April 2007 - openSUSE Programming - openSUSE Mailing Lists

[opensuse-programming] C++ inheritance issue
by Jerry Feldman 01 May '07

01 May '07

I have a case where the compiler is giving me an error: ./FOO.H: In member function Bar<ItemType>::operator()(ItemType&)': ./FOO.H:447: error: `i_key' was not declared in this scope Class bar is defined: template <class ItemType> class Bar public FuBar<ItemType, type_t> And, i_key is defined as a protected member of FuBar: template <class ItemType, class ThisType> class FuBar : public Another<ItemType>, public RWTPtrorderedVector<ItemType> ..... protected: ThisType const & i_key; I'm using g++ version 3.4.5. g++ -MMD -MP -g -O3 -Wall -fpic -D__linux_x86_64__ -I. -I/<other included dirs> -D_RWCONFIG_rmd -c Foo.C -o Foo.o I can easily work around by adding a public accessor function, but this code does work on other company builds, but this is a simply access to a protected data element that should be visible from inheriting classes. -- Jerry Feldman <gaf(a)blu.org> Boston Linux and Unix user group http://www.blu.org PGP key id:C5061EA9 PGP Key fingerprint:053C 73EC 3AC1 5C44 3E14 9245 FB00 3ED5 C506 1EA9

2 7

[opensuse-programming] Cross Compiling for ARM with openSUSE 10.2
by Jörg Hermsdorf 30 Apr '07

30 Apr '07

Hi all, I tried to compile a linux kernel for my Palm Tungsten T with openSUSE 10.2 this weekend, but I didn't succeed. At first I fetched the kernel trunk from Montavista, they also provide a 'palmtt_defconfig' file to configure the kernel (2.6.20 I think) for the Tungsten T. Next I used Yast to search for and install the packages required to build the kernel for the ARM architecture. Well, I realized, that openSUSE 10.2 just ships a binutils package (cross-arm-binutils.rpm) for arm, but no corresponding cross-compiler. I wonder if the cross-arm-binutils package is of any use without a corresponding cross-compiler? Anyway, I did not want to compile and setup my own toolchain and hoped, that openSUSE ships a toolchain out of the box. While searching for this, I recongnized icecream and the fact that openSUSE 10.2 also ships the cross-arm-gcc-icecream-backend package. It seems like this was what I wanted. I examined the archive that was placed in /usr/share/icecream-envs/cross-arm-gcc-icecream-backend_x86_64.tar.gz and it seems like it holds the complete gcc toolchain to compile software for arm. After I installed an configured icecream, I configured the Montavista kernel: # cd ~/Documents/linux-omap/ # make CC=icecc palmtt_defconfig I checked the generated Makefile for ARCH and CROSS_COMPILE variables and they seem to be set correctly: ... SUBARCH := arm ... ARCH ?= $(SUBARCH) CROSS_COMPILE ?