Re: gEDA-user: Patch to PCB build system needs testing/feedback
On Sun, 2010-04-18 at 20:16 -0700, Jared Casper wrote: I just discovered that the latest automake (1.11) has a nifty feature to create silent build rules to produce a Linux kernel style build that just displays CC file.c etc. instead of the whole command line (must have missed the memo last year). The attached patch enables this mode in PCB and edits most of the custom build rules to use the new silent type of output. Right now it turns silent build on by default. The old style can be obtained using make V=1 or with the --disable-silent-rules configure option. To make it not on by default, remove the [yes] in the call to AM_SILENT_RULES in configure.ac. I tried such a patch for gEDA (it is in my repo actually). I'm of mixed feeling about it though.. it makes the normal compile output nicer, but it doesn't honour the -s option passed to make.. so I loose my usual easy way of checking for errors without getting lost in output (make -s) -- Peter Clifton Electrical Engineering Division, Engineering Department, University of Cambridge, 9, JJ Thomson Avenue, Cambridge CB3 0FA Tel: +44 (0)7729 980173 - (No signal in the lab!) Tel: +44 (0)1223 748328 - (Shared lab phone, ask for me) ___ geda-user mailing list geda-user@moria.seul.org http://www.seul.org/cgi-bin/mailman/listinfo/geda-user
Re: gEDA-user: Patch to PCB build system needs testing/feedback
Hi Jared, On Sun, 2010-04-18 at 16:05 -0700, Jared Casper wrote: Hi all, I just discovered that the latest automake (1.11) has a nifty feature to create silent build rules to produce a Linux kernel style build that just displays CC file.c etc. instead of the whole command line (must have missed the memo last year). The attached patch enables this mode in PCB and edits most of the custom build rules to use the new silent type of output. Right now it turns silent build on by default. The old style can be obtained using make V=1 or with the --disable-silent-rules configure option. To make it not on by default, remove the [yes] in the call to AM_SILENT_RULES in configure.ac. I think it makes the build much cleaner and readable overall and, more importantly, makes the errors and warnings much easier to see. I decided to send it here instead of the patch tracker for two reasons: 1) Feedback to see if people like this style of build output. 2) Testing. I don't have a box that has autoconf 2.60 and automake 1.11 (I either have servers that have been up for ever and still on autoconf 2.60 or desktops that are very up to date and have automake = 1.11.) I'd be surprised if it broke things with automake 1.11, but that needs testing. Also, I don't have a Windows box or a box with a non-GNU tool chain to test it out on. So if any body with these environments can test this out for me I'd appreciate it. Thanks! Jared Here is my EUR 0.02 on this one I have an old Fedora Core 5 box running (which is very slw ;) and it took some time to get some results. I have automake-1.9.6 and recently upgraded to autoconf-2.63 I applied your patch and ran the usual suspects ./autogen.sh and ./configure Running make with either v=0 v=1 V=0 or V=1 gives the same results, just no silent mode over here. The other (family) box has Microsoft Windows XP (SP3) with a recent cygwin (new release and build system). This one has automake-1.11.1 and autoconf-2.65 and AFAICT here things work as you advocate ;-) I had to configure with --disable-doc and --disable-update-desktop-database to get things working here. I hope this confirms what you said in the above message. Kind regards, Bert Timmerman. ___ geda-user mailing list geda-user@moria.seul.org http://www.seul.org/cgi-bin/mailman/listinfo/geda-user
Re: gEDA-user: Patch to PCB build system needs testing/feedback
