wrote:
>>
>>>
>>> > @dansan,
>>> > The imposed random network delay will be nice for testing, and will
>>> > indeed be usable with eNet, much of the rest becomes redundant I
>>> > think. While your DataBuffer classes offers some nice functionality,
>>> > personally I'm perfectly content using memcpy() and assuming
>>> > little-endian architecture all-around.
>>> > If code that purports to solve byte order issues is added, it needs
>>> > to be tested first... and for big endian hardware I think this is
>>> > probably not likely.
>>> > As to Message structure/methodology, I'm in favour of whatever
>>> > works, for now what we have works, not as nice to use as it could be
>>> > perhaps, but it works.
>>> >
>>> > I'm currently in favour of "getting shit working". We can always
>>> > re-factor to make it look like it was well designed later ;-P
>>> >
>>> Yeah, but it's more than just endianness. There is the ABI as well.
>>> For instance, for the following structure:
>>>
>>> struct Jack {
>>> char bob;
>>> char cindy;
>>> int jill;
>>> };
>>>
>>> One compiler may lay it out like this in memory
>>>
>>> Jack fred; <-- 0x5000:
>>>
>>> 0x5000 bob
>>> 0x5001 cindy
>>> 0x5002 unused
>>> 0x5003 unused
>>> 0x5004 jill
>>> 0x5005 jill
>>> 0x5006 jill
>>> 0x5007 jill
>>>
>>> while another lays it out like this:
>>>
>>> 0x5000 bob
>>> 0x5001 cindy
>>> 0x5002 jill
>>> 0x5003 jill
>>> 0x5004 jill
>>> 0x5005 jill
>>>
>>> or even this:
>>>
>>> 0x5000 bob
>>> 0x5001 unused
>>> 0x5002 unused
>>> 0x5003 unused
>>> 0x5004 cindy
>>> 0x5005 unused
>>> 0x5006 unused
>>> 0x5007 unused
>>> 0x5008 jill
>>> 0x5009 jill
>>> 0x500a jill
>>> 0x500f jill
>>>
>>> The bottom line is that copying memory of a struct directly is never a
>>> good practice when you are leaving your machine and going across the
>>> network. I know that some of this may be solved with the pragma pack 1,
>>> but it doesn't solve all of the problems associated with it and it also
>>> inhibits the optimization possibilities of the compiler and even force
>>> slower operations (depending upon the CPU). Then, add to that the fact
>>> that NetworkString transmits the max string size in bytes all of the
>>> time, regardless of the size of the actual string. Also, don't forget
>>> that iPhone and many other phones use ARM processors, which are
>>> little-endian.
>>>
>>> Anyway, I don't remember all of the issues I unearthed of it when I was
>>> working on this before, but there are a lot of them.
>>>
>>> As for NetworkDataBuffer, it was fully tested when I was working on this
>>> in 2009. That wasn't the part that was broken -- it was the fact that I
>>> kept running off and breaking new things before I had finished fixing
>>> the last one! :)
>>>
>>> Daniel
>>>
>>>
>>> ------------------------------------------------------------------------------
>>> RSA® Conference 2012
>>> Save $700 by Nov 18
>>> Register now!
>>> http://p.sf.net/sfu/rsa-sfdev2dev1
>>> _______________________________________________
>>> Glestae-devel mailing list
>>> Gle...@li...