= arm-linux- ... Next I set ICECC_VERSION to the arm toolchain envrionment: # export ICECC_VERSION=/usr/share/icecream-envs/cross-arm-gcc-icecream-backend_x86_64.tar.gz Then I started to compile the kernel using: # make CC=icecc zImage But I get the following output: ... cc1: error: unrecognized command line option "-mlittle-endian" cc1: error: unrecognized command line option "-mapcs" cc1: error: unrecognized command line option "-mno-sched-prolog" cc1: error: unrecognized command line option "-mabi=apcs-gnu" arch/arm/kernel/asm-offsets.c:0: error: bad value (armv4t) for -march= switch arch/arm/kernel/asm-offsets.c:0: error: bad value (arm9tdmi) for -mtune= switch ... make[1]: *** [arch/arm/kernel/asm-offsets.s] Fehler 1 make: *** [prepare0] Fehler 2 Well, it seems like icecream takes the wrong cross-compiler (cc1). So how do I tell icecream that it should take the cc1 provided by the 'cross-arm-gcc-icecream-backend_x86_64.tar.gz' archive to compile the kernel? Setting it via the envrionment variable ICECC_VERSION does not work. Kind regards, Jörg. ------------------------------------------ Here's the full debug output of icecc: # make CC=icecc zImage ICECC[22178] 15:25:46: <building_local> ICECC[22178] 15:25:46: </building_local: 3ms> CHK include/linux/version.h make[1]: »include/asm-arm/mach-types.h« ist bereits aktualisiert. CHK include/linux/utsrelease.h CC arch/arm/kernel/asm-offsets.s ICECC[22255] 15:25:47: native /var/cache/icecream/native/957f0fa42db6b790e7360d4150fd7305.tar.gz ICECC[22255] 15:25:47: env: x86_64 '/var/cache/icecream/native/957f0fa42db6b790e7360d4150fd7305.tar.gz' ICECC[22255] 15:25:47: arch/arm/kernel/asm-offsets.c compiled 1 times on x86_64 ICECC[22255] 15:25:47: Have to use host 192.168.1.2:10245 - Job ID: 90 - environment: x86_64 got environment: true ICECC[22255] 15:25:47: <create CS channel> ICECC[22255] 15:25:47: </create CS channel: 7ms> ICECC[22255] 15:25:47: <send compile_file> ICECC[22255] 15:25:47: </send compile_file: 0ms> ICECC[22255] 15:25:47: <write_server_cpp from cpp> cc1: error: unrecognized command line option "-mlittle-endian" cc1: error: unrecognized command line option "-mapcs" cc1: error: unrecognized command line option "-mno-sched-prolog" cc1: error: unrecognized command line option "-mabi=apcs-gnu" arch/arm/kernel/asm-offsets.c:0: error: bad value (armv4t) for -march= switch arch/arm/kernel/asm-offsets.c:0: error: bad value (arm9tdmi) for -mtune= switch ICECC[22255] 15:25:47: </write_server_cpp from cpp: 2ms> ICECC[22255] 15:25:47: <wait for cpp> ICECC[22255] 15:25:47: </wait for cpp: 0ms> make[1]: *** [arch/arm/kernel/asm-offsets.s] Fehler 1 make: *** [prepare0] Fehler 2 --------------------------------------------------------------------- To unsubscribe, e-mail: opensuse-programming+unsubscribe(a)opensuse.org For additional commands, e-mail: opensuse-programming+help(a)opensuse.org

