-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
Hey all,
I've been using iftop on Solaris for some time now and love what it is able to do -- there are few programs that make things quite this easy to decipher.
I'd like to use this also on an HP-UX box, but I can't get it built. It balks in the pthreads department, but AFAIK HP-UX now supports pthreads properly. Any assistance would be wonderful. config.log is as follows:
configure:6294: checking POSIX threads compilation configure:6309: cc -o conftest -I/usr/local/include - -I/usr/local/include -L/u sr/local/lib conftest.c -lpcap -lm -lcurses >&5 configure:6312: $? = 0 configure:6315: test -s conftest configure:6318: $? = 0 configure:6347: result: CFLAGS= and LIBS= configure:6349: checking POSIX threads usability configure:6359: cc -o conftest -I/usr/local/include - -I/usr/local/include -L/u sr/local/lib conftest.c -lpcap -lm -lcurses >&5 configure:6362: $? = 0 configure:6364: ./conftest Function is not available configure:6367: $? = 255 configure: program exited with status 255 configure: failed program was: /* * pthread.c: * Tiny test program to see whether POSIX threads work. */ <source code snipped because you know what it says already>
There's not a whole lot of info here -- Function not available could refer to quite a number of functions.
Thanks for any time you spend on this, and for the great product that at least I can use elsewhere.
- -- ---- _ _ _ _ ___ _ _ _ |Y#| | | |/| | \ |\ | | |Ryan Novosielski - Systems Programmer III |$&| |__| | | |__/ | | _| |novosirj@umdnj.edu - 973/972.0922 (2-0922) __/ Univ. of Med. and Dent.|IST/AST - NJMS Medical Science Bldg - C630
On Thu, Apr 19, 2007 at 10:37:22AM -0400, Ryan Novosielski wrote: | | Hey all, | | I've been using iftop on Solaris for some time now and love what it is | able to do -- there are few programs that make things quite this easy to | decipher. | | I'd like to use this also on an HP-UX box, but I can't get it built. It | balks in the pthreads department, but AFAIK HP-UX now supports pthreads | properly. Any assistance would be wonderful. config.log is as follows: | | configure:6294: checking POSIX threads compilation | configure:6309: cc -o conftest -I/usr/local/include | - -I/usr/local/include -L/u | sr/local/lib conftest.c -lpcap -lm -lcurses >&5 | configure:6312: $? = 0 | configure:6315: test -s conftest | configure:6318: $? = 0 | configure:6347: result: CFLAGS= and LIBS= | configure:6349: checking POSIX threads usability | configure:6359: cc -o conftest -I/usr/local/include | - -I/usr/local/include -L/u | sr/local/lib conftest.c -lpcap -lm -lcurses >&5 | configure:6362: $? = 0 | configure:6364: ./conftest | Function is not available | configure:6367: $? = 255 | configure: program exited with status 255 | configure: failed program was: | /* | * pthread.c: | * Tiny test program to see whether POSIX threads work. | */ | <source code snipped because you know what it says already> | | There's not a whole lot of info here -- Function not available could | refer to quite a number of functions. | | Thanks for any time you spend on this, and for the great product that at | least I can use elsewhere.
Are you using gcc on HP-UX, or are you using a native HP-UX C/C++ compiler?
I don't think we've had any users attempt to build iftop on HP-UX before. We'll probably need someone with access to an HP-UX box and some autoconf-fu to tweak the configure.in file so that it can cope with the way an HP-UX system is laid out.
For example, I'd guess we'd need a -lthreads or some such there in order to pick up the threads library, but that will depend on just how HP-UX is laid out.
