search for: writetodisk

...ranlib. The native symbol table is supposed to start at offset 0x44. >> Is this data supposed to be copied out of the original file, > That's one solution but it defeats the purpose/efficiency of the mmap. Well, only the native symbol table would need to be copied, and only in the writeToDisk function, before the file gets invalidated. >> another temporary supposed to be created and then the original could >> be replaced using a file move operation instead? > Another temporary file would be even slower than copying the memory in > memory. I'm not so sure about...

...const&, std::basic_ofstream<char, std::char_traits<char> >&, bool, bool, bool) ( this=0x8356088, member=@0x8356180, ARFile=@0xbfffd630, CreateSymbolTable=false, TruncateNames=false, ShouldCompress=false) at ArchiveWriter.cpp:294 #5 0x0829d297 in llvm::Archive::writeToDisk(bool, bool, bool) ( this=0x8356088, CreateSymbolTable=true, TruncateNames=false, Compress=false) at ArchiveWriter.cpp:439 #6 0x081a5618 in main (argc=2, argv=0xbfffd9b4) at llvm-ranlib.cpp:76 #7 0x42015574 in __libc_start_main () from /lib/tls/libc.so.6 At frame #4 (Archive::writeMemb...

...b) n 686 ThrowError("Can't move '" + path + (gdb) bt #0 llvm::sys::Path::renamePathOnDisk(llvm::sys::Path const&) (this=0x22fc00, newName=@0x22fd20) at C:/msys/1.0/home/llvm_home/llvm-build/../llvm/lib/System/Win32/Path.inc:686 #1 0x004871c3 in llvm::Archive::writeToDisk(bool, bool, bool) (this=0x3d5110, CreateSymbolTable=true, TruncateNames=false, Compress=false) at C:/msys/1.0/home/llvm_home/llvm-build/../llvm/lib/Bytecode/Archive/ArchiveWriter.cpp:461 #2 0x00403d4d in doReplaceOrInsert() () at C:/msys/1.0/home/llvm_home/llvm-build/../llvm/tools/llvm...

...here the error message comes from) and llvm-ar returns error code 2 (which means an error exception was generated). We just need to find out why Win32/Path.inc is throwing an exception on rename. Possibly the temporary file has not been closed at the point of the rename. See the logic in Archive::writeToDisk in lib/Bytecode/Archive/ArchiveWriter.cpp for the details. The logic looks correct to me. The Archive file is closed and then renamed. Reid. -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 189 bytes Desc: Thi...

...ce stay valid, since the original mmap is still valid. One disadvantage of this approach is that if CreateSymbolTable is false, it will be slightly more expensive as it will build the archive in memory before writing it to disk, instead of just writing it out to disk directly. > Also, since writeToDisk() invalidates the mapped file that was > originally used to create the archive, shouldn't this function also > unmap that file and erase all its members? This would prevent and > further bugs like this from happening. This is not necessary using the replace method described above....

...an't move '" + path + > (gdb) bt > #0 llvm::sys::Path::renamePathOnDisk(llvm::sys::Path const&) > (this=0x22fc00, > newName=@0x22fd20) > at > C:/msys/1.0/home/llvm_home/llvm-build/../llvm/lib/System/Win32/Path.inc:686 > #1 0x004871c3 in llvm::Archive::writeToDisk(bool, bool, bool) > (this=0x3d5110, > CreateSymbolTable=true, TruncateNames=false, Compress=false) > at > C:/msys/1.0/home/llvm_home/llvm-build/../llvm/lib/Bytecode/Archive/ArchiveWriter.cpp:461 > #2 0x00403d4d in doReplaceOrInsert() () > at > C:/msys/1.0/home/llvm...

>From: Jeff Cohen Date: Wed, 26 Jan 2005 19:47:44 -0800 > >Fixed. Yes, now it isn't the path. I've recorded this trace: ------------------------- llvm[3]: Building Debug Bytecode Archive libc.bca /bin/rm -f /C/projects/build/MinGW/llvm-4-1/Debug/lib/libc.bca /C/projects/build/MinGW/llvm-4-1/Debug/bin/llvm-ar rcsf /C/projects/build/MinGW/llvm-4-1/Debug/lib/libc.bca

...d::basic_ofstream<char, std::char_traits<char> >&, bool, bool, bool) ( > this=0x8356088, member=@0x8356180, ARFile=@0xbfffd630, > CreateSymbolTable=false, TruncateNames=false, ShouldCompress=false) > at ArchiveWriter.cpp:294 > #5 0x0829d297 in llvm::Archive::writeToDisk(bool, bool, bool) ( > this=0x8356088, CreateSymbolTable=true, TruncateNames=false, > Compress=false) at ArchiveWriter.cpp:439 > #6 0x081a5618 in main (argc=2, argv=0xbfffd9b4) at llvm-ranlib.cpp:76 > #7 0x42015574 in __libc_start_main () from /lib/tls/libc.so.6 > > &gt...

Evan Jones wrote: > On Nov 22, 2005, at 19:10, Reid Spencer wrote: > >> 1. What is the path name associated with TmpArchive? If its the same >> as the path name associated with archPath then that's a bug, probably >> introduced when Path::makeUnique is called from >> Path::createTemporaryFileOnDisk which is called from line 377 of >> ArchiveWriter.cpp.

...ge comes from) and llvm-ar returns error code 2 (which means >an error exception was generated). We just need to find out why >Win32/Path.inc is throwing an exception on rename. Possibly the >temporary file has not been closed at the point of the rename. > >See the logic in Archive::writeToDisk in >lib/Bytecode/Archive/ArchiveWriter.cpp for the details. The logic looks >correct to me. The Archive file is closed and then renamed. > >Reid. > > > >------------------------------------------------------------------------ > >_____________________________________...

On Nov 22, 2005, at 19:10, Reid Spencer wrote: > 1. What is the path name associated with TmpArchive? If its the same > as the path name associated with archPath then that's a bug, probably > introduced when Path::makeUnique is called from > Path::createTemporaryFileOnDisk which is called from line 377 of > ArchiveWriter.cpp. This does not appear to be the problem. I

...mbolTable is false, > it will be slightly more expensive as it will build the archive in > memory before writing it to disk, instead of just writing it out to disk > directly. You can always use CreateSymbolTable as a flag to determine which mechanism to use. > >> Also, since writeToDisk() invalidates the mapped file that was >> originally used to create the archive, shouldn't this function also >> unmap that file and erase all its members? This would prevent and >> further bugs like this from happening. > > > This is not necessary using the repl...