Issue #21498 has been updated by alanwu (Alan Wu). Assignee set to windows ---------------------------------------- Bug #21498: Windows - Ruby Overrides C Library APIs thus breaking them https://bugs.ruby-lang.org/issues/21498#change-114270 * Author: cfis (Charlie Savage) * Status: Open * Assignee: windows * ruby -v: *all* on windows (testing with ruby 3.4.3 (2025-04-14 revision d0b7e5b6a0) +PRISM [x64-mswin64_140]) * Backport: 3.2: UNKNOWN, 3.3: UNKNOWN, 3.4: UNKNOWN ---------------------------------------- I am trying to wrap a simple C++ library, https://github.com/baderouaich/BitmapPlusPlus, as a Ruby extension. However when I use the extension to write a bitmap to disk the bitmap is corrupted. This is because the library uses std::ofstream which eventually uses the C API fclose to write the final bytes to the bitmap file and then closes it. The problem is that Ruby overrides `fclose` and replaces it with `rb_w32_fclose`. It then *exports* its custom version from from x64-vcruntime140-ruby340.dll. And the exported version is broken (at least from the standpoint of the C standard library). Note this has been a long standing issue. The first report I see is from 2013: https://bugs.ruby-lang.org/issues/8569 More recently in 2020 (which explains the issue very well): https://github.com/NREL/OpenStudio/issues/3942#issuecomment-610673401 I understand that Ruby is trying to provide a platform independent API. But the problem is this solution breaks any third party libraries that rely on these C APIs (which of course are very common). And there is no good workaround (see https://github.com/NREL/OpenStudio/issues/3942#issuecomment-611072774). So would it be possible for Ruby to stop exporting custom versions of basic C APIs? The code that does it is here: https://github.com/ruby/ruby/blob/master/win32/mkexports.rb#L41 Ruby of course could still use its custom versions of `fclose`, read, write etc within ruby.exe and x64-vcruntime140-ruby340.dll. But they should not be exported from x64-vcruntime140-ruby340.dll and thus be off limits to extensions. If a case comes up where an extension really needs access to rb_w32_fclose instead of fclose then an extension developer can use an `#ifdef _WIN32` to do so and work across platforms. That at least puts the developer in control versus now where I don't see any way I can wrap the bitmap library as a Ruby extension. From my experience the biggest problem is the replacing of fclose with rb_w32_fclose. This is the list of generated overrides: FD_CLR=rb_w32_fdclr FD_ISSET=rb_w32_fdisset Sleep=rb_w32_Sleep accept=rb_w32_accept access=rb_w32_uaccess bind=rb_w32_bind close=rb_w32_close connect=rb_w32_connect dup2=rb_w32_dup2 fclose=rb_w32_fclose fstat=rb_w32_fstati128 get_osfhandle=rb_w32_get_osfhandle getcwd=rb_w32_ugetcwd getenv=rb_w32_ugetenv gethostbyaddr=rb_w32_gethostbyaddr gethostbyname=rb_w32_gethostbyname gethostname=rb_w32_gethostname getpeername=rb_w32_getpeername getpid=rb_w32_getpid getppid=rb_w32_getppid getprotobyname=rb_w32_getprotobyname getprotobynumber=rb_w32_getprotobynumber getservbyname=rb_w32_getservbyname getservbyport=rb_w32_getservbyport getsockname=rb_w32_getsockname getsockopt=rb_w32_getsockopt inet_ntop=rb_w32_inet_ntop inet_pton=rb_w32_inet_pton ioctlsocket=rb_w32_ioctlsocket isatty=rb_w32_isatty listen=rb_w32_listen lseek=rb_w32_lseek lstat=rb_w32_ulstati128 mkdir=rb_w32_umkdir mmap=rb_w32_mmap mprotect=rb_w32_mprotect munmap=rb_w32_munmap pipe=rb_w32_pipe pread=rb_w32_pread pwrite=rb_w32_pwrite read=rb_w32_read recv=rb_w32_recv recvfrom=rb_w32_recvfrom rename=rb_w32_urename rmdir=rb_w32_urmdir select=rb_w32_select send=rb_w32_send sendto=rb_w32_sendto setsockopt=rb_w32_setsockopt shutdown=rb_w32_shutdown socket=rb_w32_socket stati128=rb_w32_ustati128 strcasecmp=msvcrt.stricmp strerror=rb_w32_strerror strncasecmp=msvcrt.strnicmp times=rb_w32_times unlink=rb_w32_uunlink utime=rb_w32_uutime utimensat=rb_w32_uutimensat utimes=rb_w32_uutimes write=rb_w32_write -- https://bugs.ruby-lang.org/

Issue #21498 has been updated by cfis (Charlie Savage). Thanks @alanwu. I added comments to the commit in Github, but reproducting here: This MR will result in the Ruby dll exporting the symbol _fclose instead of fclose. The exporting happens here: https://github.com/ruby/ruby/blob/c6c92bdce3ad2714a3da5f1b1d7f083b2b579be3/w... Microsoft tends to like to put _ in front of their c function names. In this case, fclose is ok but _flcoseall would not be. See https://learn.microsoft.com/en-us/cpp/c-runtime-library/reference/fclose-fcl.... So perhaps that regex should be updated to not export functions that have __ in their name? Or, should `rb_w32_fclose` be entirely removed? If standard fclose works on windows now, then maybe it no longer serves a useful purpose? ---------------------------------------- Bug #21498: Windows - Ruby Overrides C Library APIs thus breaking them https://bugs.ruby-lang.org/issues/21498#change-115100 * Author: cfis (Charlie Savage) * Status: Open * Assignee: windows * ruby -v: *all* on windows (testing with ruby 3.4.3 (2025-04-14 revision d0b7e5b6a0) +PRISM [x64-mswin64_140]) * Backport: 3.2: UNKNOWN, 3.3: UNKNOWN, 3.4: UNKNOWN ---------------------------------------- I am trying to wrap a simple C++ library, https://github.com/baderouaich/BitmapPlusPlus, as a Ruby extension. However when I use the extension to write a bitmap to disk the bitmap is corrupted. This is because the library uses std::ofstream which eventually uses the C API fclose to write the final bytes to the bitmap file and then closes it. The problem is that Ruby overrides `fclose` and replaces it with `rb_w32_fclose`. It then *exports* its custom version from from x64-vcruntime140-ruby340.dll. And the exported version is broken (at least from the standpoint of the C standard library). Note this has been a long standing issue. The first report I see is from 2013: https://bugs.ruby-lang.org/issues/8569 More recently in 2020 (which explains the issue very well): https://github.com/NREL/OpenStudio/issues/3942#issuecomment-610673401 I understand that Ruby is trying to provide a platform independent API. But the problem is this solution breaks any third party libraries that rely on these C APIs (which of course are very common). And there is no good workaround (see https://github.com/NREL/OpenStudio/issues/3942#issuecomment-611072774). So would it be possible for Ruby to stop exporting custom versions of basic C APIs? The code that does it is here: https://github.com/ruby/ruby/blob/master/win32/mkexports.rb#L41 Ruby of course could still use its custom versions of `fclose`, read, write etc within ruby.exe and x64-vcruntime140-ruby340.dll. But they should not be exported from x64-vcruntime140-ruby340.dll and thus be off limits to extensions. If a case comes up where an extension really needs access to rb_w32_fclose instead of fclose then an extension developer can use an `#ifdef _WIN32` to do so and work across platforms. That at least puts the developer in control versus now where I don't see any way I can wrap the bitmap library as a Ruby extension. From my experience the biggest problem is the replacing of fclose with rb_w32_fclose. This is the list of generated overrides: FD_CLR=rb_w32_fdclr FD_ISSET=rb_w32_fdisset Sleep=rb_w32_Sleep accept=rb_w32_accept access=rb_w32_uaccess bind=rb_w32_bind close=rb_w32_close connect=rb_w32_connect dup2=rb_w32_dup2 fclose=rb_w32_fclose fstat=rb_w32_fstati128 get_osfhandle=rb_w32_get_osfhandle getcwd=rb_w32_ugetcwd getenv=rb_w32_ugetenv gethostbyaddr=rb_w32_gethostbyaddr gethostbyname=rb_w32_gethostbyname gethostname=rb_w32_gethostname getpeername=rb_w32_getpeername getpid=rb_w32_getpid getppid=rb_w32_getppid getprotobyname=rb_w32_getprotobyname getprotobynumber=rb_w32_getprotobynumber getservbyname=rb_w32_getservbyname getservbyport=rb_w32_getservbyport getsockname=rb_w32_getsockname getsockopt=rb_w32_getsockopt inet_ntop=rb_w32_inet_ntop inet_pton=rb_w32_inet_pton ioctlsocket=rb_w32_ioctlsocket isatty=rb_w32_isatty listen=rb_w32_listen lseek=rb_w32_lseek lstat=rb_w32_ulstati128 mkdir=rb_w32_umkdir mmap=rb_w32_mmap mprotect=rb_w32_mprotect munmap=rb_w32_munmap pipe=rb_w32_pipe pread=rb_w32_pread pwrite=rb_w32_pwrite read=rb_w32_read recv=rb_w32_recv recvfrom=rb_w32_recvfrom rename=rb_w32_urename rmdir=rb_w32_urmdir select=rb_w32_select send=rb_w32_send sendto=rb_w32_sendto setsockopt=rb_w32_setsockopt shutdown=rb_w32_shutdown socket=rb_w32_socket stati128=rb_w32_ustati128 strcasecmp=msvcrt.stricmp strerror=rb_w32_strerror strncasecmp=msvcrt.strnicmp times=rb_w32_times unlink=rb_w32_uunlink utime=rb_w32_uutime utimensat=rb_w32_uutimensat utimes=rb_w32_uutimes write=rb_w32_write -- https://bugs.ruby-lang.org/