>>> https://lists.sourceforge.net/lists/listinfo/glestae-devel
>>>
>>
>>
>
|
|
From: Nathan T. <hai...@us...> - 2011-12-03 01:54:29
|
Hello Daniel, I'd like to reaffirm James' welcome. For myself I consider James to be the final word on any disagreements, although like he stated we tend to work things out in a fairly democratic manner. Cheers, Nathan On 18 November 2011 18:12, James McCulloch <si...@gm...> wrote: > > Hi all, > > Greetings Daniel, > > > I got interested in glest a while back, got busy on some other stuff, and > > recently came back to it. > > I'd really like to be able to contribute but I got some questions: > > 1- What lies in the future of GAE? I've been reading on some forum > threads > > that perhaps GAE and MegaGlest would merge into one entity. Is that the > > case? If so when? I'm asking because I can run Glest and GAE on my pc > > (albeit it's a bit slow), but MegaGlest runs to a crawl. So you'll > > understand that If I am to be soonish unable to run a game I want to work > > on, well... doesn't fare well... Besides a possible merge, where does GAE > > want to go with the tech/game? > > We are a bunch of programmers, and as such our focus is primarily on > enhancing the engine for modders, and of course adding nicities for > players. We have recently made minor modifications to magitech to make > use of some of the new features the engine supports, and will likely > do so again in the future, but in terms of extending magitech with new > units etc, we don't really have any plans. I'd personally like to make > a campaign for magitech some day, but that will take some time, and I > don't have a lot of that spare atm :( > > Plans are afoot to improve performance of various parts of the game, > and these changes involve using sensible data structures and > algorithms, rather than just caching everything in sight and wasting > memory and encouraging (hardware) cache misses, as is the case with > MG. We certainly do not intend to make it run slower, though with > active development going on such slow-downs may temporarily manifest > themselves. > > As Tom has mentioned, the merge is off, different development > philosophies and the contrasted views of certain stubborn team members > (mainly myself from GAE, and Titi from MG) have made that largely an > impossibility. > > > 2- i've got the source with TurtoiseGit and got it to compile/run and > all, > > do I really need to do all that key stuff to be able to push back unto > the > > server? I'm mainly asking so I don't start changing files and then need > to > > redo everything 'cause I wasn't set up properly. > > It will make life much easier if you do, I recently moved a local > repository from one computer to another and am still able to pull/push > with it, but tortoise complains about being unable to load my key and > then I have to enter my sourceforge username and password everytime, > which is a bit annoying. It may seem like a pain, but you only have > to do it once ;-) > > You need not worry about having to redo anything though, changes you > make now can still be committed to your local repo and patches > generated from that, which we can then apply and push (all the log > messages still show you as author, one of the many nice things about > git), and when granted repo access you can simply reset to "remote" to > the ssh url and you're ready to go, we can help out there when the > time comes. > > > 3- How are tasks assigned usually? How goes the dynamic between the team > > members? Is there a boss like person that can veto everything ans assigns > > everything? How are submissions revised? I'm trying to get a sense of how > > working with you guys will feel/look like. > > There is no boss like person who controls all, we generally operate in > a fairly democratic manner, but clashes concerning how things should > be done haven't really happened in my memory, and if you are willing > to implement something that is of benefit to modders and/or players > (or even just to make our own lives easier) then you are not likely to > hear any complaints, if there are complaints/concerns then it may need > to be made optional (if that makes sense for the feature/change in > question). > > Code review generally happens once its in the repo, and then probably > only if it causes unforeseen problems. I generally browse through the > changesets for most (if not all) commits, but this not much of a > review really, just trying to keep up with what people are doing. > > > Furthermore, I've been looking into the list of tickets for v0.5 and I > feel > > confident enough I can tackle some of the music stuff. They sound simple > > enough for introductory tasks: > > 187- non-faction music playlist > > 291 - music playlist per subfaction > > 292 - allow music playlist defined per unit > > These do indeed seem to be good candidates (although I'm unsure how > 292 is meant to work...), and we are happy to provide > assistance/explanations with/for anything that has you stumped, though > the 'turn around time' on that may not always be as prompt as you > would like ;-) > > We'd be delighted to see some patches and then promptly get you write > access to the repo, contributors are seemingly hard to come by, so > from myself and on behalf of Nathan (hailstone), Frank (Yggdrasil) and > the other Daniel (forum name daniel.santos, but I use his irc nick > 'dansan', because I like it), we hope to have you on board soon! > > Cheers, > - James (aka silnarm) > > > ------------------------------------------------------------------------------ > All the data continuously generated in your IT infrastructure > contains a definitive record of customers, application performance, > security threats, fraudulent activity, and more. Splunk takes this > data and makes sense of it. IT sense. And common sense. > http://p.sf.net/sfu/splunk-novd2d > _______________________________________________ > Glestae-devel mailing list > Gle...@li... > https://lists.sourceforge.net/lists/listinfo/glestae-devel > |
|
From: Nathan T. <hai...@us...> - 2011-11-26 10:40:05
|
On 2 November 2011 06:33, James McCulloch <si...@gm...> wrote:
> the problem
> appears to be with auto-commands being generated at different times...
> this shouldn't be possible of course, but that's what's happening.
>
Could it be the order things are done?
In Program::loop:
simulationInterface->asNetworkInterface()->update(); // client sends
messages to server
In SimulationInterface::updateWorld:
world->processFrame();
In Unit::doUpdateCommand:
getCurrCommand()->getType()->update(this); // issues auto commands by
updating unit commands (stop, move, repair, etc)
g_simInterface.updateSkillCycle(this); // syncs moves
frameProccessed(); // process commands from server (update keyframe)
The auto commands would be using the client's unit positions rather than
the server's synced ones.