Hi; On 4/18/10, Jared Casper jaredcas...@gmail.com wrote: I think it makes the build much cleaner and readable overall and, more importantly, makes the errors and warnings much easier to see. Actually, I don't think so; the build itself does not change, the readability arguably decreases (I'll need an extra step to switch on the full mode). I decided to send it here instead of the patch tracker for two reasons: 1) Feedback to see if people like this style of build output. 2) Testing. I don't have a box that has autoconf 2.60 and automake 1.11 (I either have servers that have been up for ever and still on autoconf 2.60 or desktops that are very up to date and have automake = 1.11.) I'd be surprised if it broke things with automake 1.11, but that needs testing. It works for me with autoconf-2.63, automake-1.10.2; that is, nothing changes. Some of patches from my collection will need to be rebased manually, but I have no other objections. Thanks, Ineiev ___ geda-user mailing list geda-user@moria.seul.org http://www.seul.org/cgi-bin/mailman/listinfo/geda-user
Re: gEDA-user: Patch to PCB build system needs testing/feedback
Jared Casper wrote: I think it makes the build much cleaner and readable overall and, more importantly, makes the errors and warnings much easier to see. Anything to help understandability of compile output is good by me. Your change probably won't decrease error output, just the ones that go smoothly are shortened, right? John ___ geda-user mailing list geda-user@moria.seul.org http://www.seul.org/cgi-bin/mailman/listinfo/geda-user
Re: gEDA-user: Patch to PCB build system needs testing/feedback
On Mon, Apr 19, 2010 at 7:57 AM, John Griessen j...@ecosensory.com wrote: Anything to help understandability of compile output is good by me. Your change probably won't decrease error output, just the ones that go smoothly are shortened, right? The errors/warnings themselves aren't shortened, it just shortens the command line used to compile. As an example this is the output of the build with this patch applied, versus without. (sorry for the spam of compile output, but since people without automake 1.11 won't be able to see the change, I think this will help demonstrate what this patch would do. In fact the mere fact I feel like I need to apologize for pasting compile output could be part of the argument :) With patch (warnings obvious) CC action.o CC autoplace.o CC autoroute.o ../../../pcb/pcb/src/autoroute.c:3431: warning: ‘add_clearance’ defined but not used CC buffer.o CC change.o ../../../pcb/pcb/src/change.c: In function ‘ChangeElementName’: ../../../pcb/pcb/src/change.c:1016: warning: unused variable ‘old’ CC clip.o CC command.o Without patch (very difficult to quickly see what exactly is going on, let alone catch the warnings): gcc -std=gnu99 -DLOCALEDIR=\/usr/local/share/locale\ -DHAVE_CONFIG_H -I. -I../../../pcb/pcb/src -I.. -I../../../pcb/pcb -I../../../pcb/pcb/src/icons -I../../../pcb/pcb/src/gts -I/usr/include -I/usr/include -I/usr/include -DPREFIXDIR=\/usr/local\ -DBINDIR=\/usr/local/bin\ -DHOST=\x86_64-unknown-linux-gnu\ -DPCBLIBDIR=\/usr/local/share/pcb\ -DPCBTREEDIR=\/usr/local/share/pcb/newlib\ -DPCBTREEPATH=\/usr/local/share/pcb/newlib:/usr/local/share/pcb/pcblib-newlib\ -DNDEBUG -g -O2 -I/usr/include -I/usr/include -I/usr/include -I/usr/include/glib-2.0 -I/usr/lib/glib-2.0/include -pthread -I/usr/include/gtk-2.0 -I/usr/lib/gtk-2.0/include -I/usr/include/atk-1.0 -I/usr/include/cairo -I/usr/include/pango-1.0 -I/usr/include/glib-2.0 -I/usr/lib/glib-2.0/include -I/usr/include/pixman-1 -I/usr/include/freetype2 -I/usr/include/libpng14-Wall -Wdeclaration-after-statement -MT action.o -MD -MP -MF $depbase.Tpo -c -o action.o ../../../pcb/pcb/src/action.c \ mv -f $depbase.Tpo $depbase.Po depbase=`echo autoplace.o | sed 's|[^/]*$|.deps/|;s|\.o$||'`;\ gcc -std=gnu99 -DLOCALEDIR=\/usr/local/share/locale\ -DHAVE_CONFIG_H -I. -I../../../pcb/pcb/src -I.. -I../../../pcb/pcb -I../../../pcb/pcb/src/icons -I../../../pcb/pcb/src/gts -I/usr/include -I/usr/include -I/usr/include -DPREFIXDIR=\/usr/local\ -DBINDIR=\/usr/local/bin\ -DHOST=\x86_64-unknown-linux-gnu\ -DPCBLIBDIR=\/usr/local/share/pcb\ -DPCBTREEDIR=\/usr/local/share/pcb/newlib\ -DPCBTREEPATH=\/usr/local/share/pcb/newlib:/usr/local/share/pcb/pcblib-newlib\ -DNDEBUG -g -O2 -I/usr/include -I/usr/include -I/usr/include -I/usr/include/glib-2.0 -I/usr/lib/glib-2.0/include -pthread -I/usr/include/gtk-2.0 -I/usr/lib/gtk-2.0/include -I/usr/include/atk-1.0 -I/usr/include/cairo -I/usr/include/pango-1.0 -I/usr/include/glib-2.0 -I/usr/lib/glib-2.0/include -I/usr/include/pixman-1 -I/usr/include/freetype2 -I/usr/include/libpng14-Wall -Wdeclaration-after-statement -MT autoplace.o -MD -MP -MF $depbase.Tpo -c -o autoplace.o ../../../pcb/pcb/src/autoplace.c \ mv -f $depbase.Tpo $depbase.Po depbase=`echo autoroute.o | sed 's|[^/]*$|.deps/|;s|\.o$||'`;\ gcc -std=gnu99 -DLOCALEDIR=\/usr/local/share/locale\ -DHAVE_CONFIG_H -I. -I../../../pcb/pcb/src -I.. -I../../../pcb/pcb -I../../../pcb/pcb/src/icons -I../../../pcb/pcb/src/gts -I/usr/include -I/usr/include -I/usr/include -DPREFIXDIR=\/usr/local\ -DBINDIR=\/usr/local/bin\ -DHOST=\x86_64-unknown-linux-gnu\ -DPCBLIBDIR=\/usr/local/share/pcb\ -DPCBTREEDIR=\/usr/local/share/pcb/newlib\ -DPCBTREEPATH=\/usr/local/share/pcb/newlib:/usr/local/share/pcb/pcblib-newlib\ -DNDEBUG -g -O2 -I/usr/include -I/usr/include -I/usr/include -I/usr/include/glib-2.0 -I/usr/lib/glib-2.0/include -pthread -I/usr/include/gtk-2.0 -I/usr/lib/gtk-2.0/include -I/usr/include/atk-1.0 -I/usr/include/cairo -I/usr/include/pango-1.0 -I/usr/include/glib-2.0 -I/usr/lib/glib-2.0/include -I/usr/include/pixman-1 -I/usr/include/freetype2 -I/usr/include/libpng14-Wall -Wdeclaration-after-statement -MT autoroute.o -MD -MP -MF $depbase.Tpo -c -o autoroute.o ../../../pcb/pcb/src/autoroute.c \ mv -f $depbase.Tpo $depbase.Po ../../../pcb/pcb/src/autoroute.c:3431: warning: ‘add_clearance’ defined but not used depbase=`echo buffer.o | sed 's|[^/]*$|.deps/|;s|\.o$||'`;\ gcc -std=gnu99 -DLOCALEDIR=\/usr/local/share/locale\ -DHAVE_CONFIG_H -I. -I../../../pcb/pcb/src -I.. -I../../../pcb/pcb -I../../../pcb/pcb/src/icons -I../../../pcb/pcb/src/gts -I/usr/include -I/usr/include -I/usr/include -DPREFIXDIR=\/usr/local\ -DBINDIR=\/usr/local/bin\ -DHOST=\x86_64-unknown-linux-gnu\ -DPCBLIBDIR=\/usr/local/share/pcb\ -DPCBTREEDIR=\/usr/local/share/pcb/newlib\ -DPCBTREEPATH=\/usr/local/share/pcb/newlib:/usr/local/share/pcb/pcblib-newlib\ -DNDEBUG -g -O2
Re: gEDA-user: Patch to PCB build system needs testing/feedback
On Mon, Apr 19, 2010 at 2:03 AM, Ineiev ine...@gmail.com wrote: On 4/18/10, Jared Casper jaredcas...@gmail.com wrote: I think it makes the build much cleaner and readable overall and, more importantly, makes the errors and warnings much easier to see. Actually, I don't think so; the build itself does not change, the readability arguably decreases (I'll need an extra step to switch on the full mode). I'm curious why you feel readability decreases? I still see what file is being compiled and the full output the errors. The only information I loose and need to switch on full mode for is the full compiler flags used in the compile, most of which probably aren't directly related to compiling this particular file and in my experience are very rarely needed to understand a compiler error or warning. Readability is increased because the warnings are not lost in the lines and lines of command line output (see my last reply) It works for me with autoconf-2.63, automake-1.10.2; that is, nothing changes. Thanks for testing, that is indeed what was expected. Some of patches from my collection will need to be rebased manually, but I have no other objections. I'm sure they will get in before this (if it does), in which case I'll rebase this stuff, no problem. Jared ___ geda-user mailing list geda-user@moria.seul.org http://www.seul.org/cgi-bin/mailman/listinfo/geda-user
Re: gEDA-user: Patch to PCB build system needs testing/feedback
