In preparation for the upcoming release, I thought I'd try `make distcheck`.
$ make -k distcheck make[1]: *** No rule to make target `make.dep', needed by `distdir'. make[1]: *** No rule to make target `make.files', needed by `distdir'. make[1]: *** No rule to make target `make.ofiles', needed by `distdir'. make[1]: Target `distdir' not remade because of errors. make: *** [distdir] Error 1 make: Target `distcheck' not remade because of errors.
I know the solution for this first problem is to run mkfiles.pl and mkdep.pl, but I wonder if those scripts should be thrown into the autogen.sh or configure mix somehow.
After I generate those files `make distcheck` produces the errors that can be found at: http://www.ekips.org/comp/inkscape/inx/distcheck/stderr.txt
Aaron Spike
On Fri, May 12, 2006 at 10:10:59AM -0500, Aaron Spike wrote:
In preparation for the upcoming release, I thought I'd try `make distcheck`.
$ make -k distcheck make[1]: *** No rule to make target `make.dep', needed by `distdir'. make[1]: *** No rule to make target `make.files', needed by `distdir'. make[1]: *** No rule to make target `make.ofiles', needed by `distdir'. make[1]: Target `distdir' not remade because of errors. make: *** [distdir] Error 1 make: Target `distcheck' not remade because of errors.
I know the solution for this first problem is to run mkfiles.pl and mkdep.pl, but I wonder if those scripts should be thrown into the autogen.sh or configure mix somehow.
After I generate those files `make distcheck` produces the errors that can be found at: http://www.ekips.org/comp/inkscape/inx/distcheck/stderr.txt
Are there any other distcheck issues remaining to be addressed?
Bryce
On Fri, May 12, 2006 at 10:10:59AM -0500, Aaron Spike wrote:
In preparation for the upcoming release, I thought I'd try `make distcheck`.
$ make -k distcheck make[1]: *** No rule to make target `make.dep', needed by `distdir'. make[1]: *** No rule to make target `make.files', needed by `distdir'. make[1]: *** No rule to make target `make.ofiles', needed by `distdir'. make[1]: Target `distdir' not remade because of errors. make: *** [distdir] Error 1 make: Target `distcheck' not remade because of errors.
I know the solution for this first problem is to run mkfiles.pl and mkdep.pl, but I wonder if those scripts should be thrown into the autogen.sh or configure mix somehow.
After I generate those files `make distcheck` produces the errors that can be found at: http://www.ekips.org/comp/inkscape/inx/distcheck/stderr.txt
Anyone have an idea on the above issues?
This is a critical step for being able to start doing pre-releases during the bugfix phase, so any help on getting make distcheck to pass will be very important.
Thanks, Bryce
On 22/05/06, Bryce Harrington <bryce@...961...> wrote:
On Fri, May 12, 2006 at 10:10:59AM -0500, Aaron Spike wrote:
In preparation for the upcoming release, I thought I'd try `make distcheck`.
[ snip ]
After I generate those files `make distcheck` produces the errors that can be found at: http://www.ekips.org/comp/inkscape/inx/distcheck/stderr.txt
Anyone have an idea on the above issues?
This is a critical step for being able to start doing pre-releases during the bugfix phase, ...
See http://bootstrapper.sourceforge.net/index.php?name=tutorial :
"... now for our last step, we'll check to see how our distribution works with 'make distcheck'. This does two things, it rolls a distribution tarball, and then checks it by unpacking it in a new directory, building it in another directory and installing it in a third directory."
make distcheck doesn't check Inkscape itself, it merely automates a series of steps for verifying that a candidate release tarball is (for instance) complete. The errors you are seeing are probably bogus in the sense that distcheck has to make up it idea of how to configure its untarr'd source tree from thin air, and it may not get it right for your system.
In may case, running: $ make -sC Panther distcheck goes through the steps above, and eventually reports failure as;
checking libintl.h usability... no checking libintl.h presence... no checking for libintl.h... no checking for pkg-config... /sw/bin/pkg-config checking for png_read_info in -lpng... no configure: error: libpng >= 1.2 is needed to compile inkscape
This doesn't mean that the tarball is faulty, but reflects the fact that configure cannot find my png libraries (which are in /sw/lib) without help.
Apologies if I am describing eggs to grandmothers here ...