Issue #21498 has been updated by cfis (Charlie Savage). The universal crt was introduced by Microsoft in 2015 (https://learn.microsoft.com/en-us/cpp/porting/upgrade-your-code-to-the-unive...) and Ruby for Windows has used it since Ruby 3.1 in 2021 (https://rubyinstaller.org/2021/12/31/rubyinstaller-3.1.0-1-released.html). According to Microsoft it "conforms closely to the ISO C99 standard". At this point in time, could Ruby drop some of its overrides on Windows. The UCRT function list is here: https://learn.microsoft.com/en-us/cpp/c-runtime-library/reference/crt-alphab... The ones the UCRT implements that Ruby overrides are: access=rb_w32_uaccess close=rb_w32_close dup2=rb_w32_dup2 fclose=rb_w32_fclose fstat=rb_w32_fstati128 (_fstat) get_osfhandle=rb_w32_get_osfhandle (_get_osfhandle) getpid=rb_w32_getpid isatty=rb_w32_isatty lseek=rb_w32_lseek mkdir=rb_w32_umkdir pipe=rb_w32_pipe (_pipe) read=rb_w32_read rename=rb_w32_urename rmdir=rb_w32_urmdir strerror=rb_w32_strerror unlink=rb_w32_uunlink utime=rb_w32_uutime (_utime) write=rb_w32_write ---------------------------------------- Bug #21498: Windows - Ruby Overrides C Library APIs thus breaking them https://bugs.ruby-lang.org/issues/21498#change-115102 * Author: cfis (Charlie Savage) * Status: Open * Assignee: windows * ruby -v: *all* on windows (testing with ruby 3.4.3 (2025-04-14 revision d0b7e5b6a0) +PRISM [x64-mswin64_140]) * Backport: 3.2: UNKNOWN, 3.3: UNKNOWN, 3.4: UNKNOWN ---------------------------------------- I am trying to wrap a simple C++ library, https://github.com/baderouaich/BitmapPlusPlus, as a Ruby extension. However when I use the extension to write a bitmap to disk the bitmap is corrupted. This is because the library uses std::ofstream which eventually uses the C API fclose to write the final bytes to the bitmap file and then closes it. The problem is that Ruby overrides `fclose` and replaces it with `rb_w32_fclose`. It then *exports* its custom version from from x64-vcruntime140-ruby340.dll. And the exported version is broken (at least from the standpoint of the C standard library). Note this has been a long standing issue. The first report I see is from 2013: https://bugs.ruby-lang.org/issues/8569 More recently in 2020 (which explains the issue very well): https://github.com/NREL/OpenStudio/issues/3942#issuecomment-610673401 I understand that Ruby is trying to provide a platform independent API. But the problem is this solution breaks any third party libraries that rely on these C APIs (which of course are very common). And there is no good workaround (see https://github.com/NREL/OpenStudio/issues/3942#issuecomment-611072774). So would it be possible for Ruby to stop exporting custom versions of basic C APIs? The code that does it is here: https://github.com/ruby/ruby/blob/master/win32/mkexports.rb#L41 Ruby of course could still use its custom versions of `fclose`, read, write etc within ruby.exe and x64-vcruntime140-ruby340.dll. But they should not be exported from x64-vcruntime140-ruby340.dll and thus be off limits to extensions. If a case comes up where an extension really needs access to rb_w32_fclose instead of fclose then an extension developer can use an `#ifdef _WIN32` to do so and work across platforms. That at least puts the developer in control versus now where I don't see any way I can wrap the bitmap library as a Ruby extension. From my experience the biggest problem is the replacing of fclose with rb_w32_fclose. This is the list of generated overrides: FD_CLR=rb_w32_fdclr FD_ISSET=rb_w32_fdisset Sleep=rb_w32_Sleep accept=rb_w32_accept access=rb_w32_uaccess bind=rb_w32_bind close=rb_w32_close connect=rb_w32_connect dup2=rb_w32_dup2 fclose=rb_w32_fclose fstat=rb_w32_fstati128 get_osfhandle=rb_w32_get_osfhandle getcwd=rb_w32_ugetcwd getenv=rb_w32_ugetenv gethostbyaddr=rb_w32_gethostbyaddr gethostbyname=rb_w32_gethostbyname gethostname=rb_w32_gethostname getpeername=rb_w32_getpeername getpid=rb_w32_getpid getppid=rb_w32_getppid getprotobyname=rb_w32_getprotobyname getprotobynumber=rb_w32_getprotobynumber getservbyname=rb_w32_getservbyname getservbyport=rb_w32_getservbyport getsockname=rb_w32_getsockname getsockopt=rb_w32_getsockopt inet_ntop=rb_w32_inet_ntop inet_pton=rb_w32_inet_pton ioctlsocket=rb_w32_ioctlsocket isatty=rb_w32_isatty listen=rb_w32_listen lseek=rb_w32_lseek lstat=rb_w32_ulstati128 mkdir=rb_w32_umkdir mmap=rb_w32_mmap mprotect=rb_w32_mprotect munmap=rb_w32_munmap pipe=rb_w32_pipe pread=rb_w32_pread pwrite=rb_w32_pwrite read=rb_w32_read recv=rb_w32_recv recvfrom=rb_w32_recvfrom rename=rb_w32_urename rmdir=rb_w32_urmdir select=rb_w32_select send=rb_w32_send sendto=rb_w32_sendto setsockopt=rb_w32_setsockopt shutdown=rb_w32_shutdown socket=rb_w32_socket stati128=rb_w32_ustati128 strcasecmp=msvcrt.stricmp strerror=rb_w32_strerror strncasecmp=msvcrt.strnicmp times=rb_w32_times unlink=rb_w32_uunlink utime=rb_w32_uutime utimensat=rb_w32_uutimensat utimes=rb_w32_uutimes write=rb_w32_write -- https://bugs.ruby-lang.org/

Issue #21498 has been updated by cfis (Charlie Savage). In my case I created a Ruby extension based on https://github.com/baderouaich/BitmapPlusPlus. Every time I tried to save a bitmap though it was corrupted. I submitted a bug to the project - https://github.com/baderouaich/BitmapPlusPlus/issues/6. But I later realized the real problem is Ruby overrides fclose: * bitmap library uses `std::ofstream` to write the file * ofstream eventually calls `fclose` But instead of calling `fclose` in UCRT, it calls Ruby's version of `fclose`. That happens because the Ruby dll exports `fclose` as a symbol and that *overrides* the one exported from the C runtime library (UCRT). I know this for a fact because I stepped through the code using a debugger. I can show you a screenshot of that if you would like to see it yourself. Ruby's version of fclose (rb32_fclose) is not compatible with the UCRT version, and thus the file ends up corrupted. ---------------------------------------- Bug #21498: Windows - Ruby Overrides C Library APIs thus breaking them https://bugs.ruby-lang.org/issues/21498#change-115104 * Author: cfis (Charlie Savage) * Status: Open * Assignee: windows * ruby -v: *all* on windows (testing with ruby 3.4.3 (2025-04-14 revision d0b7e5b6a0) +PRISM [x64-mswin64_140]) * Backport: 3.2: UNKNOWN, 3.3: UNKNOWN, 3.4: UNKNOWN ---------------------------------------- I am trying to wrap a simple C++ library, https://github.com/baderouaich/BitmapPlusPlus, as a Ruby extension. However when I use the extension to write a bitmap to disk the bitmap is corrupted. This is because the library uses std::ofstream which eventually uses the C API fclose to write the final bytes to the bitmap file and then closes it. The problem is that Ruby overrides `fclose` and replaces it with `rb_w32_fclose`. It then *exports* its custom version from from x64-vcruntime140-ruby340.dll. And the exported version is broken (at least from the standpoint of the C standard library). Note this has been a long standing issue. The first report I see is from 2013: https://bugs.ruby-lang.org/issues/8569 More recently in 2020 (which explains the issue very well): https://github.com/NREL/OpenStudio/issues/3942#issuecomment-610673401 I understand that Ruby is trying to provide a platform independent API. But the problem is this solution breaks any third party libraries that rely on these C APIs (which of course are very common). And there is no good workaround (see https://github.com/NREL/OpenStudio/issues/3942#issuecomment-611072774). So would it be possible for Ruby to stop exporting custom versions of basic C APIs? The code that does it is here: https://github.com/ruby/ruby/blob/master/win32/mkexports.rb#L41 Ruby of course could still use its custom versions of `fclose`, read, write etc within ruby.exe and x64-vcruntime140-ruby340.dll. But they should not be exported from x64-vcruntime140-ruby340.dll and thus be off limits to extensions. If a case comes up where an extension really needs access to rb_w32_fclose instead of fclose then an extension developer can use an `#ifdef _WIN32` to do so and work across platforms. That at least puts the developer in control versus now where I don't see any way I can wrap the bitmap library as a Ruby extension. From my experience the biggest problem is the replacing of fclose with rb_w32_fclose. This is the list of generated overrides: FD_CLR=rb_w32_fdclr FD_ISSET=rb_w32_fdisset Sleep=rb_w32_Sleep accept=rb_w32_accept access=rb_w32_uaccess bind=rb_w32_bind close=rb_w32_close connect=rb_w32_connect dup2=rb_w32_dup2 fclose=rb_w32_fclose fstat=rb_w32_fstati128 get_osfhandle=rb_w32_get_osfhandle getcwd=rb_w32_ugetcwd getenv=rb_w32_ugetenv gethostbyaddr=rb_w32_gethostbyaddr gethostbyname=rb_w32_gethostbyname gethostname=rb_w32_gethostname getpeername=rb_w32_getpeername getpid=rb_w32_getpid getppid=rb_w32_getppid getprotobyname=rb_w32_getprotobyname getprotobynumber=rb_w32_getprotobynumber getservbyname=rb_w32_getservbyname getservbyport=rb_w32_getservbyport getsockname=rb_w32_getsockname getsockopt=rb_w32_getsockopt inet_ntop=rb_w32_inet_ntop inet_pton=rb_w32_inet_pton ioctlsocket=rb_w32_ioctlsocket isatty=rb_w32_isatty listen=rb_w32_listen lseek=rb_w32_lseek lstat=rb_w32_ulstati128 mkdir=rb_w32_umkdir mmap=rb_w32_mmap mprotect=rb_w32_mprotect munmap=rb_w32_munmap pipe=rb_w32_pipe pread=rb_w32_pread pwrite=rb_w32_pwrite read=rb_w32_read recv=rb_w32_recv recvfrom=rb_w32_recvfrom rename=rb_w32_urename rmdir=rb_w32_urmdir select=rb_w32_select send=rb_w32_send sendto=rb_w32_sendto setsockopt=rb_w32_setsockopt shutdown=rb_w32_shutdown socket=rb_w32_socket stati128=rb_w32_ustati128 strcasecmp=msvcrt.stricmp strerror=rb_w32_strerror strncasecmp=msvcrt.strnicmp times=rb_w32_times unlink=rb_w32_uunlink utime=rb_w32_uutime utimensat=rb_w32_uutimensat utimes=rb_w32_uutimes write=rb_w32_write -- https://bugs.ruby-lang.org/

Issue #21498 has been updated by nobu (Nobuyoshi Nakada). cfis (Charlie Savage) wrote in #note-11:

Notice in the stack trace on the bottom of the page that the method ` bmp::Bitmap::save(const std::filesystem::path & filename)` calls fclose but ends up calling the Ruby version! Not the UCRT version.

