From: UnixOS2 Archive To: "UnixOS2 Archive" Date: Mon, 21 Oct 2002 04:39:30 EST-10EDT,10,-1,0,7200,3,-1,0,7200,3600 Subject: [UnixOS2_Archive] No. 350 ************************************************** Sunday 20 October 2002 Number 350 ************************************************** Subjects for today 1 Re: Icecast : John Poltorak 2 Icecast : Michel SUCH" 3 Re: gcc 3.0.4 : DWParsons at t-online.de (Dave Parsons) 4 Re: gcc 3.0.4 : John Poltorak 5 Re: gcc 3.0.4 : Henry Sobotka 6 Re: gcc 3.0.4 : DWParsons at t-online.de (Dave Parsons) 7 Re: gcc 3.0.4 : Henry Sobotka 8 Re: gcc 3.0.4 : DWParsons at t-online.de (Dave Parsons) **= Email 1 ==========================** Date: Mon, 21 Oct 2002 09:42:16 +0100 From: John Poltorak Subject: Re: Icecast On Mon, Oct 21, 2002 at 10:09:49AM +0100, Michel SUCH wrote: > Hi, > > There is a icecast 1.0.0 on hobbes. > Does anyone know if a newer port exists? It is supposed to be included with CoolFM although I have a seperate URL listed of:- http://os2.ru/projects/coolfm/icecastpack.zip This used to work, but no longer. Maybe some Russian subscribers here may have a better idea about where it went... > ---------------------------- > Michel SUCH TEAM OS/2 FRANCE > ICQ # 51654489 > -- John **= Email 2 ==========================** Date: Mon, 21 Oct 2002 10:09:49 +0100 (CET) From: "Michel SUCH" Subject: Icecast Hi, There is a icecast 1.0.0 on hobbes. Does anyone know if a newer port exists? ---------------------------- Michel SUCH TEAM OS/2 FRANCE ICQ # 51654489 **= Email 3 ==========================** Date: Mon, 21 Oct 2002 10:44:24 +0200 (CES) From: DWParsons at t-online.de (Dave Parsons) Subject: Re: gcc 3.0.4 On Sat, 19 Oct 2002 19:56:59 +0100 (CET), Michel SUCH wrote: > Well, > > I have downloaded 3.04 and did not feel it compiles with my current > environment. > So I don't use it. > What problems did you find? I tried it here, both as the binary based upon 2.7.2.? and I also recompiled it with 2.8.1, but in both cases it fails when I try to compile another program with it with the error message Undefined symbol __eh_frame_init in c.a(ctor.o). I get the same error when I try make bootstrap also. Setup problem or incomplete port? Anyone know whether Mikulas reads this mailing list? Dave **= Email 4 ==========================** Date: Mon, 21 Oct 2002 11:21:27 +0100 From: John Poltorak Subject: Re: gcc 3.0.4 **= Email 5 ==========================** Date: Mon, 21 Oct 2002 13:04:06 -0400 From: Henry Sobotka Subject: Re: gcc 3.0.4 Dave Parsons wrote: > > I tried it here, both as the binary based upon 2.7.2.? and I > also recompiled it with 2.8.1, but in both cases it fails > when I try to compile another program with it with the error > message Undefined symbol __eh_frame_init in c.a(ctor.o). > > I get the same error when I try make bootstrap also. > > Setup problem or incomplete port? Exception-handling is off by default in older versions of gcc, and on in newer ones (2.95+ ?). You might try compiling the other program with -fno-exeptions or turning on exceptions when building/bootstrapping with 2.8.1 (or use pgcc instead where they're on by default). In any case, I wouldn't lose sleep over 3.0.x as there's yet another new ABI with 3.2... h~ **= Email 6 ==========================** Date: Mon, 21 Oct 2002 16:30:25 +0200 (CES) From: DWParsons at t-online.de (Dave Parsons) Subject: Re: gcc 3.0.4 On Mon, 21 Oct 2002 11:21:27 +0100, John Poltorak wrote: > Nothing that I can see;-( Dave **= Email 7 ==========================** Date: Mon, 21 Oct 2002 18:32:10 -0400 From: Henry Sobotka Subject: Re: gcc 3.0.4 Dave Parsons wrote: > > I don't think that is the point. It currently fails due to > undefined symbols. __eh_frame_init is one of those behind-the-scenes symbols deployed when exception-handling is on. -fno-exceptions normally removes the demand for it. To me this suggests the undefined-symbols failure may be related to the on/offness of eh. h~ **= Email 8 ==========================** Date: Mon, 21 Oct 2002 22:37:35 +0200 (CES) From: DWParsons at t-online.de (Dave Parsons) Subject: Re: gcc 3.0.4 On Mon, 21 Oct 2002 13:04:06 -0400, Henry Sobotka wrote: > Dave Parsons wrote: > > > > I tried it here, both as the binary based upon 2.7.2.? and I > > also recompiled it with 2.8.1, but in both cases it fails > > when I try to compile another program with it with the error > > message Undefined symbol __eh_frame_init in c.a(ctor.o). > > > > > I get the same error when I try make bootstrap also. > > > > Setup problem or incomplete port? > > Exception-handling is off by default in older versions of gcc, and on in > newer ones (2.95+ ?). You might try compiling the other program with > -fno-exeptions or turning on exceptions when building/bootstrapping with > 2.8.1 (or use pgcc instead where they're on by default). > > In any case, I wouldn't lose sleep over 3.0.x as there's yet another new > ABI with 3.2... > > h~ > I don't think that is the point. It currently fails due to undefined symbols. I am aware of the changing ABI but I have no direct interest in C++. We have enough c/c++ compilers for my needs. My concern is with with making the jump from Gnat 3.14p compiled with gcc 2.8.1 to 3.15 compiled with a 3.x version of GCC. I was interested to see what Mikulas had done to see if it would help me to get a GCC 3.1 (or probably now 3.3) version working. So far I have been trying to build upon Andrew Zabolotny's 3.0.2 changes and incorporate them into a GCC 3.x which includes Gnat. After many problems with incompatible tool sets I am now able to build GCC 3.x but have not been able to get it to bootstrap. When compiling stage 2 it reverts to using gcc 2.8.1 instead of 3.1 and fails. Dave