Ben
Bryce Harrington wrote:
This is a critical step for being able to start doing pre-releases during the bugfix phase, so any help on getting make distcheck to pass will be very important.
Sorry I carelessly replaced the URL Bryce linked above. I guess that is ok, because my problem with distcheck changed. Now I getting the long strand of "What GTK?" type errors similar to those seen below. This error resulted from the changes Ralf made in revision 11886. It looks harmless enough to me, but now I can't even build autopackages.
http://svn.sourceforge.net/viewcvs.cgi/inkscape/inkscape/trunk/configure.ac?...
If we could get a couple people to run `make check` and `make distcheck` and start resolving or even just reporting the problems, that would be super.
Oh and has anybody else noticed this?
$ make clean $ svn up Restored 'po/inkscape.pot'
Shoule make clean remove a file stored in SVN? Rather should a generated file removed by make clean be stored in svn?
Aaron Spike
make[2]: Entering directory `/usr/local/src/inkscape/ap/src' if apg++ -DHAVE_CONFIG_H -I. -I. -I.. -I/usr/include/freetype2 -I/usr/local/src/inkscape/extra/static//include/loudmouth-1.0 -I/usr/include/glib-2.0 -I/usr/lib/glib-2.0/include -I/usr/include/freetype2 -DPOTRACE="potrace" -I/usr/local/src/inkscape/extra/static/include/sigc++-2.0 -I/usr/local/src/inkscape/extra/static/lib/sigc++-2.0/include -I../cxxtest -I/usr/local/src/inkscape/extra/static/include -Wall -W -Wpointer-arith -Wcast-align -Wsign-compare -Woverloaded-virtual -Wswitch -Wno-unused-parameter -O2 -march=i586 -pipe -MT arc-context.o -MD -MP -MF ".deps/arc-context.Tpo" \ -c -o arc-context.o `test -f 'arc-context.cpp' || echo './'`arc-context.cpp; \ then mv -f ".deps/arc-context.Tpo" ".deps/arc-context.Po"; \ else rm -f ".deps/arc-context.Tpo"; exit 1; \ fi arc-context.cpp:21:28: error: gdk/gdkkeysyms.h: No such file or directory arc-context.cpp:24:25: error: glibmm/i18n.h: No such file or directory In file included from arc-context.cpp:25: display/sp-canvas.h:31:27: error: gdk/gdkevents.h: No such file or directory display/sp-canvas.h:32:23: error: gdk/gdkgc.h: No such file or directory display/sp-canvas.h:33:27: error: gtk/gtkobject.h: No such file or directory display/sp-canvas.h:34:27: error: gtk/gtkwidget.h: No such file or directory In file included from arc-context.cpp:27: document.h:19:27: error: gtk/gtksignal.h: No such file or directory document.h:29:28: error: glibmm/ustring.h: No such file or directory In file included from arc-context.cpp:39: desktop.h:24:30: error: gtk/gtktypeutils.h: No such file or directory In file included from desktop.h:29, from arc-context.cpp:39: ui/view/view.h:17:26: error: gdk/gdktypes.h: No such file or directory display/sp-canvas.h:66: error: expected class-name before ‘{’ token display/sp-canvas.h:78: error: expected class-name before ‘{’ token display/sp-canvas.h:84: error: ‘GdkEvent’ has not been declared display/sp-canvas.h:87: error: ‘GtkType’ has not been declared display/sp-canvas.h:97: error: ‘GdkCursor’ has not been declared display/sp-canvas.h:116: error: ‘GtkWidget’ does not name a type display/sp-canvas.h:149: error: ‘GdkEvent’ does not name a type display/sp-canvas.h:154: error: ISO C++ forbids declaration of ‘GdkGC’ with no type display/sp-canvas.h:154: error: expected ‘;’ before ‘*’ token display/sp-canvas.h:178: error: expected constructor, destructor, or type conversion before ‘*’ token document.h:102: error: ‘Glib’ has not been declared document.h:102: error: ‘ustring’ does not name a type document.h:186: error: ‘GtkObject’ has not been declared desktop.h:172: error: ‘GtkType’ has not been declared desktop.h:173: error: ‘GtkType’ has not been declared context-fns.h:14: error: ‘GdkEvent’ has not been declared event-context.h:55: error: ISO C++ forbids declaration of ‘GdkCursor’ with no type event-context.h:55: error: expected ‘;’ before ‘*’ token event-context.h:88: error: ‘GdkEvent’ has not been declared event-context.h:89: error: ‘GdkEvent’ has not been declared event-context.h:102: error: ‘GdkEvent’ has not been declared event-context.h:103: error: ‘GdkEvent’ has not been declared event-context.h:105: error: ‘GdkEvent’ has not been declared event-context.h:112: error: ‘GdkEvent’ has not been declared event-context.h:115: error: ‘GdkEventKey’ was not declared in this scope event-context.h:115: error: ‘event’ was not declared in this scope arc-context.h:48: error: ‘GtkType’ does not name a type arc-context.cpp:50: error: ‘GdkEvent’ has not been declared arc-context.cpp:51: error: ‘GdkEvent’ has not been declared arc-context.cpp:58: error: ‘GtkType’ does not name a type arc-context.cpp: In function ‘void sp_arc_context_dispose(GObject*)’: arc-context.cpp:116: error: ‘sp_arc_context_get_type’ was not declared in this scope arc-context.cpp:116: error: expected primary-expression before ‘)’ token arc-context.cpp:116: error: ‘GTK_CHECK_CAST’ was not declared in this scope arc-context.cpp: In function ‘void sp_arc_context_selection_changed(Inkscape::Selection*, void*)’: arc-context.cpp:158: error: ‘sp_arc_context_get_type’ was not declared in this scope arc-context.cpp:158: error: expected primary-expression before ‘)’ token arc-context.cpp:158: error: ‘GTK_CHECK_CAST’ was not declared in this scope arc-context.cpp: In function ‘void sp_arc_context_setup(SPEventContext*)’: arc-context.cpp:186: error: ‘sp_arc_context_get_type’ was not declared in this scope arc-context.cpp:186: error: expected primary-expression before ‘)’ token arc-context.cpp:186: error: ‘GTK_CHECK_CAST’ was not declared in this scope arc-context.cpp: At global scope: arc-context.cpp:221: error: ‘GdkEvent’ has not been declared arc-context.cpp: In function ‘gint sp_arc_context_item_handler(SPEventContext*, SPItem*, int*)’: arc-context.cpp:226: error: request for member ‘type’ in ‘* event’, which is of non-class type ‘int’ arc-context.cpp:227: error: ‘GDK_BUTTON_PRESS’ was not declared in this scope arc-context.cpp:228: error: request for member ‘button’ in ‘* event’, which is of non-class type ‘int’ arc-context.cpp: At global scope: arc-context.cpp:245: error: ‘GdkEvent’ has not been declared arc-context.cpp: In function ‘gint sp_arc_context_root_handler(SPEventContext*, int*)’: arc-context.cpp:251: error: ‘sp_arc_context_get_type’ was not declared in this scope arc-context.cpp:251: error: expected primary-expression before ‘)’ token arc-context.cpp:251: error: ‘GTK_CHECK_CAST’ was not declared in this scope arc-context.cpp:257: error: request for member ‘type’ in ‘* event’, which is of non-class type ‘int’ arc-context.cpp:258: error: ‘GDK_BUTTON_PRESS’ was not declared in this scope arc-context.cpp:259: error: request for member ‘button’ in ‘* event’, which is of non-class type ‘int’ arc-context.cpp:266: error: expected primary-expression before ‘)’ token arc-context.cpp:267: error: ‘GDK_KEY_PRESS_MASK’ was not declared in this scope arc-context.cpp:267: error: ‘GDK_BUTTON_RELEASE_MASK’ was not declared in this scope arc-context.cpp:268: error: ‘GDK_POINTER_MOTION_MASK’ was not declared in this scope arc-context.cpp:268: error: ‘GDK_BUTTON_PRESS_MASK’ was not declared in this scope arc-context.cpp:269: error: request for member ‘button’ in ‘* event’, which is of non-class type ‘int’ arc-context.cpp:273: error: ‘GDK_MOTION_NOTIFY’ was not declared in this scope arc-context.cpp:274: error: request for member ‘motion’ in ‘* event’, which is of non-class type ‘int’ arc-context.cpp:274: error: ‘GDK_BUTTON1_MASK’ was not declared in this scope arc-context.cpp:277: error: request for member ‘motion’ in ‘* event’, which is of non-class type ‘int’ arc-context.cpp:278: error: request for member ‘motion’ in ‘* event’, which is of non-class type ‘int’ arc-context.cpp:286: error: request for member ‘motion’ in ‘* event’, which is of non-class type ‘int’ arc-context.cpp:286: error: request for member ‘motion’ in ‘* event’, which is of non-class type ‘int’ arc-context.cpp:288: error: request for member ‘motion’ in ‘* event’, which is of non-class type ‘int’ arc-context.cpp:292: error: ‘GDK_BUTTON_RELEASE’ was not declared in this scope arc-context.cpp:294: error: request for member ‘button’ in ‘* event’, which is of non-class type ‘int’ arc-context.cpp:301: error: request for member ‘button’ in ‘* event’, which is of non-class type ‘int’ arc-context.cpp:301: error: ‘GDK_SHIFT_MASK’ was not declared in this scope arc-context.cpp:315: error: expected primary-expression before ‘)’ token arc-context.cpp:315: error: request for member ‘button’ in ‘* event’, which is of non-class type ‘int’ arc-context.cpp:317: error: ‘GDK_KEY_PRESS’ was not declared in this scope arc-context.cpp:318: error: request for member ‘key’ in ‘* event’, which is of non-class type ‘int’ arc-context.cpp:318: error: ‘get_group0_keyval’ cannot be used as a function arc-context.cpp:319: error: ‘GDK_Alt_L’ was not declared in this scope arc-context.cpp:320: error: ‘GDK_Alt_R’ was not declared in this scope arc-context.cpp:321: error: ‘GDK_Control_L’ was not declared in this scope arc-context.cpp:322: error: ‘GDK_Control_R’ was not declared in this scope arc-context.cpp:323: error: ‘GDK_Shift_L’ was not declared in this scope arc-context.cpp:324: error: ‘GDK_Shift_R’ was not declared in this scope arc-context.cpp:325: error: ‘GDK_Meta_L’ was not declared in this scope arc-context.cpp:326: error: ‘GDK_Meta_R’ was not declared in this scope arc-context.cpp:328: error: ‘_’ was not declared in this scope arc-context.cpp:332: error: ‘GDK_Up’ was not declared in this scope arc-context.cpp:333: error: ‘GDK_Down’ was not declared in this scope arc-context.cpp:334: error: ‘GDK_KP_Up’ was not declared in this scope arc-context.cpp:335: error: ‘GDK_KP_Down’ was not declared in this scope arc-context.cpp:337: error: request for member ‘key’ in ‘* event’, which is of non-class type ‘int’ arc-context.cpp:337: error: ‘GDK_SHIFT_MASK’ was not declared in this scope arc-context.cpp:337: error: request for member ‘key’ in ‘* event’, which is of non-class type ‘int’ arc-context.cpp:337: error: ‘GDK_CONTROL_MASK’ was not declared in this scope arc-context.cpp:337: error: request for member ‘key’ in ‘* event’, which is of non-class type ‘int’ arc-context.cpp:337: error: ‘GDK_MOD1_MASK’ was not declared in this scope arc-context.cpp:340: error: ‘GDK_x’ was not declared in this scope arc-context.cpp:341: error: ‘GDK_X’ was not declared in this scope arc-context.cpp:342: error: request for member ‘key’ in ‘* event’, which is of non-class type ‘int’ arc-context.cpp:342: error: ‘GDK_SHIFT_MASK’ was not declared in this scope arc-context.cpp:342: error: request for member ‘key’ in ‘* event’, which is of non-class type ‘int’ arc-context.cpp:342: error: ‘GDK_CONTROL_MASK’ was not declared in this scope arc-context.cpp:342: error: request for member ‘key’ in ‘* event’, which is of non-class type ‘int’ arc-context.cpp:342: error: ‘GDK_MOD1_MASK’ was not declared in this scope arc-context.cpp:347: error: ‘GDK_Escape’ was not declared in this scope arc-context.cpp:354: error: ‘GDK_KEY_RELEASE’ was not declared in this scope arc-context.cpp:355: error: request for member ‘key’ in ‘* event’, which is of non-class type ‘int’ arc-context.cpp:356: error: ‘GDK_Alt_L’ was not declared in this scope arc-context.cpp:357: error: ‘GDK_Alt_R’ was not declared in this