Also the map height is float instead of fixed. I'm not sure if that could
cause issues setting the next pos with the pathfinder.
Nathan
On 24 November 2011 11:13, Nathan Turner
<hai...@us...>wrote:
> Hello,
>
> I said previously the network design didn't really fit how ENet works so
> I've done a redesign which I think is a lot cleaner. It's split into host
> and peer (session) with client and server specific host functions in child
> classes. ClientInterface and ServerInterface handles the connect and
> disconnect events while messages are managed by sessions.
>
> I ran a game with idle server player, idle networked player and an AI.
> When the AI defeated one of the idle players it had a sync error.
>
> That's a good point about the structs Daniel. I imagine it would be the
> same with classes.
>
> Nathan
>
>
> On 2 November 2011 13:33, Daniel Santos <dan...@at...> wrote:
>
>>
>> > @dansan,
>> > The imposed random network delay will be nice for testing, and will
>> > indeed be usable with eNet, much of the rest becomes redundant I
>> > think. While your DataBuffer classes offers some nice functionality,
>> > personally I'm perfectly content using memcpy() and assuming
>> > little-endian architecture all-around.
>> > If code that purports to solve byte order issues is added, it needs
>> > to be tested first... and for big endian hardware I think this is
>> > probably not likely.
>> > As to Message structure/methodology, I'm in favour of whatever
>> > works, for now what we have works, not as nice to use as it could be
>> > perhaps, but it works.
>> >
>> > I'm currently in favour of "getting shit working". We can always
>> > re-factor to make it look like it was well designed later ;-P
>> >
>> Yeah, but it's more than just endianness. There is the ABI as well.
>> For instance, for the following structure:
>>
>> struct Jack {
>> char bob;
>> char cindy;
>> int jill;
>> };
>>
>> One compiler may lay it out like this in memory
>>
>> Jack fred; <-- 0x5000:
>>
>> 0x5000 bob
>> 0x5001 cindy
>> 0x5002 unused
>> 0x5003 unused
>> 0x5004 jill
>> 0x5005 jill
>> 0x5006 jill
>> 0x5007 jill
>>
>> while another lays it out like this:
>>
>> 0x5000 bob
>> 0x5001 cindy
>> 0x5002 jill
>> 0x5003 jill
>> 0x5004 jill
>> 0x5005 jill
>>
>> or even this:
>>
>> 0x5000 bob
>> 0x5001 unused
>> 0x5002 unused
>> 0x5003 unused
>> 0x5004 cindy
>> 0x5005 unused
>> 0x5006 unused
>> 0x5007 unused
>> 0x5008 jill
>> 0x5009 jill
>> 0x500a jill
>> 0x500f jill
>>
>> The bottom line is that copying memory of a struct directly is never a
>> good practice when you are leaving your machine and going across the
>> network. I know that some of this may be solved with the pragma pack 1,
>> but it doesn't solve all of the problems associated with it and it also
>> inhibits the optimization possibilities of the compiler and even force
>> slower operations (depending upon the CPU). Then, add to that the fact
>> that NetworkString transmits the max string size in bytes all of the
>> time, regardless of the size of the actual string. Also, don't forget
>> that iPhone and many other phones use ARM processors, which are
>> little-endian.
>>
>> Anyway, I don't remember all of the issues I unearthed of it when I was
>> working on this before, but there are a lot of them.
>>
>> As for NetworkDataBuffer, it was fully tested when I was working on this
>> in 2009. That wasn't the part that was broken -- it was the fact that I
>> kept running off and breaking new things before I had finished fixing
>> the last one! :)
>>
>> Daniel
>>
>>
>> ------------------------------------------------------------------------------
>> RSA® Conference 2012
>> Save $700 by Nov 18
>> Register now!
>> http://p.sf.net/sfu/rsa-sfdev2dev1
>> _______________________________________________
>> Glestae-devel mailing list
>> Gle...@li...