Jon
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
Jonathan Abbey wrote:
On Thu, Apr 19, 2007 at 10:37:22AM -0400, Ryan Novosielski wrote: | | Hey all, | | I've been using iftop on Solaris for some time now and love what it is | able to do -- there are few programs that make things quite this easy to | decipher. | | I'd like to use this also on an HP-UX box, but I can't get it built. It | balks in the pthreads department, but AFAIK HP-UX now supports pthreads | properly. Any assistance would be wonderful. config.log is as follows: | | configure:6294: checking POSIX threads compilation | configure:6309: cc -o conftest -I/usr/local/include | - -I/usr/local/include -L/u | sr/local/lib conftest.c -lpcap -lm -lcurses >&5 | configure:6312: $? = 0 | configure:6315: test -s conftest | configure:6318: $? = 0 | configure:6347: result: CFLAGS= and LIBS= | configure:6349: checking POSIX threads usability | configure:6359: cc -o conftest -I/usr/local/include | - -I/usr/local/include -L/u | sr/local/lib conftest.c -lpcap -lm -lcurses >&5 | configure:6362: $? = 0 | configure:6364: ./conftest | Function is not available | configure:6367: $? = 255 | configure: program exited with status 255 | configure: failed program was: | /* | * pthread.c: | * Tiny test program to see whether POSIX threads work. | */ | <source code snipped because you know what it says already> | | There's not a whole lot of info here -- Function not available could | refer to quite a number of functions. | | Thanks for any time you spend on this, and for the great product that at | least I can use elsewhere.
Are you using gcc on HP-UX, or are you using a native HP-UX C/C++ compiler?
I am using HP's AnsiC compilers. I don't have GCC on this particular box, but I do have other HP-UX machines that do have it. I'm somewhat skeptical that that would make a difference, as it seems the compilers tend to bomb on syntactical errors when it's compiler related, not libraries.
I don't think we've had any users attempt to build iftop on HP-UX before. We'll probably need someone with access to an HP-UX box and some autoconf-fu to tweak the configure.in file so that it can cope with the way an HP-UX system is laid out.
Well, you have someone with root access on many HP-UX boxes, but I don't really have any autoconf-fu. Autoconf doesn't really seem that difficult to me, so I'd be willing to give it a shot if I knew the kind of stuff I'd be trying...
For example, I'd guess we'd need a -lthreads or some such there in order to pick up the threads library, but that will depend on just how HP-UX is laid out.
How can I be of best assistance in this particular instance?
- -- ---- _ _ _ _ ___ _ _ _ |Y#| | | |/| | \ |\ | | |Ryan Novosielski - Systems Programmer III |$&| |__| | | |__/ | | _| |novosirj@umdnj.edu - 973/972.0922 (2-0922) __/ Univ. of Med. and Dent.|IST/AST - NJMS Medical Science Bldg - C630
On Mon, Apr 23, 2007 at 11:17:22AM -0400, Ryan Novosielski wrote: | | I am using HP's AnsiC compilers. I don't have GCC on this particular | box, but I do have other HP-UX machines that do have it. I'm somewhat | skeptical that that would make a difference, as it seems the compilers | tend to bomb on syntactical errors when it's compiler related, not | libraries.
Yes, but different compilers do have the potential for taking different command line arguments to specify where those libraries are found, and etc.
I also asked because the error message you included in your first email wasn't familiar to me. I suppose the 'functions not found' message was actually generated from the configure shell script. There should be a file 'config.log' which actually includes the detritus from the builds performed by the configure script as it attempts to verify the environment.
You may find a more informative message in that file, describing what function or functions could not be found at the link stage.
| > I don't think we've had any users attempt to build iftop on HP-UX | > before. We'll probably need someone with access to an HP-UX box and | > some autoconf-fu to tweak the configure.in file so that it can cope | > with the way an HP-UX system is laid out. | | Well, you have someone with root access on many HP-UX boxes, but I don't | really have any autoconf-fu. Autoconf doesn't really seem that difficult | to me, so I'd be willing to give it a shot if I knew the kind of stuff | I'd be trying...
You'd be modifying the configure.in file, to specify new paths or new library names that should be searched for, and used in the configure tests.