`rb_w32_fclose` also calls `fclose` in the UCRT. ---------------------------------------- Bug #21498: Windows - Ruby Overrides C Library APIs thus breaking them https://bugs.ruby-lang.org/issues/21498#change-115152 * Author: cfis (Charlie Savage) * Status: Open * Assignee: windows * ruby -v: *all* on windows (testing with ruby 3.4.3 (2025-04-14 revision d0b7e5b6a0) +PRISM [x64-mswin64_140]) * Backport: 3.2: UNKNOWN, 3.3: UNKNOWN, 3.4: UNKNOWN ---------------------------------------- I am trying to wrap a simple C++ library, https://github.com/baderouaich/BitmapPlusPlus, as a Ruby extension. However when I use the extension to write a bitmap to disk the bitmap is corrupted. This is because the library uses std::ofstream which eventually uses the C API fclose to write the final bytes to the bitmap file and then closes it. The problem is that Ruby overrides `fclose` and replaces it with `rb_w32_fclose`. It then *exports* its custom version from from x64-vcruntime140-ruby340.dll. And the exported version is broken (at least from the standpoint of the C standard library). Note this has been a long standing issue. The first report I see is from 2013: https://bugs.ruby-lang.org/issues/8569 More recently in 2020 (which explains the issue very well): https://github.com/NREL/OpenStudio/issues/3942#issuecomment-610673401 I understand that Ruby is trying to provide a platform independent API. But the problem is this solution breaks any third party libraries that rely on these C APIs (which of course are very common). And there is no good workaround (see https://github.com/NREL/OpenStudio/issues/3942#issuecomment-611072774). So would it be possible for Ruby to stop exporting custom versions of basic C APIs? The code that does it is here: https://github.com/ruby/ruby/blob/master/win32/mkexports.rb#L41 Ruby of course could still use its custom versions of `fclose`, read, write etc within ruby.exe and x64-vcruntime140-ruby340.dll. But they should not be exported from x64-vcruntime140-ruby340.dll and thus be off limits to extensions. If a case comes up where an extension really needs access to rb_w32_fclose instead of fclose then an extension developer can use an `#ifdef _WIN32` to do so and work across platforms. That at least puts the developer in control versus now where I don't see any way I can wrap the bitmap library as a Ruby extension. From my experience the biggest problem is the replacing of fclose with rb_w32_fclose. This is the list of generated overrides: FD_CLR=rb_w32_fdclr FD_ISSET=rb_w32_fdisset Sleep=rb_w32_Sleep accept=rb_w32_accept access=rb_w32_uaccess bind=rb_w32_bind close=rb_w32_close connect=rb_w32_connect dup2=rb_w32_dup2 fclose=rb_w32_fclose fstat=rb_w32_fstati128 get_osfhandle=rb_w32_get_osfhandle getcwd=rb_w32_ugetcwd getenv=rb_w32_ugetenv gethostbyaddr=rb_w32_gethostbyaddr gethostbyname=rb_w32_gethostbyname gethostname=rb_w32_gethostname getpeername=rb_w32_getpeername getpid=rb_w32_getpid getppid=rb_w32_getppid getprotobyname=rb_w32_getprotobyname getprotobynumber=rb_w32_getprotobynumber getservbyname=rb_w32_getservbyname getservbyport=rb_w32_getservbyport getsockname=rb_w32_getsockname getsockopt=rb_w32_getsockopt inet_ntop=rb_w32_inet_ntop inet_pton=rb_w32_inet_pton ioctlsocket=rb_w32_ioctlsocket isatty=rb_w32_isatty listen=rb_w32_listen lseek=rb_w32_lseek lstat=rb_w32_ulstati128 mkdir=rb_w32_umkdir mmap=rb_w32_mmap mprotect=rb_w32_mprotect munmap=rb_w32_munmap pipe=rb_w32_pipe pread=rb_w32_pread pwrite=rb_w32_pwrite read=rb_w32_read recv=rb_w32_recv recvfrom=rb_w32_recvfrom rename=rb_w32_urename rmdir=rb_w32_urmdir select=rb_w32_select send=rb_w32_send sendto=rb_w32_sendto setsockopt=rb_w32_setsockopt shutdown=rb_w32_shutdown socket=rb_w32_socket stati128=rb_w32_ustati128 strcasecmp=msvcrt.stricmp strerror=rb_w32_strerror strncasecmp=msvcrt.strnicmp times=rb_w32_times unlink=rb_w32_uunlink utime=rb_w32_uutime utimensat=rb_w32_uutimensat utimes=rb_w32_uutimes write=rb_w32_write ---Files-------------------------------- Screenshot 2025-11-06 233918.png (223 KB) -- https://bugs.ruby-lang.org/

Issue #21498 has been updated by mame (Yusuke Endoh). cfis (Charlie Savage) wrote in #note-13:

if the call goes through `rb_w32_fclose` the saved file is corrupted/invalid.

@nobu is curious *why* it's getting corrupted, given that `rb_w32_fclose` just delegates to the original `fclose` when `!is_socket(sock)` is true. In that case, it shouldn't be corrupted. ---------------------------------------- Bug #21498: Windows - Ruby Overrides C Library APIs thus breaking them https://bugs.ruby-lang.org/issues/21498#change-115184 * Author: cfis (Charlie Savage) * Status: Open * Assignee: windows * ruby -v: *all* on windows (testing with ruby 3.4.3 (2025-04-14 revision d0b7e5b6a0) +PRISM [x64-mswin64_140]) * Backport: 3.2: UNKNOWN, 3.3: UNKNOWN, 3.4: UNKNOWN ---------------------------------------- I am trying to wrap a simple C++ library, https://github.com/baderouaich/BitmapPlusPlus, as a Ruby extension. However when I use the extension to write a bitmap to disk the bitmap is corrupted. This is because the library uses std::ofstream which eventually uses the C API fclose to write the final bytes to the bitmap file and then closes it. The problem is that Ruby overrides `fclose` and replaces it with `rb_w32_fclose`. It then *exports* its custom version from from x64-vcruntime140-ruby340.dll. And the exported version is broken (at least from the standpoint of the C standard library). Note this has been a long standing issue. The first report I see is from 2013: https://bugs.ruby-lang.org/issues/8569 More recently in 2020 (which explains the issue very well): https://github.com/NREL/OpenStudio/issues/3942#issuecomment-610673401 I understand that Ruby is trying to provide a platform independent API. But the problem is this solution breaks any third party libraries that rely on these C APIs (which of course are very common). And there is no good workaround (see https://github.com/NREL/OpenStudio/issues/3942#issuecomment-611072774). So would it be possible for Ruby to stop exporting custom versions of basic C APIs? The code that does it is here: https://github.com/ruby/ruby/blob/master/win32/mkexports.rb#L41 Ruby of course could still use its custom versions of `fclose`, read, write etc within ruby.exe and x64-vcruntime140-ruby340.dll. But they should not be exported from x64-vcruntime140-ruby340.dll and thus be off limits to extensions. If a case comes up where an extension really needs access to rb_w32_fclose instead of fclose then an extension developer can use an `#ifdef _WIN32` to do so and work across platforms. That at least puts the developer in control versus now where I don't see any way I can wrap the bitmap library as a Ruby extension. From my experience the biggest problem is the replacing of fclose with rb_w32_fclose. This is the list of generated overrides: FD_CLR=rb_w32_fdclr FD_ISSET=rb_w32_fdisset Sleep=rb_w32_Sleep accept=rb_w32_accept access=rb_w32_uaccess bind=rb_w32_bind close=rb_w32_close connect=rb_w32_connect dup2=rb_w32_dup2 fclose=rb_w32_fclose fstat=rb_w32_fstati128 get_osfhandle=rb_w32_get_osfhandle getcwd=rb_w32_ugetcwd getenv=rb_w32_ugetenv gethostbyaddr=rb_w32_gethostbyaddr gethostbyname=rb_w32_gethostbyname gethostname=rb_w32_gethostname getpeername=rb_w32_getpeername getpid=rb_w32_getpid getppid=rb_w32_getppid getprotobyname=rb_w32_getprotobyname getprotobynumber=rb_w32_getprotobynumber getservbyname=rb_w32_getservbyname getservbyport=rb_w32_getservbyport getsockname=rb_w32_getsockname getsockopt=rb_w32_getsockopt inet_ntop=rb_w32_inet_ntop inet_pton=rb_w32_inet_pton ioctlsocket=rb_w32_ioctlsocket isatty=rb_w32_isatty listen=rb_w32_listen lseek=rb_w32_lseek lstat=rb_w32_ulstati128 mkdir=rb_w32_umkdir mmap=rb_w32_mmap mprotect=rb_w32_mprotect munmap=rb_w32_munmap pipe=rb_w32_pipe pread=rb_w32_pread pwrite=rb_w32_pwrite read=rb_w32_read recv=rb_w32_recv recvfrom=rb_w32_recvfrom rename=rb_w32_urename rmdir=rb_w32_urmdir select=rb_w32_select send=rb_w32_send sendto=rb_w32_sendto setsockopt=rb_w32_setsockopt shutdown=rb_w32_shutdown socket=rb_w32_socket stati128=rb_w32_ustati128 strcasecmp=msvcrt.stricmp strerror=rb_w32_strerror strncasecmp=msvcrt.strnicmp times=rb_w32_times unlink=rb_w32_uunlink utime=rb_w32_uutime utimensat=rb_w32_uutimensat utimes=rb_w32_uutimes write=rb_w32_write ---Files-------------------------------- Screenshot 2025-11-06 233918.png (223 KB) -- https://bugs.ruby-lang.org/

Issue #21498 has been updated by cfis (Charlie Savage). File primitives.bmp added File primitives_bad.bmp added I have attached the "good" file and the "bad" file. The good file is generated by rewriting the extension to not use fclose, and thus not have the call go through `rb_w32_fclose`. The bad file is generated by using flcose which does end up calling `rb_w32_fclose `. ---------------------------------------- Bug #21498: Windows - Ruby Overrides C Library APIs thus breaking them https://bugs.ruby-lang.org/issues/21498#change-115206 * Author: cfis (Charlie Savage) * Status: Open * Assignee: windows * ruby -v: *all* on windows (testing with ruby 3.4.3 (2025-04-14 revision d0b7e5b6a0) +PRISM [x64-mswin64_140]) * Backport: 3.2: UNKNOWN, 3.3: UNKNOWN, 3.4: UNKNOWN ---------------------------------------- I am trying to wrap a simple C++ library, https://github.com/baderouaich/BitmapPlusPlus, as a Ruby extension. However when I use the extension to write a bitmap to disk the bitmap is corrupted. This is because the library uses std::ofstream which eventually uses the C API fclose to write the final bytes to the bitmap file and then closes it. The problem is that Ruby overrides `fclose` and replaces it with `rb_w32_fclose`. It then *exports* its custom version from from x64-vcruntime140-ruby340.dll. And the exported version is broken (at least from the standpoint of the C standard library). Note this has been a long standing issue. The first report I see is from 2013: https://bugs.ruby-lang.org/issues/8569 More recently in 2020 (which explains the issue very well): https://github.com/NREL/OpenStudio/issues/3942#issuecomment-610673401 I understand that Ruby is trying to provide a platform independent API. But the problem is this solution breaks any third party libraries that rely on these C APIs (which of course are very common). And there is no good workaround (see https://github.com/NREL/OpenStudio/issues/3942#issuecomment-611072774). So would it be possible for Ruby to stop exporting custom versions of basic C APIs? The code that does it is here: https://github.com/ruby/ruby/blob/master/win32/mkexports.rb#L41 Ruby of course could still use its custom versions of `fclose`, read, write etc within ruby.exe and x64-vcruntime140-ruby340.dll. But they should not be exported from x64-vcruntime140-ruby340.dll and thus be off limits to extensions. If a case comes up where an extension really needs access to rb_w32_fclose instead of fclose then an extension developer can use an `#ifdef _WIN32` to do so and work across platforms. That at least puts the developer in control versus now where I don't see any way I can wrap the bitmap library as a Ruby extension. From my experience the biggest problem is the replacing of fclose with rb_w32_fclose. This is the list of generated overrides: FD_CLR=rb_w32_fdclr FD_ISSET=rb_w32_fdisset Sleep=rb_w32_Sleep accept=rb_w32_accept access=rb_w32_uaccess bind=rb_w32_bind close=rb_w32_close connect=rb_w32_connect dup2=rb_w32_dup2 fclose=rb_w32_fclose fstat=rb_w32_fstati128 get_osfhandle=rb_w32_get_osfhandle getcwd=rb_w32_ugetcwd getenv=rb_w32_ugetenv gethostbyaddr=rb_w32_gethostbyaddr gethostbyname=rb_w32_gethostbyname gethostname=rb_w32_gethostname getpeername=rb_w32_getpeername getpid=rb_w32_getpid getppid=rb_w32_getppid getprotobyname=rb_w32_getprotobyname getprotobynumber=rb_w32_getprotobynumber getservbyname=rb_w32_getservbyname getservbyport=rb_w32_getservbyport getsockname=rb_w32_getsockname getsockopt=rb_w32_getsockopt inet_ntop=rb_w32_inet_ntop inet_pton=rb_w32_inet_pton ioctlsocket=rb_w32_ioctlsocket isatty=rb_w32_isatty listen=rb_w32_listen lseek=rb_w32_lseek lstat=rb_w32_ulstati128 mkdir=rb_w32_umkdir mmap=rb_w32_mmap mprotect=rb_w32_mprotect munmap=rb_w32_munmap pipe=rb_w32_pipe pread=rb_w32_pread pwrite=rb_w32_pwrite read=rb_w32_read recv=rb_w32_recv recvfrom=rb_w32_recvfrom rename=rb_w32_urename rmdir=rb_w32_urmdir select=rb_w32_select send=rb_w32_send sendto=rb_w32_sendto setsockopt=rb_w32_setsockopt shutdown=rb_w32_shutdown socket=rb_w32_socket stati128=rb_w32_ustati128 strcasecmp=msvcrt.stricmp strerror=rb_w32_strerror strncasecmp=msvcrt.strnicmp times=rb_w32_times unlink=rb_w32_uunlink utime=rb_w32_uutime utimensat=rb_w32_uutimensat utimes=rb_w32_uutimes write=rb_w32_write ---Files-------------------------------- Screenshot 2025-11-06 233918.png (223 KB) primitives.bmp (360 KB) primitives_bad.bmp (360 KB) -- https://bugs.ruby-lang.org/

Issue #21498 has been updated by mame (Yusuke Endoh). cfis (Charlie Savage) wrote in #note-15:

When I step through with a debugger, this line forces an early return from the function. So fclose is never called.

