| Message ID | 20170725144450.30516-1-aconole@bytheb.org |
|---|---|
| State | Accepted |
| Delegated to: | Aaron Conole |
| Headers | show |
Aaron Conole <aconole@bytheb.org> writes: > When cross-compiling the system, it is not possible for the esnacc binary to > correctly execute on the build system (well, not always). There is a > method to combat this, which is the approach gcc takes, to generate a stage > one version of the compiler and then execute that to generate the final > tools (stage 2). That represents quite a bit of work. > > On the other hand, people are using esnacc *now* and as Marty Galligan > reports at http://mail.esnacc.org/pipermail/dev/2017-July/000414.html there > can be problems cross building, especially related to this. > > This commit detects that a cross-compilation is occuring and disables the > generation of the asn.1 ROSE support. A future commit can enable this > when it becomes necessary. > > Reported-by: Marty Galligan <m_galligan@comcast.net> > Signed-off-by: Aaron Conole <aconole@bytheb.org> > --- So, you'll notice that this is implemented without using --enable / --disable for the ASN-ROSE feature. If someone thinks that's a better approach, I'm okay to include it, but I don't like being able to configure away these kinds of built-in features (IE: there isn't any real reason a user would ever build without --enable-asnrose - it's a separate library in C++). -Aaron
Are you planning to push this to master? Or do I need to ACK the change, or is it just on one branch (which one)? -- -Marty Galligan, m_galligan@comcast.net, c: 978-505-5185 > > On July 25, 2017 at 10:44 AM Aaron Conole <aconole@bytheb.org> wrote: > > When cross-compiling the system, it is not possible for the esnacc binary to > correctly execute on the build system (well, not always). There is a > method to combat this, which is the approach gcc takes, to generate a stage > one version of the compiler and then execute that to generate the final > tools (stage 2). That represents quite a bit of work. > > On the other hand, people are using esnacc *now* and as Marty Galligan > reports at http://mail.esnacc.org/pipermail/dev/2017-July/000414.html there > can be problems cross building, especially related to this. > > This commit detects that a cross-compilation is occuring and disables the > generation of the asn.1 ROSE support. A future commit can enable this > when it becomes necessary. > > Reported-by: Marty Galligan <m_galligan@comcast.net> > > Signed-off-by: Aaron Conole <aconole@bytheb.org> > > --- > configure.ac | 4 ++++ > cxx-lib/automake.mk | 9 ++++++--- > 2 files changed, 10 insertions(+), 3 deletions(-) > > diff --git a/configure.ac b/configure.ac > index 26007ef..65b4b48 100644 > --- a/configure.ac > +++ b/configure.ac > @@ -32,6 +32,10 @@ AC_PROG_INSTALL > AM_INIT_AUTOMAKE([-Wall foreign subdir-objects]) > AC_PROG_CC > AC_PROG_CXX > + > +AM_CONDITIONAL([SNACCROSE], [test "$cross_compiling" = no]) > +test x"$cross_compiling" == xyes && AC_DEFINE([SNACCROSE], [0], [No ROSE support]) > + > AM_PATH_PYTHON(,, [:]) > AM_CONDITIONAL([HAVE_PYTHON], [test "$PYTHON" != :]) > AC_USE_SYSTEM_EXTENSIONS > diff --git a/cxx-lib/automake.mk b/cxx-lib/automake.mk > index 2c5b7ea..fc841ca 100644 > --- a/cxx-lib/automake.mk > +++ b/cxx-lib/automake.mk > @@ -1,5 +1,9 @@ > -lib_LTLIBRARIES += cxx-lib/libcxxasn1.la \ > > * cxx-lib/libcxxasn1rose.la > +lib_LTLIBRARIES += cxx-lib/libcxxasn1.la > + > +if SNACCROSE > +lib_LTLIBRARIES += cxx-lib/libcxxasn1rose.la > +nobase_include_HEADERS += cxx-lib/inc/snaccrose.h > +endif > > BUILT_SOURCES += cxx-lib/inc/snacc.h > > @@ -17,7 +21,6 @@ nobase_include_HEADERS += cxx-lib/inc/asn-buf.h \ > cxx-lib/inc/snacc.h \ > cxx-lib/inc/snaccdll.h \ > cxx-lib/inc/snaccexcept.h \ > > * cxx-lib/inc/snaccrose.h \ > cxx-lib/inc/tcl-if.h > > cxx_lib_libcxxasn1_la_SOURCES = \ > -- > 2.9.4 > > _______________________________________________ > dev mailing list > dev@lists.esnacc.org > http://mail.esnacc.org/mailman/listinfo/dev > <!DOCTYPE html> <html><head> <meta charset="UTF-8"> </head><body><p class="io-ox-signature">Are you planning to push this to master?  Or do I need to ACK the change, or is it just on one branch (which one)?</p><p class="io-ox-signature"><br></p><p class="io-ox-signature">--<br>-Marty Galligan, m_galligan@comcast.net, c: 978-505-5185</p><p><br></p><blockquote><p>On July 25, 2017 at 10:44 AM Aaron Conole <aconole@bytheb.org> wrote:</p><p>When cross-compiling the system, it is not possible for the esnacc binary to<br>correctly execute on the build system (well, not always). There is a<br>method to combat this, which is the approach gcc takes, to generate a stage<br>one version of the compiler and then execute that to generate the final<br>tools (stage 2). That represents quite a bit of work.</p><p>On the other hand, people are using esnacc *now* and as Marty Galligan<br>reports at <a href="http://mail.esnacc.org/pipermail/dev/2017-July/000414.html">http://mail.esnacc.org/pipermail/dev/2017-July/000414.html</a> there<br>can be problems cross building, especially related to this.</p><p>This commit detects that a cross-compilation is occuring and disables the<br>generation of the asn.1 ROSE support. A future commit can enable this<br>when it becomes necessary.</p><p>Reported-by: Marty Galligan <m_galligan@comcast.net></p><p>Signed-off-by: Aaron Conole <aconole@bytheb.org></p><p>---<br> configure.ac | 4 ++++<br> cxx-lib/automake.mk | 9 ++++++---<br> 2 files changed, 10 insertions(+), 3 deletions(-)</p><p>diff --git a/configure.ac b/configure.ac<br>index 26007ef..65b4b48 100644<br>--- a/configure.ac<br>+++ b/configure.ac<br>@@ -32,6 +32,10 @@ AC_PROG_INSTALL<br> AM_INIT_AUTOMAKE([-Wall foreign subdir-objects])<br> AC_PROG_CC<br> AC_PROG_CXX<br>+<br>+AM_CONDITIONAL([SNACCROSE], [test "$cross_compiling" = no])<br>+test x"$cross_compiling" == xyes && AC_DEFINE([SNACCROSE], [0], [No ROSE support])<br>+<br> AM_PATH_PYTHON(,, [:])<br> AM_CONDITIONAL([HAVE_PYTHON], [test "$PYTHON" != :])<br> AC_USE_SYSTEM_EXTENSIONS<br>diff --git a/cxx-lib/automake.mk b/cxx-lib/automake.mk<br>index 2c5b7ea..fc841ca 100644<br>--- a/cxx-lib/automake.mk<br>+++ b/cxx-lib/automake.mk<br>@@ -1,5 +1,9 @@<br>-lib_LTLIBRARIES += cxx-lib/libcxxasn1.la \</p><ul><li><p>cxx-lib/libcxxasn1rose.la<br>+lib_LTLIBRARIES += cxx-lib/libcxxasn1.la<br>+<br>+if SNACCROSE<br>+lib_LTLIBRARIES += cxx-lib/libcxxasn1rose.la<br>+nobase_include_HEADERS += cxx-lib/inc/snaccrose.h<br>+endif</p><p>BUILT_SOURCES += cxx-lib/inc/snacc.h</p></li></ul><p>@@ -17,7 +21,6 @@ nobase_include_HEADERS += cxx-lib/inc/asn-buf.h \<br> cxx-lib/inc/snacc.h \<br> cxx-lib/inc/snaccdll.h \<br> cxx-lib/inc/snaccexcept.h \</p><ul><li><p>cxx-lib/inc/snaccrose.h \<br>cxx-lib/inc/tcl-if.h</p><p>cxx_lib_libcxxasn1_la_SOURCES = \<br>--<br>2.9.4</p></li></ul><p>_______________________________________________<br>dev mailing list<br>dev@lists.esnacc.org<br><a href="http://mail.esnacc.org/mailman/listinfo/dev">http://mail.esnacc.org/mailman/listinfo/dev</a></p></blockquote></body></html>