>> https://lists.sourceforge.net/lists/listinfo/glestae-devel
>>
>
>
|
|
From: Frank T. <te...@us...> - 2011-11-24 10:35:19
|
Hi, you can rewrite your commit locally and push with force. See: http://christoph.ruegg.name/blog/2010/5/5/git-howto-revert-a-commit-already-pushed-to-a-remote-reposit.html As mentioned there it's actually not really a good idea. Normally one should use revert instead of reset on pushed commits. It doesn't delete the commit. It reverts all effects with a new commit, like reversed patch. I recently learned about notes (see man git-notes). If you just want to change something in the commit message that's probably the easiest and cleanest way. It adds a note to the commit message, not changing it though. Btw, i mostly use `gitk --all` to look at all the commits i want to push. Very nice tool. Regards, Frank. On Wed, 23 Nov 2011 13:16:47 -0600 Daniel Santos <dan...@at...> wrote: > Hey, does anybody know how to change commit messages after they've > already been pushed? I'm not sure what happened, but I know that I > amended the commit locally before pushing it, but it still had my commit > message from when I made the initial change. :( On the bright side, it > worked on the first attempt :) |
|
From: Nathan T. <hai...@us...> - 2011-11-24 01:14:06
|
Hello,
I said previously the network design didn't really fit how ENet works so
I've done a redesign which I think is a lot cleaner. It's split into host
and peer (session) with client and server specific host functions in child
classes. ClientInterface and ServerInterface handles the connect and
disconnect events while messages are managed by sessions.
I ran a game with idle server player, idle networked player and an AI. When
the AI defeated one of the idle players it had a sync error.
That's a good point about the structs Daniel. I imagine it would be the
same with classes.
Nathan
On 2 November 2011 13:33, Daniel Santos <dan...@at...> wrote:
>
> > @dansan,
> > The imposed random network delay will be nice for testing, and will
> > indeed be usable with eNet, much of the rest becomes redundant I
> > think. While your DataBuffer classes offers some nice functionality,
> > personally I'm perfectly content using memcpy() and assuming
> > little-endian architecture all-around.
> > If code that purports to solve byte order issues is added, it needs
> > to be tested first... and for big endian hardware I think this is
> > probably not likely.
> > As to Message structure/methodology, I'm in favour of whatever
> > works, for now what we have works, not as nice to use as it could be
> > perhaps, but it works.
> >
> > I'm currently in favour of "getting shit working". We can always
> > re-factor to make it look like it was well designed later ;-P
> >
> Yeah, but it's more than just endianness. There is the ABI as well.
> For instance, for the following structure:
>
> struct Jack {
> char bob;
> char cindy;
> int jill;
> };
>
> One compiler may lay it out like this in memory
>
> Jack fred; <-- 0x5000:
>
> 0x5000 bob
> 0x5001 cindy
> 0x5002 unused
> 0x5003 unused
> 0x5004 jill
> 0x5005 jill
> 0x5006 jill
> 0x5007 jill
>
> while another lays it out like this:
>
> 0x5000 bob
> 0x5001 cindy
> 0x5002 jill
> 0x5003 jill
> 0x5004 jill
> 0x5005 jill
>
> or even this:
>
> 0x5000 bob
> 0x5001 unused
> 0x5002 unused
> 0x5003 unused
> 0x5004 cindy
> 0x5005 unused
> 0x5006 unused
> 0x5007 unused
> 0x5008 jill
> 0x5009 jill
> 0x500a jill
> 0x500f jill
>
> The bottom line is that copying memory of a struct directly is never a
> good practice when you are leaving your machine and going across the
> network. I know that some of this may be solved with the pragma pack 1,
> but it doesn't solve all of the problems associated with it and it also
> inhibits the optimization possibilities of the compiler and even force
> slower operations (depending upon the CPU). Then, add to that the fact
> that NetworkString transmits the max string size in bytes all of the
> time, regardless of the size of the actual string. Also, don't forget
> that iPhone and many other phones use ARM processors, which are
> little-endian.
>
> Anyway, I don't remember all of the issues I unearthed of it when I was
> working on this before, but there are a lot of them.
>
> As for NetworkDataBuffer, it was fully tested when I was working on this
> in 2009. That wasn't the part that was broken -- it was the fact that I
> kept running off and breaking new things before I had finished fixing
> the last one! :)
>
> Daniel
>
>
> ------------------------------------------------------------------------------
> RSA® Conference 2012
> Save $700 by Nov 18
> Register now!
> http://p.sf.net/sfu/rsa-sfdev2dev1
> _______________________________________________
> Glestae-devel mailing list
> Gle...@li...