https://github.com/ruby/ruby/blob/master/win32/win32.c#L6627

``` cpp if (fflush(fp)) return -1; ```

Thank you for debugging. The early return (not calling `fclose` when `fflush` fails) seems problematic in terms of resource leaks. However, I don't think the leak itself is related to your issue. The real question is why `fflush` is failing. @nobu said in the dev meeting that he's not opposed to removing the override, but he doesn't want to do it just because 'the problem went away.' He wants to understand the mechanism why the override doesn't work well. cfis (Charlie Savage) wrote in #note-16:

The good one is 372,736 bytes and the bad one is 368,640 bytes. Also interesting is your browser can probably render both.

Looking closely, the bad one is missing data of the last line (i.e., the topmost line of the bmp). It does indeed seem to have failed flushing. By the way, calling `fclose` while buffered data remains is somewhat risky. `fclose` flushes the buffer before closing, but it doesn't specify what happens to the data if the flush fails. It's likely discarded. ---------------------------------------- Bug #21498: Windows - Ruby Overrides C Library APIs thus breaking them https://bugs.ruby-lang.org/issues/21498#change-115210 * Author: cfis (Charlie Savage) * Status: Open * Assignee: windows * ruby -v: *all* on windows (testing with ruby 3.4.3 (2025-04-14 revision d0b7e5b6a0) +PRISM [x64-mswin64_140]) * Backport: 3.2: UNKNOWN, 3.3: UNKNOWN, 3.4: UNKNOWN ---------------------------------------- I am trying to wrap a simple C++ library, https://github.com/baderouaich/BitmapPlusPlus, as a Ruby extension. However when I use the extension to write a bitmap to disk the bitmap is corrupted. This is because the library uses std::ofstream which eventually uses the C API fclose to write the final bytes to the bitmap file and then closes it. The problem is that Ruby overrides `fclose` and replaces it with `rb_w32_fclose`. It then *exports* its custom version from from x64-vcruntime140-ruby340.dll. And the exported version is broken (at least from the standpoint of the C standard library). Note this has been a long standing issue. The first report I see is from 2013: https://bugs.ruby-lang.org/issues/8569 More recently in 2020 (which explains the issue very well): https://github.com/NREL/OpenStudio/issues/3942#issuecomment-610673401 I understand that Ruby is trying to provide a platform independent API. But the problem is this solution breaks any third party libraries that rely on these C APIs (which of course are very common). And there is no good workaround (see https://github.com/NREL/OpenStudio/issues/3942#issuecomment-611072774). So would it be possible for Ruby to stop exporting custom versions of basic C APIs? The code that does it is here: https://github.com/ruby/ruby/blob/master/win32/mkexports.rb#L41 Ruby of course could still use its custom versions of `fclose`, read, write etc within ruby.exe and x64-vcruntime140-ruby340.dll. But they should not be exported from x64-vcruntime140-ruby340.dll and thus be off limits to extensions. If a case comes up where an extension really needs access to rb_w32_fclose instead of fclose then an extension developer can use an `#ifdef _WIN32` to do so and work across platforms. That at least puts the developer in control versus now where I don't see any way I can wrap the bitmap library as a Ruby extension. From my experience the biggest problem is the replacing of fclose with rb_w32_fclose. This is the list of generated overrides: FD_CLR=rb_w32_fdclr FD_ISSET=rb_w32_fdisset Sleep=rb_w32_Sleep accept=rb_w32_accept access=rb_w32_uaccess bind=rb_w32_bind close=rb_w32_close connect=rb_w32_connect dup2=rb_w32_dup2 fclose=rb_w32_fclose fstat=rb_w32_fstati128 get_osfhandle=rb_w32_get_osfhandle getcwd=rb_w32_ugetcwd getenv=rb_w32_ugetenv gethostbyaddr=rb_w32_gethostbyaddr gethostbyname=rb_w32_gethostbyname gethostname=rb_w32_gethostname getpeername=rb_w32_getpeername getpid=rb_w32_getpid getppid=rb_w32_getppid getprotobyname=rb_w32_getprotobyname getprotobynumber=rb_w32_getprotobynumber getservbyname=rb_w32_getservbyname getservbyport=rb_w32_getservbyport getsockname=rb_w32_getsockname getsockopt=rb_w32_getsockopt inet_ntop=rb_w32_inet_ntop inet_pton=rb_w32_inet_pton ioctlsocket=rb_w32_ioctlsocket isatty=rb_w32_isatty listen=rb_w32_listen lseek=rb_w32_lseek lstat=rb_w32_ulstati128 mkdir=rb_w32_umkdir mmap=rb_w32_mmap mprotect=rb_w32_mprotect munmap=rb_w32_munmap pipe=rb_w32_pipe pread=rb_w32_pread pwrite=rb_w32_pwrite read=rb_w32_read recv=rb_w32_recv recvfrom=rb_w32_recvfrom rename=rb_w32_urename rmdir=rb_w32_urmdir select=rb_w32_select send=rb_w32_send sendto=rb_w32_sendto setsockopt=rb_w32_setsockopt shutdown=rb_w32_shutdown socket=rb_w32_socket stati128=rb_w32_ustati128 strcasecmp=msvcrt.stricmp strerror=rb_w32_strerror strncasecmp=msvcrt.strnicmp times=rb_w32_times unlink=rb_w32_uunlink utime=rb_w32_uutime utimensat=rb_w32_uutimensat utimes=rb_w32_uutimes write=rb_w32_write ---Files-------------------------------- Screenshot 2025-11-06 233918.png (223 KB) primitives.bmp (360 KB) primitives_bad.bmp (360 KB) -- https://bugs.ruby-lang.org/

Issue #21498 has been updated by YO4 (Yoshinao Muramatsu). If `rb_win32_fclose` calls ucrt's `fflush`/`fclose`, it's questionable why it doesn't work. Is it truly calling the same ucrts? The screenshot suggests the C++ code is built in Debug build. From the C++ code, the the ucrt debug DLL might be linked. On the other hand, if ruby isn't specially built, ruby's DLL will likely call the release version of the ucrt. Could you verify which DLL's `fopen`/`_wfopen` (probably called from `std::ofstream`'s constructor) and `fflush`/`fclose` (called from `rb_win32_fclose`) are being used? Of course, if ucrt debug and release builds are compatible, this wouldn't matter for this issue. I apologize, but I have no knowledge regarding their compatibility. ---------------------------------------- Bug #21498: Windows - Ruby Overrides C Library APIs thus breaking them https://bugs.ruby-lang.org/issues/21498#change-115213 * Author: cfis (Charlie Savage) * Status: Open * Assignee: windows * ruby -v: *all* on windows (testing with ruby 3.4.3 (2025-04-14 revision d0b7e5b6a0) +PRISM [x64-mswin64_140]) * Backport: 3.2: UNKNOWN, 3.3: UNKNOWN, 3.4: UNKNOWN ---------------------------------------- I am trying to wrap a simple C++ library, https://github.com/baderouaich/BitmapPlusPlus, as a Ruby extension. However when I use the extension to write a bitmap to disk the bitmap is corrupted. This is because the library uses std::ofstream which eventually uses the C API fclose to write the final bytes to the bitmap file and then closes it. The problem is that Ruby overrides `fclose` and replaces it with `rb_w32_fclose`. It then *exports* its custom version from from x64-vcruntime140-ruby340.dll. And the exported version is broken (at least from the standpoint of the C standard library). Note this has been a long standing issue. The first report I see is from 2013: https://bugs.ruby-lang.org/issues/8569 More recently in 2020 (which explains the issue very well): https://github.com/NREL/OpenStudio/issues/3942#issuecomment-610673401 I understand that Ruby is trying to provide a platform independent API. But the problem is this solution breaks any third party libraries that rely on these C APIs (which of course are very common). And there is no good workaround (see https://github.com/NREL/OpenStudio/issues/3942#issuecomment-611072774). So would it be possible for Ruby to stop exporting custom versions of basic C APIs? The code that does it is here: https://github.com/ruby/ruby/blob/master/win32/mkexports.rb#L41 Ruby of course could still use its custom versions of `fclose`, read, write etc within ruby.exe and x64-vcruntime140-ruby340.dll. But they should not be exported from x64-vcruntime140-ruby340.dll and thus be off limits to extensions. If a case comes up where an extension really needs access to rb_w32_fclose instead of fclose then an extension developer can use an `#ifdef _WIN32` to do so and work across platforms. That at least puts the developer in control versus now where I don't see any way I can wrap the bitmap library as a Ruby extension. From my experience the biggest problem is the replacing of fclose with rb_w32_fclose. This is the list of generated overrides: FD_CLR=rb_w32_fdclr FD_ISSET=rb_w32_fdisset Sleep=rb_w32_Sleep accept=rb_w32_accept access=rb_w32_uaccess bind=rb_w32_bind close=rb_w32_close connect=rb_w32_connect dup2=rb_w32_dup2 fclose=rb_w32_fclose fstat=rb_w32_fstati128 get_osfhandle=rb_w32_get_osfhandle getcwd=rb_w32_ugetcwd getenv=rb_w32_ugetenv gethostbyaddr=rb_w32_gethostbyaddr gethostbyname=rb_w32_gethostbyname gethostname=rb_w32_gethostname getpeername=rb_w32_getpeername getpid=rb_w32_getpid getppid=rb_w32_getppid getprotobyname=rb_w32_getprotobyname getprotobynumber=rb_w32_getprotobynumber getservbyname=rb_w32_getservbyname getservbyport=rb_w32_getservbyport getsockname=rb_w32_getsockname getsockopt=rb_w32_getsockopt inet_ntop=rb_w32_inet_ntop inet_pton=rb_w32_inet_pton ioctlsocket=rb_w32_ioctlsocket isatty=rb_w32_isatty listen=rb_w32_listen lseek=rb_w32_lseek lstat=rb_w32_ulstati128 mkdir=rb_w32_umkdir mmap=rb_w32_mmap mprotect=rb_w32_mprotect munmap=rb_w32_munmap pipe=rb_w32_pipe pread=rb_w32_pread pwrite=rb_w32_pwrite read=rb_w32_read recv=rb_w32_recv recvfrom=rb_w32_recvfrom rename=rb_w32_urename rmdir=rb_w32_urmdir select=rb_w32_select send=rb_w32_send sendto=rb_w32_sendto setsockopt=rb_w32_setsockopt shutdown=rb_w32_shutdown socket=rb_w32_socket stati128=rb_w32_ustati128 strcasecmp=msvcrt.stricmp strerror=rb_w32_strerror strncasecmp=msvcrt.strnicmp times=rb_w32_times unlink=rb_w32_uunlink utime=rb_w32_uutime utimensat=rb_w32_uutimensat utimes=rb_w32_uutimes write=rb_w32_write ---Files-------------------------------- Screenshot 2025-11-06 233918.png (223 KB) primitives.bmp (360 KB) primitives_bad.bmp (360 KB) -- https://bugs.ruby-lang.org/

Issue #21498 has been updated by nobu (Nobuyoshi Nakada). cfis (Charlie Savage) wrote in #note-21:

Summary - don't pass memory/crt objects across boundaries. By overriding `fclose` and other UCRT functions Ruby causes this to happen.