scope arc-context.cpp:358: error: ‘GDK_Control_L’ was not declared in this scope arc-context.cpp:359: error: ‘GDK_Control_R’ was not declared in this scope arc-context.cpp:360: error: ‘GDK_Shift_L’ was not declared in this scope arc-context.cpp:361: error: ‘GDK_Shift_R’ was not declared in this scope arc-context.cpp:362: error: ‘GDK_Meta_L’ was not declared in this scope arc-context.cpp:363: error: ‘GDK_Meta_R’ was not declared in this scope arc-context.cpp: In function ‘void sp_arc_drag(SPArcContext*, NR::Point, guint)’: arc-context.cpp:414: error: ‘_’ was not declared in this scope arc-context.cpp: At global scope: arc-context.cpp:77: warning: ‘void sp_arc_context_class_init(SPArcContextClass*)’ defined but not used arc-context.cpp:92: warning: ‘void sp_arc_context_init(SPArcContext*)’ defined but not used make[2]: *** [arc-context.o] Error 1 make[2]: Leaving directory `/usr/local/src/inkscape/ap/src' make[1]: *** [all-recursive] Error 1 make[1]: Leaving directory `/usr/local/src/inkscape/ap' make: *** [all] Error 2
Aaron Spike wrote:
This error resulted from the changes Ralf made in revision 11886. It looks harmless enough to me, but now I can't even build autopackages.
http://svn.sourceforge.net/viewcvs.cgi/inkscape/inkscape/trunk/configure.ac?...
So here's one problem with online code viewers. They compress adjacent spaces. I'll have to check and see if the missing space is causing me problems.
Aaron Spike
On 22/05/06, Aaron Spike <aaron@...749...> wrote:
Bryce Harrington wrote:
This is a critical step for being able to start doing pre-releases during the bugfix phase, so any help on getting make distcheck to pass will be very important.
Sorry I carelessly replaced the URL Bryce linked above. I guess that is ok, because my problem with distcheck changed. Now I getting the long strand of "What GTK?" type errors similar to those seen below. This error resulted from the changes Ralf made in revision 11886. It looks harmless enough to me, but now I can't even build autopackages.
http://svn.sourceforge.net/viewcvs.cgi/inkscape/inkscape/trunk/configure.ac?...
This is a good point. When I saw that a change to configure.ac had been made, I ran ./autogen.sh to create an up-to-date configure.in (this advice may need to go into the CompilingMacOSX page); I am fairly sure that make's dependency tracking does not check whether configure.in is out of date and would not know how to re-create it anyway. You may want to check the autoconf docs about this.
(and I repeat my apology concerning grandmaternal ovafrages).
I think that the autoconf folk assume that developers using configure.ac can work this out for themselves, which is not altogether unreasonable, but fails to take into account the fact that most of us like to treat it as a black box and that it may not be true that everyone working from checked out sources is an expert. Perhaps this is the main mission of 'make distcheck' - to check that the distribution tarball is complete.
If we could get a couple people to run `make check` and `make distcheck` and start resolving or even just reporting the problems, that would be super.
Oh and has anybody else noticed this?
$ make clean $ svn up Restored 'po/inkscape.pot'
Should make clean remove a file stored in SVN? Rather should a generated file removed by make clean be stored in svn?
It is not my pigeon, but I suspect that your second suggestion applies. There is nothing really wrong with committing generated files (though some people frown on it), but one should be selective. For example, if one wishes to encourage 'occasional' developers to compile from svn then the ./configure script should be in svn even though regular developers will use autoconf to generate it afresh.
(I don't do 'make clean' all that much, either!).