Marty Galligan <m_galligan@comcast.net> writes: > Are you planning to push this to master? Or do I need to ACK the change, or is it just on one branch > (which one)? I do plan to push this to master. If you post an acked-by tag, then I will apply it in the next batch, along with your tag. You should understand what 'Acked-by' means, though. You can always see the outstanding patches at: http://patchwork.esnacc.org Even further, you can download them, apply them, and see the discussion surrounding them. -Aaron > -- > -Marty Galligan, m_galligan@comcast.net, c: 978-505-5185 > > On July 25, 2017 at 10:44 AM Aaron Conole <aconole@bytheb.org> wrote: > > When cross-compiling the system, it is not possible for the esnacc binary to > correctly execute on the build system (well, not always). There is a > method to combat this, which is the approach gcc takes, to generate a stage > one version of the compiler and then execute that to generate the final > tools (stage 2). That represents quite a bit of work. > > On the other hand, people are using esnacc *now* and as Marty Galligan > reports at http://mail.esnacc.org/pipermail/dev/2017-July/000414.html there > can be problems cross building, especially related to this. > > This commit detects that a cross-compilation is occuring and disables the > generation of the asn.1 ROSE support. A future commit can enable this > when it becomes necessary. > > Reported-by: Marty Galligan <m_galligan@comcast.net> > > Signed-off-by: Aaron Conole <aconole@bytheb.org> > > --- > configure.ac | 4 ++++ > cxx-lib/automake.mk | 9 ++++++--- > 2 files changed, 10 insertions(+), 3 deletions(-) > > diff --git a/configure.ac b/configure.ac > index 26007ef..65b4b48 100644 > --- a/configure.ac > +++ b/configure.ac > @@ -32,6 +32,10 @@ AC_PROG_INSTALL > AM_INIT_AUTOMAKE([-Wall foreign subdir-objects]) > AC_PROG_CC > AC_PROG_CXX > + > +AM_CONDITIONAL([SNACCROSE], [test "$cross_compiling" = no]) > +test x"$cross_compiling" == xyes && AC_DEFINE([SNACCROSE], [0], [No ROSE support]) > + > AM_PATH_PYTHON(,, [:]) > AM_CONDITIONAL([HAVE_PYTHON], [test "$PYTHON" != :]) > AC_USE_SYSTEM_EXTENSIONS > diff --git a/cxx-lib/automake.mk b/cxx-lib/automake.mk > index 2c5b7ea..fc841ca 100644 > --- a/cxx-lib/automake.mk > +++ b/cxx-lib/automake.mk > @@ -1,5 +1,9 @@ > -lib_LTLIBRARIES += cxx-lib/libcxxasn1.la \ > > * cxx-lib/libcxxasn1rose.la > +lib_LTLIBRARIES += cxx-lib/libcxxasn1.la > + > +if SNACCROSE > +lib_LTLIBRARIES += cxx-lib/libcxxasn1rose.la > +nobase_include_HEADERS += cxx-lib/inc/snaccrose.h > +endif > > BUILT_SOURCES += cxx-lib/inc/snacc.h > > @@ -17,7 +21,6 @@ nobase_include_HEADERS += cxx-lib/inc/asn-buf.h \ > cxx-lib/inc/snacc.h \ > cxx-lib/inc/snaccdll.h \ > cxx-lib/inc/snaccexcept.h \ > > * cxx-lib/inc/snaccrose.h \ > cxx-lib/inc/tcl-if.h > > cxx_lib_libcxxasn1_la_SOURCES = \ > -- > 2.9.4 > > _______________________________________________ > dev mailing list > dev@lists.esnacc.org > http://mail.esnacc.org/mailman/listinfo/dev > > _______________________________________________ > dev mailing list > dev@lists.esnacc.org > http://mail.esnacc.org/mailman/listinfo/dev