Yes for the former half. It is the reason we embed the CRT DLL version in the ruby DLL. And "the CRT DLL version" includes release/debug, and the command line option is stored in `RbConfig::CFLAGS`. I think we should check the CRT DLL version at loading time too. ---------------------------------------- Bug #21498: Windows - Ruby Overrides C Library APIs thus breaking them https://bugs.ruby-lang.org/issues/21498#change-115221 * Author: cfis (Charlie Savage) * Status: Open * Assignee: windows * ruby -v: *all* on windows (testing with ruby 3.4.3 (2025-04-14 revision d0b7e5b6a0) +PRISM [x64-mswin64_140]) * Backport: 3.2: UNKNOWN, 3.3: UNKNOWN, 3.4: UNKNOWN ---------------------------------------- I am trying to wrap a simple C++ library, https://github.com/baderouaich/BitmapPlusPlus, as a Ruby extension. However when I use the extension to write a bitmap to disk the bitmap is corrupted. This is because the library uses std::ofstream which eventually uses the C API fclose to write the final bytes to the bitmap file and then closes it. The problem is that Ruby overrides `fclose` and replaces it with `rb_w32_fclose`. It then *exports* its custom version from from x64-vcruntime140-ruby340.dll. And the exported version is broken (at least from the standpoint of the C standard library). Note this has been a long standing issue. The first report I see is from 2013: https://bugs.ruby-lang.org/issues/8569 More recently in 2020 (which explains the issue very well): https://github.com/NREL/OpenStudio/issues/3942#issuecomment-610673401 I understand that Ruby is trying to provide a platform independent API. But the problem is this solution breaks any third party libraries that rely on these C APIs (which of course are very common). And there is no good workaround (see https://github.com/NREL/OpenStudio/issues/3942#issuecomment-611072774). So would it be possible for Ruby to stop exporting custom versions of basic C APIs? The code that does it is here: https://github.com/ruby/ruby/blob/master/win32/mkexports.rb#L41 Ruby of course could still use its custom versions of `fclose`, read, write etc within ruby.exe and x64-vcruntime140-ruby340.dll. But they should not be exported from x64-vcruntime140-ruby340.dll and thus be off limits to extensions. If a case comes up where an extension really needs access to rb_w32_fclose instead of fclose then an extension developer can use an `#ifdef _WIN32` to do so and work across platforms. That at least puts the developer in control versus now where I don't see any way I can wrap the bitmap library as a Ruby extension. From my experience the biggest problem is the replacing of fclose with rb_w32_fclose. This is the list of generated overrides: FD_CLR=rb_w32_fdclr FD_ISSET=rb_w32_fdisset Sleep=rb_w32_Sleep accept=rb_w32_accept access=rb_w32_uaccess bind=rb_w32_bind close=rb_w32_close connect=rb_w32_connect dup2=rb_w32_dup2 fclose=rb_w32_fclose fstat=rb_w32_fstati128 get_osfhandle=rb_w32_get_osfhandle getcwd=rb_w32_ugetcwd getenv=rb_w32_ugetenv gethostbyaddr=rb_w32_gethostbyaddr gethostbyname=rb_w32_gethostbyname gethostname=rb_w32_gethostname getpeername=rb_w32_getpeername getpid=rb_w32_getpid getppid=rb_w32_getppid getprotobyname=rb_w32_getprotobyname getprotobynumber=rb_w32_getprotobynumber getservbyname=rb_w32_getservbyname getservbyport=rb_w32_getservbyport getsockname=rb_w32_getsockname getsockopt=rb_w32_getsockopt inet_ntop=rb_w32_inet_ntop inet_pton=rb_w32_inet_pton ioctlsocket=rb_w32_ioctlsocket isatty=rb_w32_isatty listen=rb_w32_listen lseek=rb_w32_lseek lstat=rb_w32_ulstati128 mkdir=rb_w32_umkdir mmap=rb_w32_mmap mprotect=rb_w32_mprotect munmap=rb_w32_munmap pipe=rb_w32_pipe pread=rb_w32_pread pwrite=rb_w32_pwrite read=rb_w32_read recv=rb_w32_recv recvfrom=rb_w32_recvfrom rename=rb_w32_urename rmdir=rb_w32_urmdir select=rb_w32_select send=rb_w32_send sendto=rb_w32_sendto setsockopt=rb_w32_setsockopt shutdown=rb_w32_shutdown socket=rb_w32_socket stati128=rb_w32_ustati128 strcasecmp=msvcrt.stricmp strerror=rb_w32_strerror strncasecmp=msvcrt.strnicmp times=rb_w32_times unlink=rb_w32_uunlink utime=rb_w32_uutime utimensat=rb_w32_uutimensat utimes=rb_w32_uutimes write=rb_w32_write ---Files-------------------------------- Screenshot 2025-11-06 233918.png (223 KB) primitives.bmp (360 KB) primitives_bad.bmp (360 KB) -- https://bugs.ruby-lang.org/

Issue #21498 has been updated by mame (Yusuke Endoh). I commend YO4-san for their insight into the debug/release build discrepancy. I'm now convinced the root cause is loading multiple runtimes. Therefore, I agree with @nobu's fix: we should prohibit loading multiple runtime versions, not just stop the override. Mixing build types (e.g., debug extensions on a release core) can cause subtle failures in C APIs like `FILE *rb_io_stdio_file(rb_io_t *fptr);`. This risk alone justifies the prohibition. Furthermore, Microsoft's documentation implies using a single CRT version is safer:

It's also possible to avoid some of these issues if all of the images in your process use the same dynamically loaded version of the CRT.