> https://lists.sourceforge.net/lists/listinfo/glestae-devel
>
|
|
From: Daniel S. <dan...@at...> - 2011-11-23 19:22:17
|
Yeah Silnarn, same thing on gcc. It turns out that this is because the *full* C++11 language isn't supported. Therefore, we need to change this somehow. I'm not 100% sure what we need to check on MSVC to make sure that the C++11 features are "enabled", but I suspect that MS never even bothers with language specifications, as they seem to extend standard languages at a whim, so it may just be a matter of checking the version of the compiler. I do know, however, that on MSVC, you can disable link-time optimizations and possibly some other features by disabling the "common language runtime", so there is at least some variation based upon cl.exe's input switches. Anyway, I'm going to patch this in a bit. On GCC, I can check for it with __GXX_EXPERIMENTAL_CXX0X__, so I'll probably just change it for now to #if (__cplusplus > 199711L) || defined(_MSC_FULL_VER) || defined(__GXX_EXPERIMENTAL_CXX0X__) |
|
From: Daniel S. <dan...@at...> - 2011-11-23 19:19:24
|
Hey, does anybody know how to change commit messages after they've already been pushed? I'm not sure what happened, but I know that I amended the commit locally before pushing it, but it still had my commit message from when I made the initial change. :( On the bright side, it worked on the first attempt :) |
|
From: Daniel S. <dan...@at...> - 2011-11-19 02:00:04
|
Thanks for the response! I'll round up my patches and put them in master then. > I'm not quite sure about using > shared_ptr nearly everywhere. You'll do garbage collection with > reference counting then. I doubt that this won't affect performance. Well, you don't really want to use it "nearly everywhere", rather, it should be used "nearly everywhere that a function returns a new object", as well as other places where you've allocated something, especially some that's shared, and you want to make sure it's freed. Places you wouldn't want to use it are areas where large numbers of objects are created and destroyed frequently. For classes where you create an object in your constructor, never share it's pointer and then delete it in the destructor, it's really not too important beyond just being able to be lazy in your destructor. Usually in such cases, one would just have a data member that is of the type of that object instead of a pointer to it anyway. But one prime example where you absolutely *should* use it is GameSettings. This object is created in the various "new game" menus (or at least it used to be) and then passed around everywhere, as the state machine changes. Finally, at the end of the game, it's passed to the BattleEnd program state (well, at least it is in the old network re-write branch). Another example is the XmlNode's toString() function. Daniel |
|
From: Frank T. <te...@us...> - 2011-11-18 15:12:15
|
Hi, i made a new branch 0.4.x where we should get our stuff ready for release. So you can start to merge other things into master like boost. I'm fine with your foreach changes. Actually i don't really care as i never used it. It looks reasonable. I'm not quite sure about using shared_ptr nearly everywhere. You'll do garbage collection with reference counting then. I doubt that this won't affect performance. Regards, Frank. |
|
From: James M. <si...@gm...> - 2011-11-18 08:12:32
|
> Hi all, Greetings Daniel, > I got interested in glest a while back, got busy on some other stuff, and > recently came back to it. > I'd really like to be able to contribute but I got some questions: > 1- What lies in the future of GAE? I've been reading on some forum threads > that perhaps GAE and MegaGlest would merge into one entity. Is that the > case? If so when? I'm asking because I can run Glest and GAE on my pc > (albeit it's a bit slow), but MegaGlest runs to a crawl. So you'll > understand that If I am to be soonish unable to run a game I want to work > on, well... doesn't fare well... Besides a possible merge, where does GAE > want to go with the tech/game? We are a bunch of programmers, and as such our focus is primarily on enhancing the engine for modders, and of course adding nicities for players. We have recently made minor modifications to magitech to make use of some of the new features the engine supports, and will likely do so again in the future, but in terms of extending magitech with new units etc, we don't really have any plans. I'd personally like to make a campaign for magitech some day, but that will take some time, and I don't have a lot of that spare atm :( Plans are afoot to improve performance of various parts of the game, and these changes involve using sensible data structures and algorithms, rather than just caching everything in sight and wasting memory and encouraging (hardware) cache misses, as is the case with MG. We certainly do not intend to make it run slower, though with active development going on such slow-downs may temporarily manifest themselves. As Tom has mentioned, the merge is off, different development philosophies and the contrasted views of certain stubborn team members (mainly myself from GAE, and Titi from MG) have made that largely an impossibility. > 2- i've got the source with TurtoiseGit and got it to compile/run and all, > do I really need to do all that key stuff to be able to push back unto the > server? I'm mainly asking so I don't start changing files and then need to > redo everything 'cause I wasn't set up properly. It will make life much easier if you do, I