ack. I manually put this patch into my checkout from github, and the cross compile worked well. So I'm in favor of committing the change. -Marty At 10:44 AM 7/25/2017, Aaron Conole wrote: >When cross-compiling the system, it is not possible for the esnacc binary to >correctly execute on the build system (well, not always). There is a >method to combat this, which is the approach gcc takes, to generate a stage >one version of the compiler and then execute that to generate the final >tools (stage 2). That represents quite a bit of work. > >On the other hand, people are using esnacc *now* and as Marty Galligan >reports at http://mail.esnacc.org/pipermail/dev/2017-July/000414.html there >can be problems cross building, especially related to this. > >This commit detects that a cross-compilation is occuring and disables the >generation of the asn.1 ROSE support. A future commit can enable this >when it becomes necessary. > >Reported-by: Marty Galligan <m_galligan@comcast.net> >Signed-off-by: Aaron Conole <aconole@bytheb.org> >--- > configure.ac | 4 ++++ > cxx-lib/automake.mk | 9 ++++++--- > 2 files changed, 10 insertions(+), 3 deletions(-) > >diff --git a/configure.ac b/configure.ac >index 26007ef..65b4b48 100644 >--- a/configure.ac >+++ b/configure.ac >@@ -32,6 +32,10 @@ AC_PROG_INSTALL > AM_INIT_AUTOMAKE([-Wall foreign subdir-objects]) > AC_PROG_CC > AC_PROG_CXX >+ >+AM_CONDITIONAL([SNACCROSE], [test "$cross_compiling" = no]) >+test x"$cross_compiling" == xyes && AC_DEFINE([SNACCROSE], [0], [No >ROSE support]) >+ > AM_PATH_PYTHON(,, [:]) > AM_CONDITIONAL([HAVE_PYTHON], [test "$PYTHON" != :]) > AC_USE_SYSTEM_EXTENSIONS >diff --git a/cxx-lib/automake.mk b/cxx-lib/automake.mk >index 2c5b7ea..fc841ca 100644 >--- a/cxx-lib/automake.mk >+++ b/cxx-lib/automake.mk >@@ -1,5 +1,9 @@ >-lib_LTLIBRARIES += cxx-lib/libcxxasn1.la \ >- cxx-lib/libcxxasn1rose.la >+lib_LTLIBRARIES += cxx-lib/libcxxasn1.la >+ >+if SNACCROSE >+lib_LTLIBRARIES += cxx-lib/libcxxasn1rose.la >+nobase_include_HEADERS += cxx-lib/inc/snaccrose.h >+endif > > BUILT_SOURCES += cxx-lib/inc/snacc.h > >@@ -17,7 +21,6 @@ nobase_include_HEADERS += cxx-lib/inc/asn-buf.h \ > cxx-lib/inc/snacc.h \ > cxx-lib/inc/snaccdll.h \ > cxx-lib/inc/snaccexcept.h \ >- cxx-lib/inc/snaccrose.h \ > cxx-lib/inc/tcl-if.h > > cxx_lib_libcxxasn1_la_SOURCES = \ >-- >2.9.4 > >_______________________________________________ >dev mailing list >dev@lists.esnacc.org >http://mail.esnacc.org/mailman/listinfo/dev -Marty Galligan, m_galligan@comcast.net, c: 978-505-5185
Marty Galligan <m_galligan@comcast.net> writes: > ack. > > I manually put this patch into my checkout from github, and the cross > compile worked well. So I'm in favor of committing the change. Applied. Thanks, Marty. I've also added you to the AUTHORS file under the Reporters heading. > -Marty > > At 10:44 AM 7/25/2017, Aaron Conole wrote: >>When cross-compiling the system, it is not possible for the esnacc binary to >>correctly execute on the build system (well, not always). There is a >>method to combat this, which is the approach gcc takes, to generate a stage >>one version of the compiler and then execute that to generate the final >>tools (stage 2). That represents quite a bit of work. >> >>On the other hand, people are using esnacc *now* and as Marty Galligan >>reports at http://mail.esnacc.org/pipermail/dev/2017-July/000414.html there >>can be problems cross building, especially related to this. >> >>This commit detects that a cross-compilation is occuring and disables the >>generation of the asn.1 ROSE support. A future commit can enable this >>when it becomes necessary. >> >>Reported-by: Marty Galligan <m_galligan@comcast.net> >>Signed-off-by: Aaron Conole <aconole@bytheb.org> >>--- >> configure.ac | 4 ++++ >> cxx-lib/automake.mk | 9 ++++++--- >> 2 files changed, 10 insertions(+), 3 deletions(-) >> >>diff --git a/configure.ac b/configure.ac >>index 26007ef..65b4b48 100644 >>--- a/configure.ac >>+++ b/configure.ac >>@@ -32,6 +32,10 @@ AC_PROG_INSTALL >> AM_INIT_AUTOMAKE([-Wall foreign subdir-objects]) >> AC_PROG_CC >> AC_PROG_CXX >>+ >>+AM_CONDITIONAL([SNACCROSE], [test "$cross_compiling" = no]) >> +test x"$cross_compiling" == xyes && AC_DEFINE([SNACCROSE], [0], [No >> ROSE support]) >>+ >> AM_PATH_PYTHON(,, [:]) >> AM_CONDITIONAL([HAVE_PYTHON], [test "$PYTHON" != :]) >> AC_USE_SYSTEM_EXTENSIONS >>diff --git a/cxx-lib/automake.mk b/cxx-lib/automake.mk >>index 2c5b7ea..fc841ca 100644 >>--- a/cxx-lib/automake.mk >>+++ b/cxx-lib/automake.mk >>@@ -1,5 +1,9 @@ >>-lib_LTLIBRARIES += cxx-lib/libcxxasn1.la \ >>- cxx-lib/libcxxasn1rose.la >>+lib_LTLIBRARIES += cxx-lib/libcxxasn1.la >>+ >>+if SNACCROSE >>+lib_LTLIBRARIES += cxx-lib/libcxxasn1rose.la >>+nobase_include_HEADERS += cxx-lib/inc/snaccrose.h >>+endif >> >> BUILT_SOURCES += cxx-lib/inc/snacc.h >> >>@@ -17,7 +21,6 @@ nobase_include_HEADERS += cxx-lib/inc/asn-buf.h \ >> cxx-lib/inc/snacc.h \ >> cxx-lib/inc/snaccdll.h \ >> cxx-lib/inc/snaccexcept.h \ >>- cxx-lib/inc/snaccrose.h \ >> cxx-lib/inc/tcl-if.h >> >> cxx_lib_libcxxasn1_la_SOURCES = \ >>-- >>2.9.4 >> >>_______________________________________________ >>dev mailing list >>dev@lists.esnacc.org >>http://mail.esnacc.org/mailman/listinfo/dev > > -Marty Galligan, m_galligan@comcast.net, c: 978-505-5185 > > _______________________________________________ > dev mailing list > dev@lists.esnacc.org > http://mail.esnacc.org/mailman/listinfo/dev
diff --git a/configure.ac b/configure.ac index 26007ef..65b4b48 100644 --- a/configure.ac +++ b/configure.ac @@ -32,6 +32,10 @@ AC_PROG_INSTALL AM_INIT_AUTOMAKE([-Wall foreign subdir-objects]) AC_PROG_CC AC_PROG_CXX + +AM_CONDITIONAL([SNACCROSE], [test "$cross_compiling" = no]) +test x"$cross_compiling" == xyes && AC_DEFINE([SNACCROSE], [0], [No ROSE support]) + AM_PATH_PYTHON(,, [:]) AM_CONDITIONAL([HAVE_PYTHON], [test "$PYTHON" != :]) AC_USE_SYSTEM_EXTENSIONS diff --git a/cxx-lib/automake.mk b/cxx-lib/automake.mk index 2c5b7ea..fc841ca 100644 --- a/cxx-lib/automake.mk +++ b/cxx-lib/automake.mk @@ -1,5 +1,9 @@ -lib_LTLIBRARIES += cxx-lib/libcxxasn1.la \ - cxx-lib/libcxxasn1rose.la +lib_LTLIBRARIES += cxx-lib/libcxxasn1.la + +if SNACCROSE +lib_LTLIBRARIES += cxx-lib/libcxxasn1rose.la +nobase_include_HEADERS += cxx-lib/inc/snaccrose.h +endif BUILT_SOURCES += cxx-lib/inc/snacc.h @@ -17,7 +21,6 @@ nobase_include_HEADERS += cxx-lib/inc/asn-buf.h \ cxx-lib/inc/snacc.h \ cxx-lib/inc/snaccdll.h \ cxx-lib/inc/snaccexcept.h \ - cxx-lib/inc/snaccrose.h \ cxx-lib/inc/tcl-if.h cxx_lib_libcxxasn1_la_SOURCES = \
When cross-compiling the system, it is not possible for the esnacc binary to correctly execute on the build system (well, not always). There is a method to combat this, which is the approach gcc takes, to generate a stage one version of the compiler and then execute that to generate the final tools (stage 2). That represents quite a bit of work. On the other hand, people are using esnacc *now* and as Marty Galligan reports at http://mail.esnacc.org/pipermail/dev/2017-July/000414.html there can be problems cross building, especially related to this. This commit detects that a cross-compilation is occuring and disables the generation of the asn.1 ROSE support. A future commit can enable this when it becomes necessary. Reported-by: Marty Galligan <m_galligan@comcast.net> Signed-off-by: Aaron Conole <aconole@bytheb.org> --- configure.ac | 4 ++++ cxx-lib/automake.mk | 9 ++++++--- 2 files changed, 10 insertions(+), 3 deletions(-)