This doesn't read like a recommendation to load multiple versions. While not mixing builds may seem inconvenient, this issue itself demonstrates the trouble it causes. For your case, building the Ruby core in debug mode is a necessary overhead, not an "unnecessary overhead" you said, I think. ---------------------------------------- Bug #21498: Windows - Ruby Overrides C Library APIs thus breaking them https://bugs.ruby-lang.org/issues/21498#change-115223 * Author: cfis (Charlie Savage) * Status: Open * Assignee: windows * ruby -v: *all* on windows (testing with ruby 3.4.3 (2025-04-14 revision d0b7e5b6a0) +PRISM [x64-mswin64_140]) * Backport: 3.2: UNKNOWN, 3.3: UNKNOWN, 3.4: UNKNOWN ---------------------------------------- I am trying to wrap a simple C++ library, https://github.com/baderouaich/BitmapPlusPlus, as a Ruby extension. However when I use the extension to write a bitmap to disk the bitmap is corrupted. This is because the library uses std::ofstream which eventually uses the C API fclose to write the final bytes to the bitmap file and then closes it. The problem is that Ruby overrides `fclose` and replaces it with `rb_w32_fclose`. It then *exports* its custom version from from x64-vcruntime140-ruby340.dll. And the exported version is broken (at least from the standpoint of the C standard library). Note this has been a long standing issue. The first report I see is from 2013: https://bugs.ruby-lang.org/issues/8569 More recently in 2020 (which explains the issue very well): https://github.com/NREL/OpenStudio/issues/3942#issuecomment-610673401 I understand that Ruby is trying to provide a platform independent API. But the problem is this solution breaks any third party libraries that rely on these C APIs (which of course are very common). And there is no good workaround (see https://github.com/NREL/OpenStudio/issues/3942#issuecomment-611072774). So would it be possible for Ruby to stop exporting custom versions of basic C APIs? The code that does it is here: https://github.com/ruby/ruby/blob/master/win32/mkexports.rb#L41 Ruby of course could still use its custom versions of `fclose`, read, write etc within ruby.exe and x64-vcruntime140-ruby340.dll. But they should not be exported from x64-vcruntime140-ruby340.dll and thus be off limits to extensions. If a case comes up where an extension really needs access to rb_w32_fclose instead of fclose then an extension developer can use an `#ifdef _WIN32` to do so and work across platforms. That at least puts the developer in control versus now where I don't see any way I can wrap the bitmap library as a Ruby extension. From my experience the biggest problem is the replacing of fclose with rb_w32_fclose. This is the list of generated overrides: FD_CLR=rb_w32_fdclr FD_ISSET=rb_w32_fdisset Sleep=rb_w32_Sleep accept=rb_w32_accept access=rb_w32_uaccess bind=rb_w32_bind close=rb_w32_close connect=rb_w32_connect dup2=rb_w32_dup2 fclose=rb_w32_fclose fstat=rb_w32_fstati128 get_osfhandle=rb_w32_get_osfhandle getcwd=rb_w32_ugetcwd getenv=rb_w32_ugetenv gethostbyaddr=rb_w32_gethostbyaddr gethostbyname=rb_w32_gethostbyname gethostname=rb_w32_gethostname getpeername=rb_w32_getpeername getpid=rb_w32_getpid getppid=rb_w32_getppid getprotobyname=rb_w32_getprotobyname getprotobynumber=rb_w32_getprotobynumber getservbyname=rb_w32_getservbyname getservbyport=rb_w32_getservbyport getsockname=rb_w32_getsockname getsockopt=rb_w32_getsockopt inet_ntop=rb_w32_inet_ntop inet_pton=rb_w32_inet_pton ioctlsocket=rb_w32_ioctlsocket isatty=rb_w32_isatty listen=rb_w32_listen lseek=rb_w32_lseek lstat=rb_w32_ulstati128 mkdir=rb_w32_umkdir mmap=rb_w32_mmap mprotect=rb_w32_mprotect munmap=rb_w32_munmap pipe=rb_w32_pipe pread=rb_w32_pread pwrite=rb_w32_pwrite read=rb_w32_read recv=rb_w32_recv recvfrom=rb_w32_recvfrom rename=rb_w32_urename rmdir=rb_w32_urmdir select=rb_w32_select send=rb_w32_send sendto=rb_w32_sendto setsockopt=rb_w32_setsockopt shutdown=rb_w32_shutdown socket=rb_w32_socket stati128=rb_w32_ustati128 strcasecmp=msvcrt.stricmp strerror=rb_w32_strerror strncasecmp=msvcrt.strnicmp times=rb_w32_times unlink=rb_w32_uunlink utime=rb_w32_uutime utimensat=rb_w32_uutimensat utimes=rb_w32_uutimes write=rb_w32_write ---Files-------------------------------- Screenshot 2025-11-06 233918.png (223 KB) primitives.bmp (360 KB) primitives_bad.bmp (360 KB) -- https://bugs.ruby-lang.org/

Issue #21498 has been updated by mame (Yusuke Endoh). I feel your communication on this ticket suffers from what is known as the [XY problem](https://en.wikipedia.org/wiki/XY_problem). You are focusing on your proposed solution ("remove the override"), without first detailing the original problem and providing steps to reproduce it. This is evident even from the ticket's title. Unfortunately, this style of discussion is well-known to be unproductive. To be clear, I (and probably @nobu) am not necessarily opposed to removing the override. In fact, I was initially so sympathetic to the idea that I was even lightly trying to persuade Nobu to move in that direction. However, now that it's clear the override wasn't the direct cause of the problem you are facing, I honestly find it difficult to take that argument seriously. I certainly no longer feel motivated to persuade Nobu about this. ---------------------------------------- Bug #21498: Windows - Ruby Overrides C Library APIs thus breaking them https://bugs.ruby-lang.org/issues/21498#change-115228 * Author: cfis (Charlie Savage) * Status: Open * Assignee: windows * ruby -v: *all* on windows (testing with ruby 3.4.3 (2025-04-14 revision d0b7e5b6a0) +PRISM [x64-mswin64_140]) * Backport: 3.2: UNKNOWN, 3.3: UNKNOWN, 3.4: UNKNOWN ---------------------------------------- I am trying to wrap a simple C++ library, https://github.com/baderouaich/BitmapPlusPlus, as a Ruby extension. However when I use the extension to write a bitmap to disk the bitmap is corrupted. This is because the library uses std::ofstream which eventually uses the C API fclose to write the final bytes to the bitmap file and then closes it. The problem is that Ruby overrides `fclose` and replaces it with `rb_w32_fclose`. It then *exports* its custom version from from x64-vcruntime140-ruby340.dll. And the exported version is broken (at least from the standpoint of the C standard library). Note this has been a long standing issue. The first report I see is from 2013: https://bugs.ruby-lang.org/issues/8569 More recently in 2020 (which explains the issue very well): https://github.com/NREL/OpenStudio/issues/3942#issuecomment-610673401 I understand that Ruby is trying to provide a platform independent API. But the problem is this solution breaks any third party libraries that rely on these C APIs (which of course are very common). And there is no good workaround (see https://github.com/NREL/OpenStudio/issues/3942#issuecomment-611072774). So would it be possible for Ruby to stop exporting custom versions of basic C APIs? The code that does it is here: https://github.com/ruby/ruby/blob/master/win32/mkexports.rb#L41 Ruby of course could still use its custom versions of `fclose`, read, write etc within ruby.exe and x64-vcruntime140-ruby340.dll. But they should not be exported from x64-vcruntime140-ruby340.dll and thus be off limits to extensions. If a case comes up where an extension really needs access to rb_w32_fclose instead of fclose then an extension developer can use an `#ifdef _WIN32` to do so and work across platforms. That at least puts the developer in control versus now where I don't see any way I can wrap the bitmap library as a Ruby extension. From my experience the biggest problem is the replacing of fclose with rb_w32_fclose. This is the list of generated overrides: FD_CLR=rb_w32_fdclr FD_ISSET=rb_w32_fdisset Sleep=rb_w32_Sleep accept=rb_w32_accept access=rb_w32_uaccess bind=rb_w32_bind close=rb_w32_close connect=rb_w32_connect dup2=rb_w32_dup2 fclose=rb_w32_fclose fstat=rb_w32_fstati128 get_osfhandle=rb_w32_get_osfhandle getcwd=rb_w32_ugetcwd getenv=rb_w32_ugetenv gethostbyaddr=rb_w32_gethostbyaddr gethostbyname=rb_w32_gethostbyname gethostname=rb_w32_gethostname getpeername=rb_w32_getpeername getpid=rb_w32_getpid getppid=rb_w32_getppid getprotobyname=rb_w32_getprotobyname getprotobynumber=rb_w32_getprotobynumber getservbyname=rb_w32_getservbyname getservbyport=rb_w32_getservbyport getsockname=rb_w32_getsockname getsockopt=rb_w32_getsockopt inet_ntop=rb_w32_inet_ntop inet_pton=rb_w32_inet_pton ioctlsocket=rb_w32_ioctlsocket isatty=rb_w32_isatty listen=rb_w32_listen lseek=rb_w32_lseek lstat=rb_w32_ulstati128 mkdir=rb_w32_umkdir mmap=rb_w32_mmap mprotect=rb_w32_mprotect munmap=rb_w32_munmap pipe=rb_w32_pipe pread=rb_w32_pread pwrite=rb_w32_pwrite read=rb_w32_read recv=rb_w32_recv recvfrom=rb_w32_recvfrom rename=rb_w32_urename rmdir=rb_w32_urmdir select=rb_w32_select send=rb_w32_send sendto=rb_w32_sendto setsockopt=rb_w32_setsockopt shutdown=rb_w32_shutdown socket=rb_w32_socket stati128=rb_w32_ustati128 strcasecmp=msvcrt.stricmp strerror=rb_w32_strerror strncasecmp=msvcrt.strnicmp times=rb_w32_times unlink=rb_w32_uunlink utime=rb_w32_uutime utimensat=rb_w32_uutimensat utimes=rb_w32_uutimes write=rb_w32_write ---Files-------------------------------- Screenshot 2025-11-06 233918.png (223 KB) primitives.bmp (360 KB) primitives_bad.bmp (360 KB) -- https://bugs.ruby-lang.org/

Issue #21498 has been updated by YO4 (Yoshinao Muramatsu). Ruby has made significant efforts to achieve cross-platform support at the C level. This includes compensating for the lack of C runtime libraries. The statement you quoted is taken out of context and that seems unfair. The override of `fclose` also stems from Microsoft's past decision not to integrate the socket library into low-level I/O. While Microsoft has recently begun focusing on the CLI environment, it has historically prioritized the Win32 API over investing in the C library. For example, time zone names were influenced by the OS language settings rather than the C locale, failing to behave as expected. This required additional support in Ruby. #21144 A cooperative approach is preferable to criticism. I like the process at https://github.com/oneclick/rubyinstaller2/issues/308. The root cause of this issue lies in the coexistence of C runtime DLLs. Will your proposal solve it? Even if the override of `fclose` were removed, who could guarantee the integrity of globally shared resources? `errno`, `locale`, and Runtime support for `_beginthread()`—would these function correctly? Can someone guarantee it? Alternatively, if a mechanism like LD_PRELOAD is available (side-by-side/manifest?), forcing the loading of a C runtime DLL might work around the issue. However, the static-linked startup is present, so it's questionable whether this would work. Therefore, its possibility depends on Microsoft's ecosystem. In any case, the responsibility for ensuring the singleness of the runtime DLL lies with the user. It seems that performing the runtime check on Ruby's side is the maximum that can be done. Are you mistaking someone else's opinion for your own? Even if it's only a band-aid, we should have been able to identify the root cause of this problem earlier. Your cooperation with the investigation paved the way for that. That is undoubtedly an achievement of this issue. One more thing... Next time you encounter a similar situation, you can say it might be due to mixed C runtime DLLs. Cheers! :wink: ---------------------------------------- Bug #21498: Windows - Ruby Overrides C Library APIs thus breaking them https://bugs.ruby-lang.org/issues/21498#change-115241 * Author: cfis (Charlie Savage) * Status: Open * Assignee: windows * ruby -v: *all* on