At least, that's my best guess. Knowing what that config.log file said would be very helpful.
| > For example, I'd guess we'd need a -lthreads or some such there in | > order to pick up the threads library, but that will depend on just how | > HP-UX is laid out. | | How can I be of best assistance in this particular instance?
Are there any libraries in your system's standard library paths that reference the word 'thread' in any form? libpthreads, libposixthreads, etc.?
Jon
On Mon, Apr 23, 2007 at 10:31:49AM -0500, Jonathan Abbey wrote: | | You'd be modifying the configure.in file, to specify new paths or new | library names that should be searched for, and used in the configure | tests.
Note: once you've modified the configure.in file, you'd need to have the autoconf tools installed on your system, so that you could rebuild the configure script.
If you don't have autoconf installed, about the best you could do would be to provide details to someone who did have autoconf, so that they could generate modifications to the configure definition for you to test.
Jon
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
Jonathan Abbey wrote:
On Mon, Apr 23, 2007 at 11:17:22AM -0400, Ryan Novosielski wrote: | | I am using HP's AnsiC compilers. I don't have GCC on this particular | box, but I do have other HP-UX machines that do have it. I'm somewhat | skeptical that that would make a difference, as it seems the compilers | tend to bomb on syntactical errors when it's compiler related, not | libraries.
Yes, but different compilers do have the potential for taking different command line arguments to specify where those libraries are found, and etc.
Yeah, I'm not really arguing with you, I just have a hunch on this one. I can try this on a gcc box also, though, but it seems like if the flags are defined properly, the compiler will usually go through the compile no problem or error out.
I also asked because the error message you included in your first email wasn't familiar to me. I suppose the 'functions not found' message was actually generated from the configure shell script. There should be a file 'config.log' which actually includes the detritus from the builds performed by the configure script as it attempts to verify the environment.
The output that is in this thread is from configure.log.
You may find a more informative message in that file, describing what function or functions could not be found at the link stage.
| > I don't think we've had any users attempt to build iftop on HP-UX | > before. We'll probably need someone with access to an HP-UX box and | > some autoconf-fu to tweak the configure.in file so that it can cope | > with the way an HP-UX system is laid out. | | Well, you have someone with root access on many HP-UX boxes, but I don't | really have any autoconf-fu. Autoconf doesn't really seem that difficult | to me, so I'd be willing to give it a shot if I knew the kind of stuff | I'd be trying...
You'd be modifying the configure.in file, to specify new paths or new library names that should be searched for, and used in the configure tests.
At least, that's my best guess. Knowing what that config.log file said would be very helpful.
Would you like the whole file attached to the list? I generally try to avoid that sort of thing as I feel like I'm being a pest, but if that would be the best course, I'm happy to do so.
| > For example, I'd guess we'd need a -lthreads or some such there in | > order to pick up the threads library, but that will depend on just how | > HP-UX is laid out. | | How can I be of best assistance in this particular instance?
Are there any libraries in your system's standard library paths that reference the word 'thread' in any form? libpthreads, libposixthreads, etc.?
Yup, from what I can see looks like everything that would be needed is here. That's actually why I didn't know how to proceed -- nothing is in a non-standard location, so there are no -L or -I flags that I figure I'd have needed to add. Note HP-UX uses .sl for .so files.
/usr/lib/libpthread.1 /usr/lib/libp/libpthread.a /usr/lib/libpthread.a /usr/lib/libpthread.sl /usr/lib/libpthread_tr.1 /usr/lib/libpthread_tr.a
- From my cursory browsing on the web, 11.11 definitely does have a reasonably good POSIX threads library set (wasn't the case in some of the older releases -- not so sure about 11.00).
- -- ---- _ _ _ _ ___ _ _ _ |Y#| | | |/| | \ |\ | | |Ryan Novosielski - Systems Programmer III |$&| |__| | | |__/ | | _| |novosirj@umdnj.edu - 973/972.0922 (2-0922) __/ Univ. of Med. and Dent.|IST/AST - NJMS Medical Science Bldg - C630