From: Scott C. <can...@os...> - 2007-08-24 06:21:35
|
> Index: include/log4cpp/FixedContextCategory.hh > =================================================================== > RCS file: > /cvsroot/log4cpp/log4cpp/include/log4cpp/FixedContextCategory.hh,v > retrieving revision 1.12 > diff -r1.12 FixedContextCategory.hh > 114a115,116 > > using Category::ownsAppender; Cancel that fix, for some reason it worked or seemed to the first time I rebuilt, and now it's failing, so we can leave it with the warning and forget it. There are a number of makefile issues with headers missing from the package if I build from make dist, so I'll report the full set of fixes to that once it's working, along with some RPM spec improvements. -- Scott |
From: Konstantin L. <1da...@ma...> - 2007-08-24 06:47:23
|
Hello, Scott! You wrote to <log...@li...> on Fri, 24 Aug 2007 02:21:30 -0400: ??>> Index: include/log4cpp/FixedContextCategory.hh ??>> =================================================================== ??>> RCS file: ??>> /cvsroot/log4cpp/log4cpp/include/log4cpp/FixedContextCategory.hh,v ??>> retrieving revision 1.12 ??>> diff -r1.12 FixedContextCategory.hh ??>> 114a115,116 ??>>> using Category::ownsAppender; SC> Cancel that fix, for some reason it worked or seemed to the first time SC> I rebuilt, and now it's failing, so we can leave it with the warning SC> and forget it. This is strange. Fix is seams OK. Anyway, I have looked on this and I can say that ownsAppender is broken in this context. To work as intended this function must be virtual. So I will fix it. SC> There are a number of makefile issues with headers missing from the SC> package if I build from make dist, so I'll report the full set of fixes SC> to that once it's working, along with some RPM spec improvements. OK. Thanks ;) With best regards, Konstantin Litvinenko. |
From: Scott C. <can...@os...> - 2007-08-24 07:03:51
|
> This is strange. Fix is seams OK. Anyway, I have looked on this and I > can say that ownsAppender is broken in this context. To work as intended > this function must be virtual. So I will fix it. Usually "using" works for me when there's a single function override signature, but there were overloads here, so I probably could qualify it, but whatever. Here are the next round of packaging cleanups. I added two missing headers to the makefiles, and revised log4cpp.spec.in using a process I'm trying on several libraries. I'm making them work on Solaris, which reveals a lot of hidden Linux assumptions and missing macros in the specfile, and verifies that it will work without gcc/g++ and when installed to a different root. This specfile was in good shape, I just fixed a few small issues. The side effect is that you can actually install it as an RPM on Solaris using the Sun supplied rpm port. -- Scott Index: include/log4cpp/Makefile.am =================================================================== RCS file: /cvsroot/log4cpp/log4cpp/include/log4cpp/Makefile.am,v retrieving revision 1.40 diff -r1.40 Makefile.am 45a46 > convenience.h \ Index: src/Makefile.am =================================================================== RCS file: /cvsroot/log4cpp/log4cpp/src/Makefile.am,v retrieving revision 1.36 diff -r1.36 Makefile.am 5c5 < noinst_HEADERS = snprintf.c --- > noinst_HEADERS = snprintf.c Localtime.hh Index: log4cpp.spec.in =================================================================== RCS file: /cvsroot/log4cpp/log4cpp/log4cpp.spec.in,v retrieving revision 1.12 diff -r1.12 log4cpp.spec.in 12c12 < Copyright: LGPL --- > License: LGPL 41a42 > %{!?_without_doxygenrpm:BuildRequires: doxygen} 48c49 < rm -rf $RPM_BUILD_ROOT --- > %{__rm} -rf $RPM_BUILD_ROOT 51c52 < CFLAGS="$RPM_OPT_FLAGS" ./configure --prefix=%{prefix} --enable-doxygen --- > CC=%{__cc} CXX=%{__cxx} ./configure --prefix=%{prefix} --enable-doxygen 54c55 < make --- > %{__make} 57c58 < rm -rf $RPM_BUILD_ROOT --- > %{__rm} -rf $RPM_BUILD_ROOT 59c60 < make prefix=$RPM_BUILD_ROOT%{prefix} docdir=$RPM_BUILD_ROOT/%{manualdir} install --- > %{__make} prefix=$RPM_BUILD_ROOT%{prefix} docdir=$RPM_BUILD_ROOT/%{manualdir} install 61c62 < rm -rf $RPM_BUILD_ROOT --- > %{__rm} -rf $RPM_BUILD_ROOT 63,64c64,66 < %post < /sbin/ldconfig --- > %ifnos solaris2.8 solaris2.9 solaris2.10 > %post -p /sbin/ldconfig > %endif 68,69c70,71 < perl -pi -e "s|^libdir='[^\']*'|libdir='$RPM_INSTALL_PREFIX0/lib'|" $RPM_INSTALL_PREFIX0/lib/liblog4cpp.la < perl -pi -e "s|^prefix=\"[^\"]*\"|prefix=\"$RPM_INSTALL_PREFIX0\"|" $RPM_INSTALL_PREFIX0/bin/log4cpp-config --- > %{__perl} -pi -e "s|^libdir='[^\']*'|libdir='$RPM_INSTALL_PREFIX0/lib'|" $RPM_INSTALL_PREFIX0/lib/liblog4cpp.la > %{__perl} -pi -e "s|^prefix=\"[^\"]*\"|prefix=\"$RPM_INSTALL_PREFIX0\"|" $RPM_INSTALL_PREFIX0/bin/log4cpp-config 71,73c73,76 < < %postun < /sbin/ldconfig --- > > %ifnos solaris2.8 solaris2.9 solaris2.10 > %postun -p /sbin/ldconfig > %endif |
From: Konstantin L. <1da...@ma...> - 2007-08-25 09:44:06
|
Hello, Scott! You wrote to "'Konstantin Litvinenko'" <1da...@ma...>; <log...@li...> on Fri, 24 Aug 2007 03:03:45 -0400: SC> Here are the next round of packaging cleanups. I added two missing SC> headers to the makefiles, and revised log4cpp.spec.in using a process SC> I'm trying on several libraries. I'm making them work on Solaris, which SC> reveals a lot of hidden Linux assumptions and missing macros in the SC> specfile, and verifies that it will work without gcc/g++ and when SC> installed to a different root. SC> This specfile was in good shape, I just fixed a few small issues. SC> The side effect is that you can actually install it as an RPM on SC> Solaris using the Sun supplied rpm port. Fixed. With best regards, Konstantin Litvinenko. |