recently moved a local repository from one computer to another and am still able to pull/push with it, but tortoise complains about being unable to load my key and then I have to enter my sourceforge username and password everytime, which is a bit annoying. It may seem like a pain, but you only have to do it once ;-) You need not worry about having to redo anything though, changes you make now can still be committed to your local repo and patches generated from that, which we can then apply and push (all the log messages still show you as author, one of the many nice things about git), and when granted repo access you can simply reset to "remote" to the ssh url and you're ready to go, we can help out there when the time comes. > 3- How are tasks assigned usually? How goes the dynamic between the team > members? Is there a boss like person that can veto everything ans assigns > everything? How are submissions revised? I'm trying to get a sense of how > working with you guys will feel/look like. There is no boss like person who controls all, we generally operate in a fairly democratic manner, but clashes concerning how things should be done haven't really happened in my memory, and if you are willing to implement something that is of benefit to modders and/or players (or even just to make our own lives easier) then you are not likely to hear any complaints, if there are complaints/concerns then it may need to be made optional (if that makes sense for the feature/change in question). Code review generally happens once its in the repo, and then probably only if it causes unforeseen problems. I generally browse through the changesets for most (if not all) commits, but this not much of a review really, just trying to keep up with what people are doing. > Furthermore, I've been looking into the list of tickets for v0.5 and I feel > confident enough I can tackle some of the music stuff. They sound simple > enough for introductory tasks: > 187- non-faction music playlist > 291 - music playlist per subfaction > 292 - allow music playlist defined per unit These do indeed seem to be good candidates (although I'm unsure how 292 is meant to work...), and we are happy to provide assistance/explanations with/for anything that has you stumped, though the 'turn around time' on that may not always be as prompt as you would like ;-) We'd be delighted to see some patches and then promptly get you write access to the repo, contributors are seemingly hard to come by, so from myself and on behalf of Nathan (hailstone), Frank (Yggdrasil) and the other Daniel (forum name daniel.santos, but I use his irc nick 'dansan', because I like it), we hope to have you on board soon! Cheers, - James (aka silnarm) |
|
From: Tom R. <to...@me...> - 2011-11-17 20:12:37
|
Hi Daniel, On 16.11.2011 16:49 Daniel Carrier wrote: > I'd really like to be able to contribute but I got some questions: > 1- What lies in the future of GAE? I've been reading on some forum threads > that perhaps GAE and MegaGlest would merge into one entity. Is that the > case? If so when? I'm asking because I can run Glest and GAE on my pc > (albeit it's a bit slow), but MegaGlest runs to a crawl. So you'll > understand that If I am to be soonish unable to run a game I want to work > on, well... doesn't fare well... Besides a possible merge, where does GAE > want to go with the tech/game? being part of the MegaGlest (but not GAE team) I could not answer this last question of yours (nor the other question you have which I have not quoted). Regarding the merge, my understanding is that it is not currently likely to happen (and the more GAE and MegaGlest are diverting the less likely it gets). The real reason I'm responding, though, is to offer you to help sorting out why 'MegaGlest runs to a crawl' on your computer. If you would like to look into this, please describe the issues as well as your PC specs on the MegaGlest forums at http://forums.megaglest.org - we will get back to you. Tom |
|
From: Daniel S. <dan...@at...> - 2011-11-16 23:47:43
|
I see that dev has been merged. Are we ready to begin requiring boost in master or should I branch for this? I'm thinking that foreach3 would be a good name to use to replace the current foreach macro (the three for the number of arguments). Then, the definitions of the new foreach macro will go in lang_features.h (see attachment). Since Boost's foreach syntax was adopted into C++11, their foreach.hpp doesn't need to be included if it's supported. Currently, I think this is only gcc 4.6 and later, but I'm sure MSVC will be there shortly. I'm also adding a new foreach_col macro, in case you prefer to use it over the ranged-for. It accepts two arguments and the type of the iterator is automatically determined. #define foreach_col(it, collection) for(BOOST_AUTO (it, (collection).begin()); it != (collection).end(); ++it) Anyway, if we move forward with this now, we can begin changing functions that return newly allocated objects to return a shared_ptr instead of raw pointers to objects (where reasonable of course). shared_ptr has very little overhead and is a nice way assure that memory leaks don't occur. The places it isn't appropriate is areas where large numbers of objects are being created, like with particle systems. Since shared_ptr is also in TR1, I should probably add a check for it in lang_features.h and pull in the boost implementation if it's not supported by the compiler's standard C++ library. Daniel |
6 messages has been excluded from this view by a project administrator.