windows (testing with ruby 3.4.3 (2025-04-14 revision d0b7e5b6a0) +PRISM [x64-mswin64_140]) * Backport: 3.2: UNKNOWN, 3.3: UNKNOWN, 3.4: UNKNOWN ---------------------------------------- I am trying to wrap a simple C++ library, https://github.com/baderouaich/BitmapPlusPlus, as a Ruby extension. However when I use the extension to write a bitmap to disk the bitmap is corrupted. This is because the library uses std::ofstream which eventually uses the C API fclose to write the final bytes to the bitmap file and then closes it. The problem is that Ruby overrides `fclose` and replaces it with `rb_w32_fclose`. It then *exports* its custom version from from x64-vcruntime140-ruby340.dll. And the exported version is broken (at least from the standpoint of the C standard library). Note this has been a long standing issue. The first report I see is from 2013: https://bugs.ruby-lang.org/issues/8569 More recently in 2020 (which explains the issue very well): https://github.com/NREL/OpenStudio/issues/3942#issuecomment-610673401 I understand that Ruby is trying to provide a platform independent API. But the problem is this solution breaks any third party libraries that rely on these C APIs (which of course are very common). And there is no good workaround (see https://github.com/NREL/OpenStudio/issues/3942#issuecomment-611072774). So would it be possible for Ruby to stop exporting custom versions of basic C APIs? The code that does it is here: https://github.com/ruby/ruby/blob/master/win32/mkexports.rb#L41 Ruby of course could still use its custom versions of `fclose`, read, write etc within ruby.exe and x64-vcruntime140-ruby340.dll. But they should not be exported from x64-vcruntime140-ruby340.dll and thus be off limits to extensions. If a case comes up where an extension really needs access to rb_w32_fclose instead of fclose then an extension developer can use an `#ifdef _WIN32` to do so and work across platforms. That at least puts the developer in control versus now where I don't see any way I can wrap the bitmap library as a Ruby extension. From my experience the biggest problem is the replacing of fclose with rb_w32_fclose. This is the list of generated overrides: FD_CLR=rb_w32_fdclr FD_ISSET=rb_w32_fdisset Sleep=rb_w32_Sleep accept=rb_w32_accept access=rb_w32_uaccess bind=rb_w32_bind close=rb_w32_close connect=rb_w32_connect dup2=rb_w32_dup2 fclose=rb_w32_fclose fstat=rb_w32_fstati128 get_osfhandle=rb_w32_get_osfhandle getcwd=rb_w32_ugetcwd getenv=rb_w32_ugetenv gethostbyaddr=rb_w32_gethostbyaddr gethostbyname=rb_w32_gethostbyname gethostname=rb_w32_gethostname getpeername=rb_w32_getpeername getpid=rb_w32_getpid getppid=rb_w32_getppid getprotobyname=rb_w32_getprotobyname getprotobynumber=rb_w32_getprotobynumber getservbyname=rb_w32_getservbyname getservbyport=rb_w32_getservbyport getsockname=rb_w32_getsockname getsockopt=rb_w32_getsockopt inet_ntop=rb_w32_inet_ntop inet_pton=rb_w32_inet_pton ioctlsocket=rb_w32_ioctlsocket isatty=rb_w32_isatty listen=rb_w32_listen lseek=rb_w32_lseek lstat=rb_w32_ulstati128 mkdir=rb_w32_umkdir mmap=rb_w32_mmap mprotect=rb_w32_mprotect munmap=rb_w32_munmap pipe=rb_w32_pipe pread=rb_w32_pread pwrite=rb_w32_pwrite read=rb_w32_read recv=rb_w32_recv recvfrom=rb_w32_recvfrom rename=rb_w32_urename rmdir=rb_w32_urmdir select=rb_w32_select send=rb_w32_send sendto=rb_w32_sendto setsockopt=rb_w32_setsockopt shutdown=rb_w32_shutdown socket=rb_w32_socket stati128=rb_w32_ustati128 strcasecmp=msvcrt.stricmp strerror=rb_w32_strerror strncasecmp=msvcrt.strnicmp times=rb_w32_times unlink=rb_w32_uunlink utime=rb_w32_uutime utimensat=rb_w32_uutimensat utimes=rb_w32_uutimes write=rb_w32_write ---Files-------------------------------- Screenshot 2025-11-06 233918.png (223 KB) primitives.bmp (360 KB) primitives_bad.bmp (360 KB) -- https://bugs.ruby-lang.org/

Issue #21498 has been updated by YO4 (Yoshinao Muramatsu). The discussion you cited seemed to me to be ignoring the actual problem by pinning the cause solely on ruby's behavior. You used it to solidify your argument within the ruby community, and even tried to borrow Microsoft's authority to do so. That's how I felt about it. No need to apologize, but we must always keep learning. I still don't understand how calls to overridden functions that are using the original routine could cause problems in a healthy setup. What do you mean 'the problems introduced by the ucrt functions that are overridden'? I don't know what the difficulties faced by the OpenStudio people actually stemmed from, since they haven't provided that information. However, I can examine the OpenStudio-3.1 binaries of that time. ```

objdump -p openstudio.so | grep "DLL Name" DLL Name: openstudiolib.dll DLL Name: x64-msvcrt-ruby250.dll DLL Name: USER32.dll DLL Name: KERNEL32.dll DLL Name: MSVCP140.dll DLL Name: VCRUNTIME140.dll DLL Name: VCRUNTIME140_1.dll DLL Name: api-ms-win-crt-runtime-l1-1-0.dll DLL Name: api-ms-win-crt-stdio-l1-1-0.dll DLL Name: api-ms-win-crt-heap-l1-1-0.dll DLL Name: api-ms-win-crt-string-l1-1-0.dll DLL Name: api-ms-win-crt-time-l1-1-0.dll

I should note that even back then, it was possible to build ruby using ucrt with msvc.

Lets move on to the technical details.
I've already mentioned that ruby has C-level cross-platform support.
For example, winsock library lacks support for errno, but ruby's wrapper makes it usable.
File and socket fds can be handled uniformly on other operating systems, but on windows, you must use _close() and closesocket() appropriately depending on the target.
This is what ruby is currently tackling.

Should we allow selecting whether to include overrides during build time?
The moment we do, we'll be flooded with reports of mysterious errors here and to gem developers. Maybe Windows-specific error reports will flood useful gem developers using other OSes, even though their code is correct.
Furthermore, setup differences could make troubleshooting difficult. I don't want to see a future where gem developers come to hate the Windows platform.
If you continue to propose overriding deletions, I believe you need to provide an alternative solution.

If removing the override made the test pass, then the test is likely insufficient.
As mentioned above, it works fine on other operating systems. They don't need that test.
It's no secret that the Windows version has platform-specific issues and relies on community contributions. We can contribute as well.

The debug version of the Microsoft C runtime library is not redistributable. That is to say, the user is the developer.
What I meant by "user" is that you are responsible for matching the C runtime DLL in this case.

I don't want to burden the ruby core team any further.
My apologies, but the basis for your argument in this issue has been lost, and I believe continuing discussion here would be unproductive.
If you have a better solutions to what ruby is currently doing, or if you have made progress in your own efforts, they would be welcome.

I apologize for the lengthy message.
Regards

----------------------------------------
Bug #21498: Windows - Ruby Overrides C Library APIs thus breaking them
https://bugs.ruby-lang.org/issues/21498#change-115245

* Author: cfis (Charlie Savage)
* Status: Open
* Assignee: windows
* ruby -v: *all* on windows (testing with ruby 3.4.3 (2025-04-14 revision d0b7e5b6a0) +PRISM [x64-mswin64_140])
* Backport: 3.2: UNKNOWN, 3.3: UNKNOWN, 3.4: UNKNOWN
----------------------------------------
I am trying to wrap a simple C++ library, https://github.com/baderouaich/BitmapPlusPlus, as a Ruby extension.

However when I use the extension to write a bitmap to disk the bitmap is corrupted. This is because the library uses std::ofstream which eventually uses the C API fclose to write the final bytes to the bitmap file and then closes it. The problem is that Ruby overrides `fclose` and replaces it with `rb_w32_fclose`. It then *exports* its custom version from from x64-vcruntime140-ruby340.dll. And the exported version is broken (at least from the standpoint of the C standard library).

Note this has been a long standing issue. The first report I see is from 2013:

https://bugs.ruby-lang.org/issues/8569

More recently in 2020 (which explains the issue very well):

https://github.com/NREL/OpenStudio/issues/3942#issuecomment-610673401

I understand that Ruby is trying to provide a platform independent API. But the problem is this solution breaks any third party libraries that rely on these C APIs (which of course are very common). And there is no good workaround (see https://github.com/NREL/OpenStudio/issues/3942#issuecomment-611072774).

So would it be possible for Ruby to stop exporting custom versions of basic C APIs? The code that does it is here:

https://github.com/ruby/ruby/blob/master/win32/mkexports.rb#L41

Ruby of course could still use its custom versions of `fclose`, read, write etc within ruby.exe and x64-vcruntime140-ruby340.dll. But they should not be exported from x64-vcruntime140-ruby340.dll and thus be off limits to extensions. If a case comes up where an extension really needs access to rb_w32_fclose instead of fclose then an extension developer can use an `#ifdef _WIN32` to do so and work across platforms. That at least puts the developer in control versus now where I don't see any way I can wrap the bitmap library as a Ruby extension.

From my experience the biggest problem is the replacing of fclose with rb_w32_fclose.

This is the list of generated overrides:

FD_CLR=rb_w32_fdclr
FD_ISSET=rb_w32_fdisset
Sleep=rb_w32_Sleep
accept=rb_w32_accept
access=rb_w32_uaccess
bind=rb_w32_bind
close=rb_w32_close
connect=rb_w32_connect
dup2=rb_w32_dup2
fclose=rb_w32_fclose
fstat=rb_w32_fstati128
get_osfhandle=rb_w32_get_osfhandle
getcwd=rb_w32_ugetcwd
getenv=rb_w32_ugetenv
gethostbyaddr=rb_w32_gethostbyaddr
gethostbyname=rb_w32_gethostbyname
gethostname=rb_w32_gethostname
getpeername=rb_w32_getpeername
getpid=rb_w32_getpid
getppid=rb_w32_getppid
getprotobyname=rb_w32_getprotobyname
getprotobynumber=rb_w32_getprotobynumber
getservbyname=rb_w32_getservbyname
getservbyport=rb_w32_getservbyport
getsockname=rb_w32_getsockname
getsockopt=rb_w32_getsockopt
inet_ntop=rb_w32_inet_ntop
inet_pton=rb_w32_inet_pton
ioctlsocket=rb_w32_ioctlsocket
isatty=rb_w32_isatty
listen=rb_w32_listen
lseek=rb_w32_lseek
lstat=rb_w32_ulstati128
mkdir=rb_w32_umkdir
mmap=rb_w32_mmap
mprotect=rb_w32_mprotect
munmap=rb_w32_munmap
pipe=rb_w32_pipe
pread=rb_w32_pread
pwrite=rb_w32_pwrite
read=rb_w32_read
recv=rb_w32_recv
recvfrom=rb_w32_recvfrom
rename=rb_w32_urename
rmdir=rb_w32_urmdir
select=rb_w32_select
send=rb_w32_send
sendto=rb_w32_sendto
setsockopt=rb_w32_setsockopt
shutdown=rb_w32_shutdown
socket=rb_w32_socket
stati128=rb_w32_ustati128
strcasecmp=msvcrt.stricmp
strerror=rb_w32_strerror
strncasecmp=msvcrt.strnicmp
times=rb_w32_times
unlink=rb_w32_uunlink
utime=rb_w32_uutime
utimensat=rb_w32_uutimensat
utimes=rb_w32_uutimes
write=rb_w32_write

---Files--------------------------------
Screenshot 2025-11-06 233918.png (223 KB)
primitives.bmp (360 KB)
primitives_bad.bmp (360 KB)


-- 
https://bugs.ruby-lang.org/

Issue #21498 has been updated by cfis (Charlie Savage). Hi YO4 - thank you for your last comment - I appreciate it. I too need to strive for better communication, because clearly I am not communicating very well. I hope we can all assume the best intentions in others, and realize communication, especially online, in tickets and with cultural differences is difficult. I think your comment shows where we are not communicating:

I still don't understand how calls to overridden functions that are using the original routine could cause problems in a healthy setup.

