bitwiseworks/gcc-os2
1
0
Fork 0
mirror of https://github.com/bitwiseworks/gcc-os2.git synced 2026-04-22 22:09:52 +00:00
Code Issues 22 Packages Projects Releases Wiki Activity

Make sure special wchar_t support is enabled #21

New Issue
Closed
opened 2026-01-29 22:23:28 +00:00 by claunia · 9 comments
claunia commented 2026-01-29 22:23:28 +00:00
Owner
Copy Link

Originally created by @dmik on GitHub (Jan 29, 2020).

Originally assigned to: @dmik on GitHub.

From https://github.com/psmedley/gcc/issues/30:

IIRC this is disabled by configure now because kLIBC doesn't implement all necessary wide character functions (so the respective configure test fails).

When enabled, gcc defines _GLIBCXX_USE_WCHAR_T which enables special code for wide characters. This code includes e.g. a std::is_integral<wchar_t> specialization returning true instead of false and a lot of other things.

The problem was found while working on the QTypeInfo<wchar_t> specialization for Qt 5 within https://github.com/bitwiseworks/qtbase-os2/issues/29.

Note that we may force such support with --enable-wchar_t but a test build needs to be done to check if it will work with our kLIBC.

Originally created by @dmik on GitHub (Jan 29, 2020). Originally assigned to: @dmik on GitHub. From https://github.com/psmedley/gcc/issues/30: IIRC this is disabled by configure now because kLIBC doesn't implement all necessary wide character functions (so the respective configure test fails). When enabled, gcc defines `_GLIBCXX_USE_WCHAR_T` which enables special code for wide characters. This code includes e.g. a `std::is_integral<wchar_t>` specialization returning `true` instead of `false` and a lot of other things. The problem was found while working on the `QTypeInfo<wchar_t>` specialization for Qt 5 within https://github.com/bitwiseworks/qtbase-os2/issues/29. Note that we may force such support with `--enable-wchar_t` but a test build needs to be done to check if it will work with our kLIBC.
claunia commented 2026-01-29 22:23:30 +00:00
Author
Owner
Copy Link

@dmik commented on GitHub (Jan 29, 2020):

Read the original ticket for more info. In particular, a few changes should be undone one this issue is resolved. These include:

Note that the LIBC part of wide char support is done within https://github.com/bitwiseworks/libc/issues/8.

Also note that we should keep wchar_t and wint_t definitions in the EMX target for GCC in sync with LIBC. wchar_t is currently fine (unsigned short in both cases) but wint_t is not (int in LIBC and unsigned int in GCC). I will fix it within this ticket as well.

@dmik commented on GitHub (Jan 29, 2020): Read the original ticket for more info. In particular, a few changes should be undone one this issue is resolved. These include: - https://github.com/bitwiseworks/qtbase-os2/commit/9e4df832a62d7fa857fba02e34a71b20dc457bae - https://github.com/bitwiseworks/qtdeclarative-os2/commit/e7a97582cd3174aaa5c8533df37fe86efdc5f740 Note that the LIBC part of wide char support is done within https://github.com/bitwiseworks/libc/issues/8. Also note that we should keep `wchar_t` and `wint_t` definitions in the EMX target for GCC in sync with LIBC. `wchar_t` is currently fine (`unsigned short` in both cases) but `wint_t` is not (`int` in LIBC and `unsigned int` in GCC). I will fix it within this ticket as well.
claunia commented 2026-01-29 22:23:30 +00:00
Author
Owner
Copy Link

@dmik commented on GitHub (Feb 28, 2020):

std::is_integral<wchar_t>::value is false ATM, this breaks Chromium and needs to be dealt with. I guess we must force --enable-wchar_t for now and pray that it doesn't blow somewhere. And if it blows, #if 0 such places with a reference to this ticket as a temporary solution.

