[Bug java/26538] New: 'make install' to non-standard prefix fails with java enabled
Hi, I am using Gentoo Linux. I followed the following steps: Make directory ~/src/packages/gcc-4.1-svn Change directory to ~/src/packages/gcc-4.1-svn Get sources via svn co svn://gcc.gnu.org/svn/gcc/branches/gcc-4_1-branch gcc-4.1-branch make directory ~/src/packages/gcc-4.1-svn/test_build change to directory ~/src/packages/gcc-4.1-svn/test_build run bash -c '../gcc-4.1-branch/configure --prefix=/home/ayqazi/src/packages/gcc-4.1-svn/test_root --disable-nls --enable-version-specific-runtime-libs --enable-__cxa_atexit --enable-languages=c,c++,java --with-system-zlib --enable-java-awt=gtk --enable-gtk-cairo nice make profiledbootstrap make install' ../test_build.log Here is the end of the file test_build.log: make[3]: Entering directory `/home/ayqazi/src/packages/gcc-4.1-svn/test_build/i686-pc-linu x-gnu/libjava/libltdl' make[4]: Entering directory `/home/ayqazi/src/packages/gcc-4.1-svn/test_build/i686-pc-linu x-gnu/libjava/libltdl' test -z /home/ayqazi/src/packages/gcc-4.1-svn/test_root/lib || mkdir -p -- /home/ayqazi /src/packages/gcc-4.1-svn/test_root/lib test -z /home/ayqazi/src/packages/gcc-4.1-svn/test_root/include || mkdir -p -- /home/ay qazi/src/packages/gcc-4.1-svn/test_root/include make[4]: Leaving directory `/home/ayqazi/src/packages/gcc-4.1-svn/test_build/i686-pc-linux -gnu/libjava/libltdl' make[3]: Leaving directory `/home/ayqazi/src/packages/gcc-4.1-svn/test_build/i686-pc-linux -gnu/libjava/libltdl' Making install in gcj make[3]: Entering directory `/home/ayqazi/src/packages/gcc-4.1-svn/test_build/i686-pc-linu x-gnu/libjava/gcj' make[4]: Entering directory `/home/ayqazi/src/packages/gcc-4.1-svn/test_build/i686-pc-linu x-gnu/libjava/gcj' make[4]: Nothing to be done for `install-exec-am'. test -z /include/c++/gcj || mkdir -p -- /include/c++/gcj mkdir: cannot create directory `/include': Permission denied make[4]: *** [install-gcjHEADERS] Error 1 make[4]: Leaving directory `/home/ayqazi/src/packages/gcc-4.1-svn/test_build/i686-pc-linux -gnu/libjava/gcj' make[3]: *** [install-am] Error 2 make[3]: Leaving directory `/home/ayqazi/src/packages/gcc-4.1-svn/test_build/i686-pc-linux -gnu/libjava/gcj' make[2]: *** [install-recursive] Error 1 make[2]: Leaving directory `/home/ayqazi/src/packages/gcc-4.1-svn/test_build/i686-pc-linux -gnu/libjava' make[1]: *** [install-target-libjava] Error 2 make[1]: Leaving directory `/home/ayqazi/src/packages/gcc-4.1-svn/test_build' make: *** [install] Error 2 Full build log (contents of 'test_build.log') available on request. -- Summary: 'make install' to non-standard prefix fails with java enabled Product: gcc Version: 4.1.1 Status: UNCONFIRMED Severity: critical Priority: P3 Component: java AssignedTo: unassigned at gcc dot gnu dot org ReportedBy: ayqazi at yahoo dot co dot uk GCC build triplet: i686-pc-linux GCC host triplet: i686-pc-linux GCC target triplet: i686-pc-linux http://gcc.gnu.org/bugzilla/show_bug.cgi?id=26538
[Bug java/26538] 'make install' to non-standard prefix fails with java enabled
--- Comment #1 from ayqazi at yahoo dot co dot uk 2006-03-03 06:49 --- Created an attachment (id=10958) -- (http://gcc.gnu.org/bugzilla/attachment.cgi?id=10958action=view) Log of just the 'make install' command make install executed separately again, and its complete output logged, bzip2'ed, and attached here. -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=26538
[Bug c++/18418] [3.4 only] GCC 3.4.3 builds worse code than GCC 3.3.4 using template expressions
--- Additional Comments From ayqazi at yahoo dot co dot uk 2005-02-10 17:23 --- Once GCC 4.0 is out, I'll experiment with it and submit test cases etc. for it. No use trying to fix an older release's optimisations IMHO. Thanks anyway. -- What|Removed |Added Status|WAITING |SUSPENDED http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18418
[Bug c++/18415] New: Heavily template-expression using program crashes on -O0 but not on -O1
When running a program that makes heavy use of expression templates, it works
fine with any -O option. However, if no -O option is used, then it segfaults.
