@dohashi I have already tried -Xss......but it doesn't work

On linux, I have to issue "ulimit -s ${some desire stack max size}" first, then the -Xss will help.

But on windows, there is no such command that has similar functionality with "ulimit -s". And no matter how big stack size I specified  via -Xss, the stack size that maple kernel could use remains the same(893 on my mechine)....

So any thoughts about this situation?

@tomleslie very grateful for your explaination...since I start to work with maple which is only one month, I happen to find two significant bugs, including this one...

i wonder where i can find the maple bug trace site if there is any. So i can check the bugs' state.

@tomleslie try to compare the outcome when issue the following commands :

 sys := {x = abs(y-1), 5^(2*x+2) = 25^(2*x-3), 5^(3*x+1)/5^(x-1) = 25^(2*x-3)};

solve(sys,{x,y});

 sys2 := {x^2 = (abs(y-1))^2, 5^(2*x+2) = 25^(2*x-3), 5^(3*x+1)/5^(x-1) = 25^(2*x-3)};

solve(sys2,{x,y});

@acer i am using maple 2015 installed on windows 7.

"kernelopts(stacklimit) " gives me 130941 if issued in maple GUI or command line.

but gives me 889 if issued by java code.

Groebner[Basis](numer([arean1-(1/2)*lenc1*lenk1, arean1-(1/2)*lenc1*lenk1, areaf4-(1/2)*lenk1*lenj1, areag4-(1/2)*lenk1*lenr2, lens2-lenn2-lenr2-lenk1, lenf2-lenb1-lenj1-lenk1, trig1-lenk1/lenb1, trig2-lenk1/lenb1, trig3-lenj1/lenk1, trig4-lenk1/lenj1, -lenb1^2+lenj1^2+lenk1^2, trig5-lenk1/lenn2, trig6-lenk1/lenn2, trig7-lenr2/lenk1, trig8-lenk1/lenr2, lenk1^2-lenn2^2+lenr2^2, lenk1-lenj1, lenj1/lenk1-lenb1*(1/lenb1), lenj1/lenk1-lenk1/lenj1, lenj1/lenk1-lenk1/lenj1, lenk1/lenj1-lenb1*(1/lenb1), lenj1*(1/lenj1)-lenk1*(1/lenk1), lenb1*(1/lenb1)-lenk1*(1/lenk1), lenl2/lenj1-lenc1/lenk1, lenc1/lenk1-lenu1/lenb1, lenc1/lenj1-lenl2/lenk1, lenu1/lenb1-lenl2/lenk1, lenv1/lenj1-lenw1/lenk1, lenu1/lenb1-lenw1/lenk1, lenw1/lenj1-lenv1/lenk1, lenu1/lenb1-lenv1/lenk1, lenl3/lenr2-lenp1/lenk1, lent2/lenn2-lenp1/lenk1, lene3/lenr2-leno1/lenk1, lenc2/lenn2-leno1/lenk1, areal1-(1/2)*lenb1*leno1, areac4-(1/2)*leno1*lene3, areab4-(1/2)*leno1*leng3, lenh3-leno1-leng3-lenb2, lenf3-lene3-leno1-lenc2, trig17-leno1/lenb2, trig18-leno1/lenb2, trig19-leno1/leng3, trig20-leng3/leno1, -lenb2^2+leng3^2+leno1^2, trig21-leno1/lenc2, trig22-leno1/lenc2, trig23-lene3/leno1, trig24-leno1/lene3, -lenc2^2+lene3^2+leno1^2, lene3/lenl3-leno1/lenp1, lenc2/lent2-leno1/lenp1, lene3/lenr2-leno1/lenk1, lenc2/lenn2-leno1/lenk1, aream1-(1/2)*lenc1*lenp1, areae4-(1/2)*lenp1*lenl3, aread4-(1/2)*lenp1*lenn3, leno3-lenp1-lenn3-lenb2, lenm3-lenp1-lenl3-lent2, trig14-lenp1/lenb2, trig13-lenp1/lenb2, trig12-lenp1/lenn3, trig11-lenn3/lenp1, -lenb2^2+lenn3^2+lenp1^2, trig5-lenp1/lent2, trig15-lenp1/lent2, trig8-lenp1/lenl3, trig16-lenl3/lenp1, lenl3^2+lenp1^2-lent2^2, lene3/lenl3-leno1/lenp1, lenc2/lent2-leno1/lenp1, lenl3/lenr2-lenp1/lenk1, lent2/lenn2-lenp1/lenk1, degi1-90, dege1-45, lenb1-1, lenk1-(1/2)*2^(1/2), arean1-areal1-aream1, areal1-(1/2)*lenb1*leno1, aream1-(1/2)*lenc1*lenp1, arean1-(1/2)*lenc1*lenk1, (1/2)*lenc1*lenk1-(1/2)*lenb1*leno1-(1/2)*lenc1*lenp1, lenb1-1]), 'tord')

issue this command in maple(GUI or commandline) will give me result...but will crash jvm if i use java code to call it.

@Markiyan Hirnyk 

consider these two commands:

a. RealDomain[solve]({(2*b+c)^(1/2) >= 0, 2*b+c >= 0, a-b = 0, 2*b+c = 0, c^2-c+1/4 = 0,(a*b > 0 or a*b<0)});

b. RealDomain[solve]({(2*b+c)^(1/2) >= 0, 2*b+c >= 0, a-b = 0, 2*b+c = 0, c^2-c+1/4 = 0, a*b <> 0});

command a gives me : {a = -(1/2)*c, b = -(1/2)*c, c = c}

command b gives me : {a = -1/4, b = -1/4, c = 1/2}

any idea?

@Markiyan Hirnyk your approach solving this equation inspires me another one. since your way ensentially is to find the global (or local) minimun of the left side of this equation...So why not solve the system of gradient equation? here is an example:

with(linalg):

f:= (a-b)^2+(2b+c)^2+c^2-c+1/4;

gf := grad(f,[a,b,c]);

solve({gf[1],gf[2],gf[3]},[a,b,c]);

the only problem is it cannot deal with absolute or square. but we always can square a absolute or square value to make them go away, like transform |a-b|=c to (a-b)^2=c^2 and sqrt(b)=c to b=c^2,b>=0...

@Adri van der Meer what if I don't want to tell maple the constraints explicitily? because all the expression or equation i give maple hasing meaning, like denominator cant be zero etc, and i want maple can take these things into account.

@Markiyan Hirnyk But when I try to execute "RealDomain:-solve({x^2 = 1, x <> 1})", maple gives me one solution.

Why？

great thanks to you, first...

but i cant find any help document about this DirectSearch package. So could you please help me understand the meaning of the output? Would appretiate that!