1 0

[opensuse-programming] GAS
by M Harris 26 Apr '07

26 Apr '07

I have been writing some routines in assembler using nasm & ld (compile and link) and gdb for debuggin. I am going to be converting these routines to the at&t syntax and pass them off to gcc(gas) so that I will have access to libc. I have been reading about the gcc compiler, and have been examining its .s files (at&t syntax) which are fed to GAS. GAS is supposed to ship with binutils, but I do not find it on my system as a separate utility... I use gcc every day so obviously gas is there somewhere. Can someone please tell me whether gcc still uses gas? ... is gas internal to gcc now? Also, I have read rumors that the latest gas code supports the intel syntax... true? Thanks -- Kind regards, M Harris <>< --------------------------------------------------------------------- To unsubscribe, e-mail: opensuse-programming+unsubscribe(a)opensuse.org For additional commands, e-mail: opensuse-programming+help(a)opensuse.org

3 5

[opensuse-programming] eclipse on SuSE 10.2 x86_64
by Andre Truter 26 Apr '07

26 Apr '07

I have been using eclipse for a while now. Since SuSE 9.x and it normally worked fine. But, since upgrading to 10.2, I could not get it to work. It gives me horrendous java errors that I don't understand. When I delete the $HOME/.eclipse directory, it starts up fine and things work. I import a project and I can use it. But, when I exit Eclipse, it looks like it is exiting, but the java process never die. I have to kill -9 it to get rid of it, and then my config is scewed up. If I try to start eclipse again, the java process exit with -1. I have to delete the $HOME/.eclipse directory again before I can continue. I have java 1.5.0 installed with patches (java-1_5_0-sun-1.5.0_update10-2.1) eclipse : --------- eclipse-platform-3.2.1-24 eclipse-scripts-3.2.1-24 eclipse-3.2.1-24 eclipse-jdt-3.2.1-24 eclipse-gtk2-3.2.1-24 I ran the latest eclipse updates, but it made no difference. I have googled for a solution, but it seems that google-fu is not good on this. Has anybody had a similar situation, or does anybody have any idea what could be wrong? Thanks -- Andre Truter | Software Consultant | Registered Linux user #185282 Jabber: andre.truter(a)gmail.com | http://www.trusoft.co.za ~ I am never wrong except when I am not right ~ --------------------------------------------------------------------- To unsubscribe, e-mail: opensuse-programming+unsubscribe(a)opensuse.org For additional commands, e-mail: opensuse-programming+help(a)opensuse.org

2 3

[opensuse-programming] GCC and make depends
by Jerry Feldman 25 Apr '07

25 Apr '07

The GCC -MM and -MMD options generate dependencies. However, when I try to build with Rogue Wave, not only to I get the local dependencies, but I also get all the STL and Rogue Wave dependencies. Note that RogueWave has its own version of the STL. The C++ file's header lists is: #include <iostream> #include <rw/tpordvec.h> #include <string> #include <iterator> #include "thisfile.h" The command line is: g++-I$(RWPATH)/include -I$(RWPATH) -D_RWCONFIG_rmd -MMD -c thisfile.cc What I should see in thisfile.d is: thisfile.o: thisfile.cc thisfile.h What I do see is: thisfile.o: thisfile.cc <RWPATH>/include/iostream \ <RWPATH>/include/istream \ .... thisfile.h According to the GCC man page, the -MM (and -MMD) options should exclude system headers (eg. those files included in angle brackets). I've tried a few different options. My partner in Toronto actually found that it was trying to compile some of the RogueWave files such as vector.cc since it is implicitly included as <RWPATH>/include/vector.cc into the .i file by the preprocessor. I'm not worried about the performance implications regarding the inclusion of all the header files, but I am concerned about make implicitly wanting to compile the RW .cc files. One possible solution is to create a script to remove all the lines with <RWPATH>, but I'm looking for a more elegant solution. -- Jerry Feldman <gaf(a)blu.org> Boston Linux and Unix user group http://www.blu.org PGP key id:C5061EA9 PGP Key fingerprint:053C 73EC 3AC1 5C44 3E14 9245 FB00 3ED5 C506 1EA9

1 0

[opensuse-programming] gpgme_set_passphrase_cb() does not work (but only on OpenSUSE?)
by Jules Colding 19 Apr '07

19 Apr '07

Hi, I've written a small keyring utility(*) to store passwords and such. It is using gpgme to interface with gnupg and works wonderfully on Gentoo and Fedora. However, on OpenSUSE 10.2 it doesn't. The problem is that gpgme_set_passphrase_cb() doesn't have any effect on that platform. I'm seeing pinentry-gtk-2 prompting me for a passphrase no matter what I do. This naturally leads to my keyring not working as the gpgme framework can't get the correct passphrase. Is there anything I can do to force gpgme to use the callback that I provides in gpgme_set_passphrase_cb() or is this something that has been hard coded into gnupg on OpenSUSE? Thanks a lot in advance, jules (*) Full source is here: http://www.omesc.com/content/downloads/dist/testing/brutus-snapshot.tar.bz2 Look in <brutus/idl/products/evolution/2.4/brutus-keyring/> for the keyring source. A small test program is in <../keyring-test/>. --------------------------------------------------------------------- To unsubscribe, e-mail: opensuse-programming+unsubscribe(a)opensuse.org For additional commands, e-mail: opensuse-programming+help(a)opensuse.org