Note that if the expression used is changed to something like arg + arg, then it
doesn't crash but gives wrong results if no -O option is specified.
typedef unsigned int size_t;
struct AddOp
{
templateclass I1, class I2 static inline float
call(size_t i, const I1 i1, const I2 i2) {return i1[i] + i2[i];}
};
struct MultOp
{
templateclass I1, class I2 static inline float
call(size_t i, const I1 i1, const I2 i2) {return i1[i] * i2[i];}
};
class Expr
{
float data;
public:
Expr(float arg_data) : data(arg_data) {}
float
operator[](size_t i) const {return data;}
};
templateclass I1, class I2, class Op
class Expr3
{
const I1 i1;
const I2 i2;
public:
Expr3(const I1 ai1, const I2 ai2)
: i1(ai1), i2(ai2)
{
}
float
operator[](size_t i) const
{
return Op::call(i, i1, i2);
}
templateclass E
Expr3Expr3, E, AddOp
operator+(const E e) const
{
return Expr3Expr3, E, AddOp(*this, e);
}
};
struct Vec4
{
Vec4(float all)
: x(all), y(all), z(all), w(all)
{
}
Vec4(float ax, float ay, float az, float aw = 1)
: x(ax), y(ay), z(az), w(aw)
{
}
float x, y, z, w;
float
operator[](size_t i) const
{
switch(i) {
case 0: return x;
case 1: return y;
case 2: return z;
case 3: return w;
default: return 0xdeadbeef;
}
}
const Vec4
operator=(float arg)
{
x = y = z = w = arg; return *this;
}
Expr3Vec4, Expr, MultOp
operator*(float e) const
{
return Expr3Vec4, Expr, MultOp(*this, Expr(e));
}
templateclass E
const Vec4
operator=(const E e)
{
x = e[0]; y = e[1]; z = e[2]; w = e[3]; return *this;
}
};
templateclass T
Expr3Expr3Vec4, Expr, MultOp, Expr3Vec4, Expr, MultOp, AddOp
dostuff(const T arg)
{
return (arg * 0.f) + (arg * 0.1f);
}
int
main()
{
const float ARG1 = 3.40282347e+38f;
Vec4 v1(ARG1), v2(ARG1);
v2 = dostuff(v1);
}
--
Summary: Heavily template-expression using program crashes on -O0
but not on -O1
Product: gcc
Version: 3.4.3
Status: UNCONFIRMED
Severity: normal
Priority: P2
Component: c++
AssignedTo: unassigned at gcc dot gnu dot org
ReportedBy: ayqazi at yahoo dot co dot uk
CC: gcc-bugs at gcc dot gnu dot org
GCC build triplet: i686-pc-linux
GCC host triplet: i686-pc-linux
GCC target triplet: i686-pc-linux
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18415
[Bug c++/18415] Heavily template-expression using program crashes on -O0 but not on -O1
--- Additional Comments From ayqazi at yahoo dot co dot uk 2004-11-10 12:38 --- Created an attachment (id=7515) -- (http://gcc.gnu.org/bugzilla/attachment.cgi?id=7515action=view) Test case to reproduce bug -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18415
[Bug c++/18418] New: GCC 3.4.3 builds worse code than GCC 3.3.4 using template expressions
Hi, Using a template expressions 3D maths library I'm developing, I have discovered that GCC 3.4.3 builds worse code than GCC 3.3.4 (and predecessors.) I have attached a test case file using a part of the template expression library that highlights this case. Regardless of which -march flag I used to compile it (i386, etc.) and regardless of the -msse flag being present or not, the asm output was clearly longer by about 20% in the gcc 3.4.3 compiled code. p.s. it will work if an optimisation flag is present (I have verified it.) Without an optimisation flag, GCC doesn't generate proper code (this bug has already been reported.) -- Summary: GCC 3.4.3 builds worse code than GCC 3.3.4 using template expressions Product: gcc Version: 3.4.3 Status: UNCONFIRMED Severity: normal Priority: P2 Component: c++ AssignedTo: unassigned at gcc dot gnu dot org ReportedBy: ayqazi at yahoo dot co dot uk CC: gcc-bugs at gcc dot gnu dot org GCC build triplet: i686-linux GCC host triplet: i686-linux GCC target triplet: i686-linux http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18418
[Bug c++/18418] GCC 3.4.3 builds worse code than GCC 3.3.4 using template expressions
--- Additional Comments From ayqazi at yahoo dot co dot uk 2004-11-10 13:54 --- Created an attachment (id=7516) -- (http://gcc.gnu.org/bugzilla/attachment.cgi?id=7516action=view) Test case for differences in GCC 3.4.3 and 3.3.4 template expression code gen Compile to assembler (e.g. g++ -S -O) using same flags with both a GCC 3.3 series compiler, and a GCC 3.4 series compiler and spot the file length difference - you MUST use an optimisation flag for it to compile correctly (another bug, already submitted.) -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18418
[Bug c++/18415] Template-expression using program crashes on -O0 but not on -O1
-- What|Removed |Added Summary|Heavily template-expression |Template-expression using |using program crashes on -O0|program crashes on -O0 but |but not on -O1 |not on -O1 http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18415