@dmik commented on GitHub (Feb 28, 2020): `std::is_integral<wchar_t>::value` is `false` ATM, this breaks Chromium and needs to be dealt with. I guess we must force `--enable-wchar_t` for now and pray that it doesn't blow somewhere. And if it blows, `#if 0` such places with a reference to this ticket as a temporary solution.
claunia commented 2026-01-29 22:23:32 +00:00
Author
Owner
Copy Link

@dmik commented on GitHub (Feb 28, 2020):

With the latest wprinf changes to LIBC, the number of the missing functions has shortened to this in conftest responsible for this setting to be automatically enabled:

t2.cpp:40:12: error: '::fgetwc' has not been declared
   40 |    using ::fgetwc;
      |            ^~~~~~
t2.cpp:41:12: error: '::fgetws' has not been declared
   41 |    using ::fgetws;
      |            ^~~~~~
t2.cpp:43:12: error: '::fputws' has not been declared
   43 |    using ::fputws;
      |            ^~~~~~
t2.cpp:46:12: error: '::fwscanf' has not been declared
   46 |    using ::fwscanf;
      |            ^~~~~~~
t2.cpp:47:12: error: '::getwc' has not been declared
   47 |    using ::getwc;
      |            ^~~~~
t2.cpp:48:12: error: '::getwchar' has not been declared
   48 |    using ::getwchar;
      |            ^~~~~~~~
t2.cpp:56:12: error: '::swscanf' has not been declared
   56 |    using ::swscanf;
      |            ^~~~~~~
t2.cpp:57:12: error: '::ungetwc' has not been declared
   57 |    using ::ungetwc;
      |            ^~~~~~~
t2.cpp:90:12: error: '::wscanf' has not been declared
   90 |    using ::wscanf;
      |            ^~~~~~

fputws was somehow lost when doing 59105a4d89. The rest are scan/get functions that were never done so far. It may take a while to do them.

@dmik commented on GitHub (Feb 28, 2020): With the latest `wprinf` changes to LIBC, the number of the missing functions has shortened to this in conftest responsible for this setting to be automatically enabled: ``` t2.cpp:40:12: error: '::fgetwc' has not been declared 40 | using ::fgetwc; | ^~~~~~ t2.cpp:41:12: error: '::fgetws' has not been declared 41 | using ::fgetws; | ^~~~~~ t2.cpp:43:12: error: '::fputws' has not been declared 43 | using ::fputws; | ^~~~~~ t2.cpp:46:12: error: '::fwscanf' has not been declared 46 | using ::fwscanf; | ^~~~~~~ t2.cpp:47:12: error: '::getwc' has not been declared 47 | using ::getwc; | ^~~~~ t2.cpp:48:12: error: '::getwchar' has not been declared 48 | using ::getwchar; | ^~~~~~~~ t2.cpp:56:12: error: '::swscanf' has not been declared 56 | using ::swscanf; | ^~~~~~~ t2.cpp:57:12: error: '::ungetwc' has not been declared 57 | using ::ungetwc; | ^~~~~~~ t2.cpp:90:12: error: '::wscanf' has not been declared 90 | using ::wscanf; | ^~~~~~ ``` `fputws` was somehow lost when doing https://github.com/bitwiseworks/libc/commit/59105a4d893e83721c4649a0bafd0b9d84d01845. The rest are scan/get functions that were never done so far. It may take a while to do them.
claunia commented 2026-01-29 22:23:32 +00:00
Author
Owner
Copy Link

@dmik commented on GitHub (Feb 28, 2020):

No, forcing --enable-wchar_t w/o bringing the above functions to LIBC is not an option. It brings a lot of side effects. E.g. simply #include <cwchar> in the C++ source code will bring virtually the same error message as above (expected). We have to add those functions to LIBC to continue on Chromium.

@dmik commented on GitHub (Feb 28, 2020): No, forcing `--enable-wchar_t` w/o bringing the above functions to LIBC is not an option. It brings a lot of side effects. E.g. simply `#include <cwchar>` in the C++ source code will bring virtually the same error message as above (expected). We have to add those functions to LIBC to continue on Chromium.
claunia commented 2026-01-29 22:23:33 +00:00
Author
Owner
Copy Link

@dmik commented on GitHub (Mar 4, 2020):

With https://github.com/bitwiseworks/libc/issues/8 resolved, this issue goes away: libstdc++v3 configure can detect wchar_t function presence now and imply --enable-wchar_t by default. Closing this.