If the Ruby shared library on windows did not include this line in its DEF file - `fclose=rb_w32_fclose` - then the extension would use `fclose` from its debug version of the UCRT and not Ruby's release build of UCRT. Thus the problem is solved, at least in this one case. Second, by "healthy setup" I think you mean that Ruby and all extensions share the same UCRT. I think that is the best case, and likely true 99% of the time, but I am trying to explore if that restriction is absolutely necessary. Because in real life sometimes UCRT's are mismatched (static builds with UCRT, debug builds, before UCRT different C runtimes per MSVC version). Which is why I keep coming back to the same question - does Ruby still need to export (or as I have said above overrides) these function today? Ruby has done amazing work providing a consistent API across platforms, and windows is quite different that other operating systems. Your example of networking and winsock are excellent ones - those clearly are needed to be standardized across operating systems and it makes sense for Ruby to for example export `select`. So I am only discussing the exports that override URCT:

access=rb_w32_uaccess close=rb_w32_close dup2=rb_w32_dup2 fclose=rb_w32_fclose fstat=rb_w32_fstati128 (_fstat) get_osfhandle=rb_w32_get_osfhandle (_get_osfhandle) getpid=rb_w32_getpid isatty=rb_w32_isatty lseek=rb_w32_lseek mkdir=rb_w32_umkdir pipe=rb_w32_pipe (_pipe) read=rb_w32_read rename=rb_w32_urename rmdir=rb_w32_urmdir strerror=rb_w32_strerror unlink=rb_w32_uunlink

I hope that clarifies why I think this is an issue worth of discussion and resolution? I did propose a concrete path forward that I hoped would address your concerns (see if we can use a debug build to catch these errors). Do you think that is worth pursuing? ---------------------------------------- Bug #21498: Windows - Ruby Overrides C Library APIs thus breaking them https://bugs.ruby-lang.org/issues/21498#change-115260 * Author: cfis (Charlie Savage) * Status: Open * Assignee: windows * ruby -v: *all* on windows (testing with ruby 3.4.3 (2025-04-14 revision d0b7e5b6a0) +PRISM [x64-mswin64_140]) * Backport: 3.2: UNKNOWN, 3.3: UNKNOWN, 3.4: UNKNOWN ---------------------------------------- I am trying to wrap a simple C++ library, https://github.com/baderouaich/BitmapPlusPlus, as a Ruby extension. However when I use the extension to write a bitmap to disk the bitmap is corrupted. This is because the library uses std::ofstream which eventually uses the C API fclose to write the final bytes to the bitmap file and then closes it. The problem is that Ruby overrides `fclose` and replaces it with `rb_w32_fclose`. It then *exports* its custom version from from x64-vcruntime140-ruby340.dll. And the exported version is broken (at least from the standpoint of the C standard library). Note this has been a long standing issue. The first report I see is from 2013: https://bugs.ruby-lang.org/issues/8569 More recently in 2020 (which explains the issue very well): https://github.com/NREL/OpenStudio/issues/3942#issuecomment-610673401 I understand that Ruby is trying to provide a platform independent API. But the problem is this solution breaks any third party libraries that rely on these C APIs (which of course are very common). And there is no good workaround (see https://github.com/NREL/OpenStudio/issues/3942#issuecomment-611072774). So would it be possible for Ruby to stop exporting custom versions of basic C APIs? The code that does it is here: https://github.com/ruby/ruby/blob/master/win32/mkexports.rb#L41 Ruby of course could still use its custom versions of `fclose`, read, write etc within ruby.exe and x64-vcruntime140-ruby340.dll. But they should not be exported from x64-vcruntime140-ruby340.dll and thus be off limits to extensions. If a case comes up where an extension really needs access to rb_w32_fclose instead of fclose then an extension developer can use an `#ifdef _WIN32` to do so and work across platforms. That at least puts the developer in control versus now where I don't see any way I can wrap the bitmap library as a Ruby extension. From my experience the biggest problem is the replacing of fclose with rb_w32_fclose. This is the list of generated overrides: FD_CLR=rb_w32_fdclr FD_ISSET=rb_w32_fdisset Sleep=rb_w32_Sleep accept=rb_w32_accept access=rb_w32_uaccess bind=rb_w32_bind close=rb_w32_close connect=rb_w32_connect dup2=rb_w32_dup2 fclose=rb_w32_fclose fstat=rb_w32_fstati128 get_osfhandle=rb_w32_get_osfhandle getcwd=rb_w32_ugetcwd getenv=rb_w32_ugetenv gethostbyaddr=rb_w32_gethostbyaddr gethostbyname=rb_w32_gethostbyname gethostname=rb_w32_gethostname getpeername=rb_w32_getpeername getpid=rb_w32_getpid getppid=rb_w32_getppid getprotobyname=rb_w32_getprotobyname getprotobynumber=rb_w32_getprotobynumber getservbyname=rb_w32_getservbyname getservbyport=rb_w32_getservbyport getsockname=rb_w32_getsockname getsockopt=rb_w32_getsockopt inet_ntop=rb_w32_inet_ntop inet_pton=rb_w32_inet_pton ioctlsocket=rb_w32_ioctlsocket isatty=rb_w32_isatty listen=rb_w32_listen lseek=rb_w32_lseek lstat=rb_w32_ulstati128 mkdir=rb_w32_umkdir mmap=rb_w32_mmap mprotect=rb_w32_mprotect munmap=rb_w32_munmap pipe=rb_w32_pipe pread=rb_w32_pread pwrite=rb_w32_pwrite read=rb_w32_read recv=rb_w32_recv recvfrom=rb_w32_recvfrom rename=rb_w32_urename rmdir=rb_w32_urmdir select=rb_w32_select send=rb_w32_send sendto=rb_w32_sendto setsockopt=rb_w32_setsockopt shutdown=rb_w32_shutdown socket=rb_w32_socket stati128=rb_w32_ustati128 strcasecmp=msvcrt.stricmp strerror=rb_w32_strerror strncasecmp=msvcrt.strnicmp times=rb_w32_times unlink=rb_w32_uunlink utime=rb_w32_uutime utimensat=rb_w32_uutimensat utimes=rb_w32_uutimes write=rb_w32_write ---Files-------------------------------- Screenshot 2025-11-06 233918.png (223 KB) primitives.bmp (360 KB) primitives_bad.bmp (360 KB) -- https://bugs.ruby-lang.org/

Issue #21498 has been updated by cfis (Charlie Savage).

However, to focus on the original issue, I think it would be good to clarify whether the removal of the override targets a single C runtime environment.

The removal of the overrides would target each Ruby extension that is written in C or C++ (but not Rust, GO, etc). So if a Ruby process loads 5 extensions, and each of the extensions is developed in C/C++, then the removal of the overrides would allow each of those extensions to use its own version of the C Runtime and in theory not run into conflicts with Ruby's C Runtime version. If an extension uses the same C Runtime as Ruby (best case and much more common now with the UCRT), then they would share the same copy of the C runtime. Removing the overrides would then make no difference. However, I personally think for philosophical reasons they should still be removed. Maybe a Ruby example will make this clearer - sometimes you need to monkey patch a Ruby class to solve a problem. But if there is an alternative solution then you probably should not monkey patch a class. I do understand others might think differently about this. Hope that helps? ---------------------------------------- Bug #21498: Windows - Ruby Overrides C Library APIs thus breaking them https://bugs.ruby-lang.org/issues/21498#change-115304 * Author: cfis (Charlie Savage) * Status: Open * Assignee: windows * ruby -v: *all* on windows (testing with ruby 3.4.3 (2025-04-14 revision d0b7e5b6a0) +PRISM [x64-mswin64_140]) * Backport: 3.2: UNKNOWN, 3.3: UNKNOWN, 3.4: UNKNOWN ---------------------------------------- I am trying to wrap a simple C++ library, https://github.com/baderouaich/BitmapPlusPlus, as a Ruby extension. However when I use the extension to write a bitmap to disk the bitmap is corrupted. This is because the library uses std::ofstream which eventually uses the C API fclose to write the final bytes to the bitmap file and then closes it. The problem is that Ruby overrides `fclose` and replaces it with `rb_w32_fclose`. It then *exports* its custom version from from x64-vcruntime140-ruby340.dll. And the exported version is broken (at least from the standpoint of the C standard library). Note this has been a long standing issue. The first report I see is from 2013: https://bugs.ruby-lang.org/issues/8569 More recently in 2020 (which explains the issue very well): https://github.com/NREL/OpenStudio/issues/3942#issuecomment-610673401 I understand that Ruby is trying to provide a platform independent API. But the problem is this solution breaks any third party libraries that rely on these C APIs (which of course are very common). And there is no good workaround (see https://github.com/NREL/OpenStudio/issues/3942#issuecomment-611072774). So would it be possible for Ruby to stop exporting custom versions of basic C APIs? The code that does it is here: https://github.com/ruby/ruby/blob/master/win32/mkexports.rb#L41 Ruby of course could still use its custom versions of `fclose`, read, write etc within ruby.exe and x64-vcruntime140-ruby340.dll. But they should not be exported from x64-vcruntime140-ruby340.dll and thus be off limits to extensions. If a case comes up where an extension really needs access to rb_w32_fclose instead of fclose then an extension developer can use an `#ifdef _WIN32` to do so and work across platforms. That at least puts the developer in control versus now where I don't see any way I can wrap the bitmap library as a Ruby extension. From my experience the biggest problem is the replacing of fclose with rb_w32_fclose. This is the list of generated overrides: FD_CLR=rb_w32_fdclr FD_ISSET=rb_w32_fdisset Sleep=rb_w32_Sleep accept=rb_w32_accept access=rb_w32_uaccess bind=rb_w32_bind close=rb_w32_close connect=rb_w32_connect dup2=rb_w32_dup2 fclose=rb_w32_fclose fstat=rb_w32_fstati128 get_osfhandle=rb_w32_get_osfhandle getcwd=rb_w32_ugetcwd getenv=rb_w32_ugetenv gethostbyaddr=rb_w32_gethostbyaddr gethostbyname=rb_w32_gethostbyname gethostname=rb_w32_gethostname getpeername=rb_w32_getpeername getpid=rb_w32_getpid getppid=rb_w32_getppid getprotobyname=rb_w32_getprotobyname getprotobynumber=rb_w32_getprotobynumber getservbyname=rb_w32_getservbyname getservbyport=rb_w32_getservbyport getsockname=rb_w32_getsockname getsockopt=rb_w32_getsockopt inet_ntop=rb_w32_inet_ntop inet_pton=rb_w32_inet_pton ioctlsocket=rb_w32_ioctlsocket isatty=rb_w32_isatty listen=rb_w32_listen lseek=rb_w32_lseek lstat=rb_w32_ulstati128 mkdir=rb_w32_umkdir mmap=rb_w32_mmap mprotect=rb_w32_mprotect munmap=rb_w32_munmap pipe=rb_w32_pipe pread=rb_w32_pread pwrite=rb_w32_pwrite read=rb_w32_read recv=rb_w32_recv recvfrom=rb_w32_recvfrom rename=rb_w32_urename rmdir=rb_w32_urmdir select=rb_w32_select send=rb_w32_send sendto=rb_w32_sendto setsockopt=rb_w32_setsockopt shutdown=rb_w32_shutdown socket=rb_w32_socket stati128=rb_w32_ustati128 strcasecmp=msvcrt.stricmp strerror=rb_w32_strerror strncasecmp=msvcrt.strnicmp times=rb_w32_times unlink=rb_w32_uunlink utime=rb_w32_uutime utimensat=rb_w32_uutimensat utimes=rb_w32_uutimes write=rb_w32_write ---Files-------------------------------- Screenshot 2025-11-06 233918.png (223 KB) primitives.bmp (360 KB) primitives_bad.bmp (360 KB) -- https://bugs.ruby-lang.org/