1 0

Re: [opensuse-programming] C++ overload question.
by Nathan Hearn 04 Apr '07

04 Apr '07

Hi Jerry, You could set up a zero-length vector as a global object outside of a function. You could then use it as the default reference via: std::vector<double> zero(0); int foo(std::vector<double> & x, std::vector<double> & y=zero); However, this is rather unsafe, as a non-const y (and, therefore, zero) could be changed by foo (and the global, non-const zero could be changed by any part of the code). If zero can be changed, we can not guarantee that it will remain a zero-length vector. There are two ways to address this problem. First, if y is not altered within foo, then just make both zero and y const quantities: const std::vector<double> zero(0); int foo(std::vector<double> & x, const std::vector<double> & y=zero); The second way, as Verdi suggsted, is to overload foo: int foo(std::vector<double> & x, std::vector<double> & y); int foo(std::vector<double> & x) { std::vector<double> y; return foo(x, y); } Of these two options, the second is safest, as the y vector declared within foo can not interfere with any other parts of your code (i.e., it can not be used anywhere else, unlike a global). When possible, I try to find ways to reduce the number of functions I write (and thus maximize code reuse), but I do not see the advantage of using a single function definition with a default value as in the first case. (And, in effect, there is only one function foo in the second case, as both functions end up using the code within foo(x, y).) Have you come across a situation like this where overloading is not preferable? I would be interested in finding out. Take care! - Nathan -- Nathan C. Hearn nhearn(a)uchicago.edu ASC Flash Center Computational Physics Group University of Chicago On 4/3/07, Jerry Feldman <gaf(a)blu.org> wrote: > In porting some code I have a function (this is the original code). The > second parameter has a default. Array is a home grown class. > int foo(Array<aaa> &x, Array<bbb> *y = NULL); > > What I want is to replace the home grown array classes with the STL > vector class. I also want to pass them by reference. > int foo(std::vector<aaa> &x, std::vector<bbb> &y); > But, this requires 2 parameters. Using vector, I could overload foo > with an inline single argument function that would achieve the desired > result. > inline int foo(std::vector<aaa> &x) { > std::vector<bbb> y; // empty vector > return foo(x, y); > } > > Is there a good way where I could use a single function call, using > vector where the second parameter has a default value as the empty > vector. This would also achieve the desired result, but with a > performance hit if vector y is populated. > int foo(std::vector<aaa> &x, std::vector<bbb> *y =NULL); > -- > Jerry Feldman <gaf(a)blu.org> > Boston Linux and Unix user group > http://www.blu.org PGP key id:C5061EA9 > PGP Key fingerprint:053C 73EC 3AC1 5C44 3E14 9245 FB00 3ED5 C506 1EA9 --------------------------------------------------------------------- To unsubscribe, e-mail: opensuse-programming+unsubscribe(a)opensuse.org For additional commands, e-mail: opensuse-programming+help(a)opensuse.org

2 2

[opensuse-programming] C++ overload question.
by Jerry Feldman 04 Apr '07

04 Apr '07

In porting some code I have a function (this is the original code). The second parameter has a default. Array is a home grown class. int foo(Array<aaa> &x, Array<bbb> *y = NULL); What I want is to replace the home grown array classes with the STL vector class. I also want to pass them by reference. int foo(std::vector<aaa> &x, std::vector<bbb> &y); But, this requires 2 parameters. Using vector, I could overload foo with an inline single argument function that would achieve the desired result. inline int foo(std::vector<aaa> &x) { std::vector<bbb> y; // empty vector return foo(x, y); } Is there a good way where I could use a single function call, using vector where the second parameter has a default value as the empty vector. This would also achieve the desired result, but with a performance hit if vector y is populated. int foo(std::vector<aaa> &x, std::vector<bbb> *y =NULL); -- Jerry Feldman <gaf(a)blu.org> Boston Linux and Unix user group http://www.blu.org PGP key id:C5061EA9 PGP Key fingerprint:053C 73EC 3AC1 5C44 3E14 9245 FB00 3ED5 C506 1EA9

2 1