@dmik commented on GitHub (Mar 4, 2020): With https://github.com/bitwiseworks/libc/issues/8 resolved, this issue goes away: libstdc++v3 configure can detect wchar_t function presence now and imply `--enable-wchar_t` by default. Closing this.
claunia commented 2026-01-29 22:23:34 +00:00
Author
Owner
Copy Link

@dmik commented on GitHub (Mar 17, 2020):

I need to reopen this as I get build errors when trying to rebuild GCC from scratch. Some conflict with 2eafb765c8 it seems:

ctype_members.cc:137:3: error: redefinition of 'bool std::ctype<wchar_t>::do_is(std::ctype_base::mask, std::ctype<wchar_t>::char_type) const'
  137 |   ctype<wchar_t>::
      |   ^~~~~~~~~~~~~~
In file included from D:/Users/dmik/rpmbuild/BUILD/gcc-9.2.0/libstdc++-v3/include/bits/locale_facets.h:1538,
                 from D:/Users/dmik/rpmbuild/BUILD/gcc-9.2.0/libstdc++-v3/include/std/locale:40,
                 from ctype_members.cc:31:
D:/Users/dmik/rpmbuild/BUILD/gcc-9.2.0/libstdc++-v3/config/os/os2/ctype_inline.h:109:3: note: 'virtual bool std::ctype<wchar_t>::do_is(std::ctype_base::mask, wchar_t) const' previously defined here
  109 |   ctype<wchar_t>::
      |   ^~~~~~~~~~~~~~
ctype_members.cc:155:3: error: redefinition of 'const wchar_t* std::ctype<wchar_t>::do_is(const wchar_t*, const wchar_t*, std::ctype_base::mask*) const'
  155 |   ctype<wchar_t>::
      |   ^~~~~~~~~~~~~~
In file included from D:/Users/dmik/rpmbuild/BUILD/gcc-9.2.0/libstdc++-v3/include/bits/locale_facets.h:1538,
                 from D:/Users/dmik/rpmbuild/BUILD/gcc-9.2.0/libstdc++-v3/include/std/locale:40,
                 from ctype_members.cc:31:
D:/Users/dmik/rpmbuild/BUILD/gcc-9.2.0/libstdc++-v3/config/os/os2/ctype_inline.h:116:3: note: 'virtual const wchar_t* std::ctype<wchar_t>::do_is(const wchar_t*, const wchar_t*, std::ctype_base::mask*) const' previously defined here
  116 |   ctype<wchar_t>::
      |   ^~~~~~~~~~~~~~
ctype_members.cc:173:3: error: redefinition of 'const wchar_t* std::ctype<wchar_t>::do_scan_is(std::ctype_base::mask, const wchar_t*, const wchar_t*) const'
  173 |   ctype<wchar_t>::
      |   ^~~~~~~~~~~~~~
In file included from D:/Users/dmik/rpmbuild/BUILD/gcc-9.2.0/libstdc++-v3/include/bits/locale_facets.h:1538,
                 from D:/Users/dmik/rpmbuild/BUILD/gcc-9.2.0/libstdc++-v3/include/std/locale:40,
                 from ctype_members.cc:31:
D:/Users/dmik/rpmbuild/BUILD/gcc-9.2.0/libstdc++-v3/config/os/os2/ctype_inline.h:140:3: note: 'virtual const wchar_t* std::ctype<wchar_t>::do_scan_is(std::ctype_base::mask, const wchar_t*, const wchar_t*) const' previously defined here
  140 |   ctype<wchar_t>::
      |   ^~~~~~~~~~~~~~
ctype_members.cc:182:3: error: redefinition of 'const wchar_t* std::ctype<wchar_t>::do_scan_not(std::ctype_base::mask, const char_type*, const char_type*) const'
  182 |   ctype<wchar_t>::
      |   ^~~~~~~~~~~~~~
In file included from D:/Users/dmik/rpmbuild/BUILD/gcc-9.2.0/libstdc++-v3/include/bits/locale_facets.h:1538,
                 from D:/Users/dmik/rpmbuild/BUILD/gcc-9.2.0/libstdc++-v3/include/std/locale:40,
                 from ctype_members.cc:31:
D:/Users/dmik/rpmbuild/BUILD/gcc-9.2.0/libstdc++-v3/config/os/os2/ctype_inline.h:149:3: note: 'virtual const wchar_t* std::ctype<wchar_t>::do_scan_not(std::ctype_base::mask, const char_type*, const char_type*) const' previously defined here
  149 |   ctype<wchar_t>::
      |   ^~~~~~~~~~~~~~
ctype_members.cc: In member function 'virtual char std::ctype<wchar_t>::do_narrow(wchar_t, char) const':
ctype_members.cc:212:14: warning: comparison is always true due to limited range of data type [-Wtype-limits]
  212 |     if (__wc >= 0 && __wc < 128 && _M_narrow_ok)
      |         ~~~~~^~~~
ctype_members.cc: In member function 'virtual const wchar_t* std::ctype<wchar_t>::do_narrow(const wchar_t*, const wchar_t*, char, char*) const':