Leaving out the commandline in a build can open all sorts of cans with worms. Most commandline flags control parts of header files (-DHAVE_STRANGE_MOFOS -I/other/weired/headers) and getting this wrong is a show-stopper with my bugs... I feel make V=0.5 would be ideal: omit the stuff if everything goes well and show the commandline (and the directory this make was started in) if an error occurs (not for warnings). Clearly I know this is wishful thinking. I wrote an extra tool errlog, to conserve the commandline in automated build logs for my projects. It's sort of a tee(1) that calls the compiler, sends it's output to log and stdout and terminates with the code the compiler terminated. It's trivial to invoke: instead of CC = gcc use CC = errlog gcc. Armin Jared Casper wrote: On Mon, Apr 19, 2010 at 2:03 AM, Ineiev ine...@gmail.com wrote: On 4/18/10, Jared Casper jaredcas...@gmail.com wrote: I think it makes the build much cleaner and readable overall and, more importantly, makes the errors and warnings much easier to see. Actually, I don't think so; the build itself does not change, the readability arguably decreases (I'll need an extra step to switch on the full mode). I'm curious why you feel readability decreases? I still see what file is being compiled and the full output the errors. The only information I loose and need to switch on full mode for is the full compiler flags used in the compile, most of which probably aren't directly related to compiling this particular file and in my experience are very rarely needed to understand a compiler error or warning. Readability is increased because the warnings are not lost in the lines and lines of command line output (see my last reply) It works for me with autoconf-2.63, automake-1.10.2; that is, nothing changes. Thanks for testing, that is indeed what was expected. Some of patches from my collection will need to be rebased manually, but I have no other objections. I'm sure they will get in before this (if it does), in which case I'll rebase this stuff, no problem. Jared ___ geda-user mailing list geda-user@moria.seul.org http://www.seul.org/cgi-bin/mailman/listinfo/geda-user ___ geda-user mailing list geda-user@moria.seul.org http://www.seul.org/cgi-bin/mailman/listinfo/geda-user
Re: gEDA-user: Patch to PCB build system needs testing/feedback
On Mon, Apr 19, 2010 at 5:27 PM, Armin Faltl armin.fa...@aon.at wrote: I feel make V=0.5 would be ideal: omit the stuff if everything goes well and show the commandline (and the directory this make was started in) if an error occurs (not for warnings). ... which you should get if you run make || make V=1. In all likelihood, the errors will still be there the second time. I dunno, this looks suspiciously similar to what the Linux kernel's kbuild system has been doing for a number of years now. I don't think it's too much to ask users to re-post build logs with V=1. -- - Charles Lepple ___ geda-user mailing list geda-user@moria.seul.org http://www.seul.org/cgi-bin/mailman/listinfo/geda-user
gEDA-user: Patch to PCB build system needs testing/feedback
Hi all, I just discovered that the latest automake (1.11) has a nifty feature to create silent build rules to produce a Linux kernel style build that just displays CC file.c etc. instead of the whole command line (must have missed the memo last year). The attached patch enables this mode in PCB and edits most of the custom build rules to use the new silent type of output. Right now it turns silent build on by default. The old style can be obtained using make V=1 or with the --disable-silent-rules configure option. To make it not on by default, remove the [yes] in the call to AM_SILENT_RULES in configure.ac. I think it makes the build much cleaner and readable overall and, more importantly, makes the errors and warnings much easier to see. I decided to send it here instead of the patch tracker for two reasons: 1) Feedback to see if people like this style of build output. 2) Testing. I don't have a box that has autoconf 2.60 and automake 1.11 (I either have servers that have been up for ever and still on autoconf 2.60 or desktops that are very up to date and have automake = 1.11.) I'd be surprised if it broke things with automake 1.11, but that needs testing. Also, I don't have a Windows box or a box with a non-GNU tool chain to test it out on. So if any body with these environments can test this out for me I'd appreciate it. Thanks! Jared ___ geda-user mailing list geda-user@moria.seul.org http://www.seul.org/cgi-bin/mailman/listinfo/geda-user
