Re: [Warzone-dev] excessive check in actionUpdateDroid
On 7/7/07, Per Inge Mathisen <[EMAIL PROTECTED]> wrote: Then the died flag will also have a random value. It just so happens that > 0 is a more common random value than == 0. This is, IMHO, not a fix at all. We should fix the root problem, not remove the warnings indicating that there is a problem. I've been playing the campaign lately with the reference counting patch without target problems (lots of *other* problems, though), which makes me believe the patch may be ready for inclusion soon. - Per ___ Warzone-dev mailing list Warzone-dev@gna.org https://mail.gna.org/listinfo/warzone-dev hehe glad to hear the target problem is 'gone' with the few fixes in ai.cand stuff.Theoretically that assert in actionUpdateDroid.c shouldnt happen again if we have fixed all of those missing 'died' checks. The campaign is very broken indeed,though it's script-driven,so many problems might be script engine related,but most of us are not very familiar with that rather complicated scripting implementation. ___ Warzone-dev mailing list Warzone-dev@gna.org https://mail.gna.org/listinfo/warzone-dev
Re: [Warzone-dev] excessive check in actionUpdateDroid
On 7/7/07, The Watermelon <[EMAIL PROTECTED]> wrote: > > Sorry, maybe I'm just a little slow, but why would it have invalid > > facing because it died? > > it shouldnt but sometimes the psActionTarget got freed without setting to > 0,thus holding some random value and causes the assert.Luckily 'died' flag > check detects that problem promptly and serves as a crash-proof check I > guess. Then the died flag will also have a random value. It just so happens that > 0 is a more common random value than == 0. This is, IMHO, not a fix at all. We should fix the root problem, not remove the warnings indicating that there is a problem. I've been playing the campaign lately with the reference counting patch without target problems (lots of *other* problems, though), which makes me believe the patch may be ready for inclusion soon. - Per ___ Warzone-dev mailing list Warzone-dev@gna.org https://mail.gna.org/listinfo/warzone-dev
Re: [Warzone-dev] excessive check in actionUpdateDroid
On 7/7/07, Per Inge Mathisen <[EMAIL PROTECTED]> wrote: On 7/7/07, The Watermelon <[EMAIL PROTECTED]> wrote: > On 7/6/07, Per Inge Mathisen <[EMAIL PROTECTED]> wrote: > > On 7/6/07, The Watermelon <[EMAIL PROTECTED]> wrote: > > > > removed the excessive CHECK_DROID at the beginning of > > > actionUpdateDroid,because it already checks died flag/clear died ones. > > > > I do not understand. CHECK_DROID does not check the died flag. > > CHECK_DROID checks psDroid's psActionTarget's direction,but the > psActionTarget might be 'died' and has invalid facing,because the > psActionTarget died check is performance in actionUpdateDroid after that > check. Sorry, maybe I'm just a little slow, but why would it have invalid facing because it died? - Per it shouldnt but sometimes the psActionTarget got freed without setting to 0,thus holding some random value and causes the assert.Luckily 'died' flag check detects that problem promptly and serves as a crash-proof check I guess. ___ Warzone-dev mailing list Warzone-dev@gna.org https://mail.gna.org/listinfo/warzone-dev
Re: [Warzone-dev] excessive check in actionUpdateDroid
On 7/7/07, The Watermelon <[EMAIL PROTECTED]> wrote: > On 7/6/07, Per Inge Mathisen <[EMAIL PROTECTED]> wrote: > > On 7/6/07, The Watermelon <[EMAIL PROTECTED]> wrote: > > > > removed the excessive CHECK_DROID at the beginning of > > > actionUpdateDroid,because it already checks died flag/clear died ones. > > > > I do not understand. CHECK_DROID does not check the died flag. > > CHECK_DROID checks psDroid's psActionTarget's direction,but the > psActionTarget might be 'died' and has invalid facing,because the > psActionTarget died check is performance in actionUpdateDroid after that > check. Sorry, maybe I'm just a little slow, but why would it have invalid facing because it died? - Per ___ Warzone-dev mailing list Warzone-dev@gna.org https://mail.gna.org/listinfo/warzone-dev
Re: [Warzone-dev] excessive check in actionUpdateDroid
On 7/6/07, Per Inge Mathisen <[EMAIL PROTECTED]> wrote: On 7/6/07, The Watermelon <[EMAIL PROTECTED]> wrote: > > removed the excessive CHECK_DROID at the beginning of > actionUpdateDroid,because it already checks died flag/clear died ones. I do not understand. CHECK_DROID does not check the died flag. - Per CHECK_DROID checks psDroid's psActionTarget's direction,but the psActionTarget might be 'died' and has invalid facing,because the psActionTarget died check is performance in actionUpdateDroid after that check. ___ Warzone-dev mailing list Warzone-dev@gna.org https://mail.gna.org/listinfo/warzone-dev
Re: [Warzone-dev] excessive check in actionUpdateDroid
On 7/6/07, The Watermelon <[EMAIL PROTECTED]> wrote: > > removed the excessive CHECK_DROID at the beginning of > actionUpdateDroid,because it already checks died flag/clear died ones. I do not understand. CHECK_DROID does not check the died flag. - Per ___ Warzone-dev mailing list Warzone-dev@gna.org https://mail.gna.org/listinfo/warzone-dev
Re: [Warzone-dev] excessive check in actionUpdateDroid
On 7/6/07, The Watermelon <[EMAIL PROTECTED]> wrote: removed the excessive CHECK_DROID at the beginning of actionUpdateDroid,because it already checks died flag/clear died ones. oops forgot to attach the patch. exdroidcheck.patch Description: Binary data ___ Warzone-dev mailing list Warzone-dev@gna.org https://mail.gna.org/listinfo/warzone-dev