similar to: [LLVMdev] Which files does ./config.status -d produce?

Hi I've started to port LLVM to the MinGW Win32 platform. To successfully recognize the platform, following statements should be inserted into the configure script: ---------- Index: configure ================================================================== RCS file: /var/cvs/llvm/llvm/configure,v retrieving revision 1.114 diff -u -r1.114 configure --- configure 7 Sep 2004 18:04:45 -0000

Hi When compiling UnixLocalInferiorProcess.cpp, I get these errors: ----------------------- C:/MinGW/msys/local/projects/src/llvm/lib/Debugger/UnixLocalInferiorProcess.cpp:41:22: sys/wait.h: No such file or directory C:/MinGW/msys/local/projects/src/llvm/lib/Debugger/UnixLocalInferiorProcess.cpp: In member function `void <unnamed>::IP::startChild(llvm::Module*, const

Hi I'm compiling: /usr/local/src/llvm/lib/Support/SlowOperationInformer.cpp on MinGW. However, it stops complaining about that SIGALRM is undeclared: -------------------------- @ /usr/local/build/llvm/mklib --tag=disable-shared --silent --tag=CXX --mode=compile g++ -c -I/usr/local/build/llvm/lib/Support -I/usr/local/src/llvm/lib/Support -I/usr/local/build/llvm/include

Reid, >Well, if it doesn't break anything else, I'd fix the header file. The >standard type name is supposed to uint64_t not u_int64_t. I would just >change the header file to define both of them, something like: > >typedef u_int64_t uint64_t; > >You could do that in /usr/include/types.h, I've tried it and it doesn't break anything, but ... >or we could

This file (UnixLocalInferiorProcess.cpp) is due for porting and placement in lib/System but I haven't gotten there yet. If you come up with something that works on MINGW, please let me know. Thanks, Reid. Henrik Bach wrote: > Hi > > When compiling UnixLocalInferiorProcess.cpp, I get these errors: > ----------------------- >

Hi, I get below error: --------------------------------- In file included from /usr/local/build/llvm/lib/System/platform/Path.cpp:23, from /usr/local/src/llvm/lib/System/Path.cpp:27: /usr/local/build/llvm/lib/System/platform/../Unix/Path.cpp: In member function `bool llvm::sys::Path::create_directory(bool)': /usr/local/build/llvm/lib/System/platform/../Unix/Path.cpp:343:

Hi As shown below, the .\configure script found a version of stat.h: --------------------- configure:4382: checking for sys/stat.h configure:4399: gcc -c -g -O2 conftest.c >&5 configure:4402: $? = 0 configure:4405: test -s conftest.o configure:4408: $? = 0 configure:4419: result: yes --------------------- However, I got this error when compiling: --------------------- FileUtilities.cpp:

Hi Reid, I downloaded the latest source code and ran configure (I suppose it was that you did mean). The result is: --------------- checking build system type... i586-pc-interix3 --------------- Otherwise, please correct me. /Henrik --- Got Freedom? Software Freedom Day 2004 - 28th of August http://www.softwarefreedomday.org/ --- >From: Reid Spencer <reid at x10sys.com>

Yes, that's my understanding, too. But I'm not controlling where my library -ldbghelp is put when g++ is called. I just put it in the TOOLLINKOPTS variable in Makefile.config. Henrik >From: Reid Spencer <reid at x10sys.com> > >John Criswell wrote: >> >>It seems that the -L path options are specified before the LLVM libraries >>(libSystem and

Henrik Bach wrote: > Hi > > I'm compiling: /usr/local/src/llvm/lib/Support/SlowOperationInformer.cpp > on MinGW. However, it stops complaining about that SIGALRM is undeclared: Is there an alarm() syscall on MinGW? And if so, what signal does it send (according to the MinGW docs)? -- John T. > -------------------------- > @ /usr/local/build/llvm/mklib

Hi, Has some one committed these patches into CVS? Paolo, thanks for the patches. In the near future I'll try to compile them. I've installed STLport 6.4.2 with vc6. Henrik --- Got Freedom? Software Freedom Day 2004 - 28th of August http://www.softwarefreedomday.org/ --- >From: Paolo Invernizzi <arathorn at fastwebnet.it> >Reply-To: LLVM Developers Mailing List

Ultimately, this is another function that needs to go into lib/System. An alternate approach is to fork a thread, sleep, and when the thread wakes up, "ring the alarm". Reid. John Criswell wrote: > Henrik Bach wrote: > >> Hi >> >> I'm compiling: >> /usr/local/src/llvm/lib/Support/SlowOperationInformer.cpp on MinGW. >> However, it stops

Hi, I get this error: ------------------ ToolRunner.cpp:396: error: `SHLIBEXT' undeclared (first use this function) ToolRunner.cpp:396: error: (Each undeclared identifier is reported only once for each function it appears in.) ------------------ And the config.log shows that configure has detected that ld on my platform (Interix) doesn't support shared libraries: ------------------

Hi again Does cygwin support shared libraries. And if not, how did you port llvm on this issue? /Henrik >From: "Henrik Bach" <henrik_bach_llvm at hotmail.com> >Date: Tue, 27 Jul 2004 00:41:53 +0200 > >Hi, > >I get this error: >------------------ >ToolRunner.cpp:396: error: `SHLIBEXT' undeclared (first use this function) >ToolRunner.cpp:396:

There's simply no equivalent to signals on Windows. There is no way to asynchronously interrupt a thread's processing to execute some handler. The only thing you can asynchronously do to a thread is kill it, and that's generally frowned upon (who knows what critical sections it might be holding, etc...). Stuff like alarms is supposed to be done using the "event-driven"

>From: Jeff Cohen <jeffc at jolt-lang.org> >Date: Mon, 30 Aug 2004 21:46:42 -0700 >FileParser.tab.c: In function `int Fileparse()': >FileParser.tab.c:2043: error: syntax error before `goto' > >The offending lines bison generated are: > >/*----------------------------------------------------. >| yyerrlab1 -- error raised explicitly by an action. |

Reid, >From: Reid Spencer <reid at x10sys.com> >Date: Tue, 31 Aug 2004 11:08:37 -0700 > >Henrik, > >Since Interix is POSIX and SUS compatible, I've separated its Another posting on Interix forum, has revealed that Interix conforms to POSIX.1 and SUS. But, I don't think this will break your platform separation. /Henrik

Who takes action on this? >From: John Criswell <criswell at cs.uiuc.edu> >Date: Wed, 01 Sep 2004 15:16:53 -0500 > >Henrik Bach wrote: >>Reid, >> >>>Well, if it doesn't break anything else, I'd fix the header file. The >>>standard type name is supposed to uint64_t not u_int64_t. I would just >>>change the header file to define both

>From: Jeff Cohen <jeffc at jolt-lang.org> >Date: Tue, 14 Sep 2004 19:43:34 -0700 > >OK, I stand corrected: mingw can be used to create honest-to-goodness >Windows applications using Win32. But you're still using it as a Unix >emulator :) > OK, if POSIX.1 is unix emulation then we have that all *nix emulates all *nix and that's fine with me :) OK, I admit,

>From: Jeff Cohen <jeffc at jolt-lang.org> >Date: Wed, 15 Sep 2004 10:35:36 -0700 > >What's a "compiling mesh?" What I meant, was that there are some implicit defines in mingw (like __GCC) and vcX (like _MVC) but possibly also other unsupported? internal structures. As I stated earlier mingw should be win32 api compliant, but not for complicating matters. But