Aaron Spike
make[2]: Entering directory `/usr/local/src/inkscape/ap/src' if apg++ -DHAVE_CONFIG_H -I. -I. -I.. -I/usr/include/freetype2 -I/usr/local/src/inkscape/extra/static//include/loudmouth-1.0 -I/usr/include/glib-2.0 -I/usr/lib/glib-2.0/include -I/usr/include/freetype2 -DPOTRACE="potrace" -I/usr/local/src/inkscape/extra/static/include/sigc++-2.0 -I/usr/local/src/inkscape/extra/static/lib/sigc++-2.0/include -I../cxxtest -I/usr/local/src/inkscape/extra/static/include -Wall -W -Wpointer-arith -Wcast-align -Wsign-compare -Woverloaded-virtual -Wswitch -Wno-unused-parameter -O2 -march=i586 -pipe -MT arc-context.o -MD -MP -MF ".deps/arc-context.Tpo" \ -c -o arc-context.o `test -f 'arc-context.cpp' || echo './'`arc-context.cpp; \ then mv -f ".deps/arc-context.Tpo" ".deps/arc-context.Po"; \ else rm -f ".deps/arc-context.Tpo"; exit 1; \ fi arc-context.cpp:21:28: error: gdk/gdkkeysyms.h: No such file or directory
I think that this means that make cannot find your gtk include directory. Looking at the above, I would guess that you have a directory /usr/include/gtk-2.0 (mine is /sw/include/gtk-2.0/) and that there are a pair of -I flags missing, cognate to the ones for glib-2.0. I would hazard a guess that either ./configure did not run properly. or more reasonably, that it was missing something that should enable it to find your required gtk.
You may want to track down why (if this is the case) ./configure works correctly when called by hand, but not when called by 'make distcheck'. (If you are using the -k flag you may have told 'make distcheck' to carry on regardless after an unsuccessful ./configure; in which case you may have to scroll back a bit to find any errors reported by ./configure). You may find that ./configure left a 'config.log' which will almost certainly be helpful here.
I may be getting into tiger country, but I would be tempted to review the pkg-config installation, perhaps with a command like:
$ pkg-config --cflags gtk
You get the idea.
Please post again if you think I can help, but I don't claim unusual expertise!
Ben
Ben Fowler wrote:
On 22/05/06, Aaron Spike <aaron@...749...> wrote:
Bryce Harrington wrote:
This is a critical step for being able to start doing pre-releases during the bugfix phase, so any help on getting make distcheck to pass will be very important.
Sorry I carelessly replaced the URL Bryce linked above. I guess that is ok, because my problem with distcheck changed. Now I getting the long strand of "What GTK?" type errors similar to those seen below. This error resulted from the changes Ralf made in revision 11886. It looks harmless enough to me, but now I can't even build autopackages.
http://svn.sourceforge.net/viewcvs.cgi/inkscape/inkscape/trunk/configure.ac?...
This is a good point. When I saw that a change to configure.ac had been made, I ran ./autogen.sh to create an up-to-date configure.in
Yeah, I go a bit overboard with my autopackage making script. I run autogen.sh and configure and make clean every time.
Aaron Spike
On 22/05/06, Aaron Spike <aaron@...749...> wrote:
Ben Fowler wrote:
On 22/05/06, Aaron Spike <aaron@...749...> wrote:
Bryce Harrington wrote:
This is a critical step for being able to start doing pre-releases during the bugfix phase, so any help on getting make distcheck to pass will be very important.
[ snip ] This is a good point. When I saw that a change to configure.ac had been made, I ran ./autogen.sh to create an up-to-date configure.in
Yeah, I go a bit overboard with my autopackage making script. I run autogen.sh and configure and make clean every time.
If it helps, I think that for making a 'deliverable' or for packaging that is a very good idea, verging on mandatory, because you are removing one variable (the dependency tracking software) and potentially making life simpler for anybody downstream who wishes to build from sources by starting from a known 'zero' state.
Personally, I would use this as a stepping stone to a 'one step build' - a build operation that was fully automated, in which case that fact that parts were strictly unnecessary would not matter, as it would be being done without human attention.
Ben
Aaron Spike wrote:
Ben Fowler wrote:
On 22/05/06, Aaron Spike <aaron@...749...> wrote:
Bryce Harrington wrote:
This is a critical step for being able to start doing pre-releases during the bugfix phase, so any help on getting make distcheck to pass will be very important.
Sorry I carelessly replaced the URL Bryce linked above. I guess that is ok, because my problem with distcheck changed. Now I getting the long strand of "What GTK?" type errors similar to those seen below. This error resulted from the changes Ralf made in revision 11886. It looks harmless enough to me, but now I can't even build autopackages.
http://svn.sourceforge.net/viewcvs.cgi/inkscape/inkscape/trunk/configure.ac?...
