On 01/05/13 15:03, Carsten Haitzler (The Rasterman) wrote: > On Wed, 01 May 2013 15:00:57 +0100 Tom Hacohen <tom.haco...@samsung.com> said: > >> On 01/05/13 14:47, Carsten Haitzler (The Rasterman) wrote: >>> On Wed, 01 May 2013 14:37:37 +0100 Tom Hacohen <tom.haco...@samsung.com> >>> said: >>> >>>> On 01/05/13 14:17, Carsten Haitzler (The Rasterman) wrote: >>>>> On Wed, 01 May 2013 14:03:42 +0100 Tom Hacohen <tom.haco...@samsung.com> >>>>> said: >>>>> >>>>>> On 01/05/13 14:07, Carsten Haitzler (The Rasterman) wrote: >>>>>>> On Wed, 01 May 2013 13:52:50 +0100 Tom Hacohen <tom.haco...@samsung.com> >>>>>>> said: >>>>>>> >>>>>>>> On 01/05/13 13:54, Carsten Haitzler (The Rasterman) wrote: >>>>>>>>> On Wed, 01 May 2013 11:00:01 +0100 Tom Hacohen >>>>>>>>> <tom.haco...@samsung.com> said: >>>>>>>>> >>>>>>>>>> On 01/05/13 10:58, Carsten Haitzler (The Rasterman) wrote: >>>>>>>>>>> On Wed, 01 May 2013 10:08:48 +0100 Tom Hacohen >>>>>>>>>>> <tom.haco...@samsung.com> said: >>>>>>>>>>> >>>>>>>>>>>> On 30/04/13 18:48, Carsten Haitzler (The Rasterman) wrote: >>>>>>>>>>>>> On Tue, 30 Apr 2013 15:15:05 +0100 Tom Hacohen >>>>>>>>>>>>> <tom.haco...@samsung.com> said: >>>>>>>>>>>>> >>>>>>>>>>>>>> Where did you get that on? >>>>>>>>>>>>>> Anyhow, what do you think about changing it to unsigned wchar_t? >>>>>>>>>>>>> >>>>>>>>>>>>> on my pentium-m test machine... unicode val 0 was < 0 and thus >>>>>>>>>>>>> walked below the array. yes. literally a negative. wouldnt that >>>>>>>>>>>>> be wuchar_t or something? as wchar_t .. is a typedef... :) >>>>>>>>>>>>> >>>>>>>>>>>> >>>>>>>>>>>> Hm... Annoying. There's no wuchar_t though. >>>>>>>>>>> >>>>>>>>>>> then we have... a problem... and it requires we check for < 0. :( >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>> >>>>>>>>>> I think me might be better off casting to unsigned. Damn you people >>>>>>>>>> for not doing all the char type unsigned, wth?! >>>>>>>>> >>>>>>>>> chances are the compiler will produce the exact same code >>>>>>>>> regardless... a cast or what is there now. >>>>>>>>> >>>>>>>> >>>>>>>> Nah, the whole point of the cast is to convert it to unsigned, I'm >>>>>>>> quite certain the compiler is capable of doing that. >>>>>>> >>>>>>> but your casting inside the func to just avoid if (x < 0)... which a >>>>>>> compiler will figure out to be the same as the cast to unsigned... >>>>>>> >>>>>>> >>>>>> >>>>>> The cast will work for unicode values that are greater than the signed >>>>>> limit (less than 0), while the if just fail for them. >>>>> >>>>> there are no unicode values of that magnitude... unicode by definition >>>>> doesnt even get close to using the most significant bit... :) it's by >>>>> definition an invalid code if < 0 (for 32bit signed)... >>>>> >>>>> >>>> >>>> Oh, forgot to mention, I was thinking about boxes where wchar_t is >>>> signed and 16bit. :) There we'll have trouble. >>>> If you got negative values it must mean you've reached big enough >>>> unicode values, so the issue I'm describing is indeed real. >>> >>> on boxes where its 16bit.. we will have problems... because unicode does not >>> fit into 16bit.... we explicitly MUSt have it be a 32bit type in order to >>> have enough space to store the hmmm... 22? bits needed for unicode? quick >>> check... 10ffff is the top unicode value... that means 21bits... so >>> unicode needs 21bits. if we have 16bit wchar_t's we are not able to do >>> unicode. signed or not is irrelevant here. if its 32bit... we don't care :) >>> >> >> Well, a subset of... :) >> >> But anyhow, how did you get your issue then? That it was negative? >> That's what I'm interested in, as that means it's a path we actually get to. > > it looked like a garbage buffer... but evas shouldnt segv if there is an > invalid unicode value there... :) > >
Haha, so you were hiding your bugs using my code. Making me an accessory to segfault! OK though, I agree. -- Tom. ------------------------------------------------------------------------------ Introducing AppDynamics Lite, a free troubleshooting tool for Java/.NET Get 100% visibility into your production application - at no cost. Code-level diagnostics for performance bottlenecks with <2% overhead Download for free and get started troubleshooting in minutes. http://p.sf.net/sfu/appdyn_d2d_ap1 _______________________________________________ enlightenment-devel mailing list enlightenment-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/enlightenment-devel