ctype_members.cc:226:14: warning: comparison is always true due to limited range of data type [-Wtype-limits]
  226 |    if (*__lo >= 0 && *__lo < 128)
      |        ~~~~~~^~~~
@dmik commented on GitHub (Mar 17, 2020): I need to reopen this as I get build errors when trying to rebuild GCC from scratch. Some conflict with 2eafb765c8c7e63f7d481b2ac32fa51f34085899 it seems: ``` ctype_members.cc:137:3: error: redefinition of 'bool std::ctype<wchar_t>::do_is(std::ctype_base::mask, std::ctype<wchar_t>::char_type) const' 137 | ctype<wchar_t>:: | ^~~~~~~~~~~~~~ In file included from D:/Users/dmik/rpmbuild/BUILD/gcc-9.2.0/libstdc++-v3/include/bits/locale_facets.h:1538, from D:/Users/dmik/rpmbuild/BUILD/gcc-9.2.0/libstdc++-v3/include/std/locale:40, from ctype_members.cc:31: D:/Users/dmik/rpmbuild/BUILD/gcc-9.2.0/libstdc++-v3/config/os/os2/ctype_inline.h:109:3: note: 'virtual bool std::ctype<wchar_t>::do_is(std::ctype_base::mask, wchar_t) const' previously defined here 109 | ctype<wchar_t>:: | ^~~~~~~~~~~~~~ ctype_members.cc:155:3: error: redefinition of 'const wchar_t* std::ctype<wchar_t>::do_is(const wchar_t*, const wchar_t*, std::ctype_base::mask*) const' 155 | ctype<wchar_t>:: | ^~~~~~~~~~~~~~ In file included from D:/Users/dmik/rpmbuild/BUILD/gcc-9.2.0/libstdc++-v3/include/bits/locale_facets.h:1538, from D:/Users/dmik/rpmbuild/BUILD/gcc-9.2.0/libstdc++-v3/include/std/locale:40, from ctype_members.cc:31: D:/Users/dmik/rpmbuild/BUILD/gcc-9.2.0/libstdc++-v3/config/os/os2/ctype_inline.h:116:3: note: 'virtual const wchar_t* std::ctype<wchar_t>::do_is(const wchar_t*, const wchar_t*, std::ctype_base::mask*) const' previously defined here 116 | ctype<wchar_t>:: | ^~~~~~~~~~~~~~ ctype_members.cc:173:3: error: redefinition of 'const wchar_t* std::ctype<wchar_t>::do_scan_is(std::ctype_base::mask, const wchar_t*, const wchar_t*) const' 173 | ctype<wchar_t>:: | ^~~~~~~~~~~~~~ In file included from D:/Users/dmik/rpmbuild/BUILD/gcc-9.2.0/libstdc++-v3/include/bits/locale_facets.h:1538, from D:/Users/dmik/rpmbuild/BUILD/gcc-9.2.0/libstdc++-v3/include/std/locale:40, from ctype_members.cc:31: D:/Users/dmik/rpmbuild/BUILD/gcc-9.2.0/libstdc++-v3/config/os/os2/ctype_inline.h:140:3: note: 'virtual const wchar_t* std::ctype<wchar_t>::do_scan_is(std::ctype_base::mask, const wchar_t*, const wchar_t*) const' previously defined here 140 | ctype<wchar_t>:: | ^~~~~~~~~~~~~~ ctype_members.cc:182:3: error: redefinition of 'const wchar_t* std::ctype<wchar_t>::do_scan_not(std::ctype_base::mask, const char_type*, const char_type*) const' 182 | ctype<wchar_t>:: | ^~~~~~~~~~~~~~ In file included from D:/Users/dmik/rpmbuild/BUILD/gcc-9.2.0/libstdc++-v3/include/bits/locale_facets.h:1538, from D:/Users/dmik/rpmbuild/BUILD/gcc-9.2.0/libstdc++-v3/include/std/locale:40, from ctype_members.cc:31: D:/Users/dmik/rpmbuild/BUILD/gcc-9.2.0/libstdc++-v3/config/os/os2/ctype_inline.h:149:3: note: 'virtual const wchar_t* std::ctype<wchar_t>::do_scan_not(std::ctype_base::mask, const char_type*, const char_type*) const' previously defined here 149 | ctype<wchar_t>:: | ^~~~~~~~~~~~~~ ctype_members.cc: In member function 'virtual char std::ctype<wchar_t>::do_narrow(wchar_t, char) const': ctype_members.cc:212:14: warning: comparison is always true due to limited range of data type [-Wtype-limits] 212 | if (__wc >= 0 && __wc < 128 && _M_narrow_ok) | ~~~~~^~~~ ctype_members.cc: In member function 'virtual const wchar_t* std::ctype<wchar_t>::do_narrow(const wchar_t*, const wchar_t*, char, char*) const': ctype_members.cc:226:14: warning: comparison is always true due to limited range of data type [-Wtype-limits] 226 | if (*__lo >= 0 && *__lo < 128) | ~~~~~~^~~~ ```
claunia commented 2026-01-29 22:23:35 +00:00
Author
Owner
Copy Link