Ok, this must be some issue with the way these macros work. The patch below works for me. Can anyone explain why?
Aaron Spike
Index: configure.ac =================================================================== --- configure.ac (revision 11886) +++ configure.ac (working copy) @@ -495,10 +495,10 @@ dnl *** NOTE: when we move to gtk 2.6 or later, we can remove the dnl ********* the override for g_ascii_strtod below... dnl sigc++-2.0 >= 2.0.12: using "visit_each" not available in 2.0.10 -PKG_CHECK_MODULES(INKSCAPE, gdkmm-2.4 glibmm-2.4 gtkmm-2.4 gtk+-2.0
= 2.4.0 libxml-2.0 >= 2.6.11 libxslt >= 1.0.15 sigc++-2.0 >= 2.0.11
$ink_spell_pkg gthread-2.0 >= 2.0) - if test $cc_vers_major -gt 3; then - PKG_CHECK_MODULES(INKSCAPE, sigc++-2.0 >= 2.0.12) + PKG_CHECK_MODULES(INKSCAPE, gdkmm-2.4 glibmm-2.4 gtkmm-2.4 gtk+-2.0 >= 2.4.0 libxml-2.0 >= 2.6.11 libxslt >= 1.0.15 sigc++-2.0
= 2.0.12 $ink_spell_pkg gthread-2.0 >= 2.0)
+else + PKG_CHECK_MODULES(INKSCAPE, gdkmm-2.4 glibmm-2.4 gtkmm-2.4 gtk+-2.0 >= 2.4.0 libxml-2.0 >= 2.6.11 libxslt >= 1.0.15 sigc++-2.0
= 2.0.11 $ink_spell_pkg gthread-2.0 >= 2.0)
fi
dnl Shouldn't we test for libpng and libz?
On 22/05/06, Aaron Spike <aaron@...749...> wrote:
Aaron Spike wrote:
Ben Fowler wrote:
On 22/05/06, Aaron Spike <aaron@...749...> wrote:
Bryce Harrington wrote:
http://svn.sourceforge.net/viewcvs.cgi/inkscape/inkscape/trunk/configure.ac?...
Ok, this must be some issue with the way these macros work. The patch below works for me. Can anyone explain why?
Aaron Spike
Index: configure.ac
--- configure.ac (revision 11886) +++ configure.ac (working copy) @@ -495,10 +495,10 @@ dnl *** NOTE: when we move to gtk 2.6 or later, we can remove the dnl ********* the override for g_ascii_strtod below... dnl sigc++-2.0 >= 2.0.12: using "visit_each" not available in 2.0.10 -PKG_CHECK_MODULES(INKSCAPE, gdkmm-2.4 glibmm-2.4 gtkmm-2.4 gtk+-2.0
= 2.4.0 libxml-2.0 >= 2.6.11 libxslt >= 1.0.15 sigc++-2.0 >= 2.0.11
$ink_spell_pkg gthread-2.0 >= 2.0)
- if test $cc_vers_major -gt 3; then
- PKG_CHECK_MODULES(INKSCAPE, sigc++-2.0 >= 2.0.12)
- PKG_CHECK_MODULES(INKSCAPE, gdkmm-2.4 glibmm-2.4 gtkmm-2.4 gtk+-2.0 >= 2.4.0 libxml-2.0 >= 2.6.11 libxslt >= 1.0.15 sigc++-2.0 = 2.0.12 $ink_spell_pkg gthread-2.0 >= 2.0)
+else
PKG_CHECK_MODULES(INKSCAPE, gdkmm-2.4 glibmm-2.4 gtkmm-2.4 gtk+-2.0 >= 2.4.0 libxml-2.0 >= 2.6.11 libxslt >= 1.0.15 sigc++-2.0 = 2.0.11 $ink_spell_pkg gthread-2.0 >= 2.0) fi
dnl Shouldn't we test for libpng and libz?
I am getting a little impression that I am answering questions that you haven't asked - do warn me if I am not helping.
Could you confirm which way the '-' and '+' markers are pointing. Looking at revision 11951, I am working on the basis that you run two calls to PKG_CHECK_MODULES (instead of picking one).
Does the second one over-write the results of the first one?
Does your ./configure really do what you want?
Ben
participants (3)
-
Aaron Spike -
Ben Fowler -
Bryce Harrington