@dmik commented on GitHub (Mar 18, 2020):

Got one more problem when linking STDCPP6.DLL now:

weakld: error: Unresolved symbol (UNDEF) '_wcsxfrm'.
weakld: info: The symbol is referenced by:
    D:\Temp\ldconv_collate_members_o_6ff95e7155bf183d60.obj
    D:\Temp\ldconv_collate_members_cow_o_6ff95e7155bf187fc8.obj
Ignoring unresolved externals reported from weak prelinker.
Error! E2028: _wcsxfrm is an undefined reference
file D:/Temp\ldconv_collate_members_o_6ff95e7155bf183d60.obj(collate_members.cc): undefined symbol _wcsxfrm
file D:/Temp\ldconv_collate_members_cow_o_6ff95e7155bf187fc8.obj(collate_members_cow.cc): undefined symbol _wcsxfrm

Which means that this one was not required by configure so it didn't fall. Sigh. I will have to add it to LIBC or make GCC not use it. I.e. more time on this.

@dmik commented on GitHub (Mar 18, 2020): Got one more problem when linking STDCPP6.DLL now: ``` weakld: error: Unresolved symbol (UNDEF) '_wcsxfrm'. weakld: info: The symbol is referenced by: D:\Temp\ldconv_collate_members_o_6ff95e7155bf183d60.obj D:\Temp\ldconv_collate_members_cow_o_6ff95e7155bf187fc8.obj Ignoring unresolved externals reported from weak prelinker. Error! E2028: _wcsxfrm is an undefined reference file D:/Temp\ldconv_collate_members_o_6ff95e7155bf183d60.obj(collate_members.cc): undefined symbol _wcsxfrm file D:/Temp\ldconv_collate_members_cow_o_6ff95e7155bf187fc8.obj(collate_members_cow.cc): undefined symbol _wcsxfrm ``` Which means that this one was not required by configure so it didn't fall. Sigh. I will have to add it to LIBC or make GCC not use it. I.e. more time on this.
claunia commented 2026-01-29 22:23:35 +00:00
Author
Owner
Copy Link

@dmik commented on GitHub (Mar 19, 2020):

Now all builds but I've faced https://github.com/bitwiseworks/libc/issues/35 during another round. Need to investigate it further.

@dmik commented on GitHub (Mar 19, 2020): Now all builds but I've faced https://github.com/bitwiseworks/libc/issues/35 during another round. Need to investigate it further.
claunia commented 2026-01-29 22:23:36 +00:00
Author
Owner
Copy Link

@dmik commented on GitHub (Mar 19, 2020):

Couldn't move further with the make break investigation, it just worked upon the next time. This is irrelevant to this ticket though so I'm going to close it — rebuilt GCC fully supports wchar_t in all my tests (QtBase/QtDeclarative with rolled back hacks, QtWebEngine, even wregex works).

@dmik commented on GitHub (Mar 19, 2020): Couldn't move further with the make break investigation, it just worked upon the next time. This is irrelevant to this ticket though so I'm going to close it — rebuilt GCC fully supports wchar_t in all my tests (QtBase/QtDeclarative with rolled back hacks, QtWebEngine, even wregex works).
Sign in to join this conversation.
No Branch/Tag Specified
Branches Tags
Labels
Clear labels
postponed
No Label
Milestone
No items
No Milestone
Projects
Clear projects
No project
Assignees
Clear assignees
claunia
No Assignees
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: bitwiseworks/gcc-os2#21
Delete Branch "%!s()"

Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?