Re: gEDA-user: Patch to PCB build system needs testing/feedback
This time with the patch. On Sun, Apr 18, 2010 at 4:05 PM, Jared Casper jaredcas...@gmail.com wrote: Hi all, I just discovered that the latest automake (1.11) has a nifty feature to create silent build rules to produce a Linux kernel style build that just displays CC file.c etc. instead of the whole command line (must have missed the memo last year). The attached patch enables this mode in PCB and edits most of the custom build rules to use the new silent type of output. Right now it turns silent build on by default. The old style can be obtained using make V=1 or with the --disable-silent-rules configure option. To make it not on by default, remove the [yes] in the call to AM_SILENT_RULES in configure.ac. I think it makes the build much cleaner and readable overall and, more importantly, makes the errors and warnings much easier to see. I decided to send it here instead of the patch tracker for two reasons: 1) Feedback to see if people like this style of build output. 2) Testing. I don't have a box that has autoconf 2.60 and automake 1.11 (I either have servers that have been up for ever and still on autoconf 2.60 or desktops that are very up to date and have automake = 1.11.) I'd be surprised if it broke things with automake 1.11, but that needs testing. Also, I don't have a Windows box or a box with a non-GNU tool chain to test it out on. So if any body with these environments can test this out for me I'd appreciate it. Thanks! Jared ___ geda-user mailing list geda-user@moria.seul.org http://www.seul.org/cgi-bin/mailman/listinfo/geda-user 0001-Enable-linux-kernel-style-silent-build-if-available.patch Description: Binary data ___ geda-user mailing list geda-user@moria.seul.org http://www.seul.org/cgi-bin/mailman/listinfo/geda-user
Re: gEDA-user: Patch to PCB build system needs testing/feedback
Jared - Sent privately, maybe you can summarize responses (or if I'm the only one, tell me, and I'll post to the list). On Sun, Apr 18, 2010 at 04:05:16PM -0700, Jared Casper wrote: I think it makes the build much cleaner and readable overall and, more importantly, makes the errors and warnings much easier to see. I don't mind the short form, until I run into trouble. Then I want a documented way to turn it off, so I can see what make is actually trying to do. The explicit command lines are wonderful sources of cutting-and-pasting experiments on the command line. - Larry ___ geda-user mailing list geda-user@moria.seul.org http://www.seul.org/cgi-bin/mailman/listinfo/geda-user
Re: gEDA-user: Patch to PCB build system needs testing/feedback
On Sun, Apr 18, 2010 at 08:38:28PM -0700, Larry Doolittle wrote: Sent privately, maybe you can summarize responses Yeah, well, I got distracted at exactly the wrong moment. So kick me. - Larry ___ geda-user mailing list geda-user@moria.seul.org http://www.seul.org/cgi-bin/mailman/listinfo/geda-user
Re: gEDA-user: Patch to PCB build system needs testing/feedback
On Sun, Apr 18, 2010 at 8:38 PM, Larry Doolittle ldool...@recycle.lbl.gov wrote: I don't mind the short form, until I run into trouble. Then I want a documented way to turn it off, so I can see what make is actually trying to do. The explicit command lines are wonderful sources of cutting-and-pasting experiments on the command line. Absolutely. Like I said it is pretty easy to turn off. Using make V=1 instead of just make will display all the full command lines for a single run of make. ./configure'ing with --disable-silent-rules will turn off the silent builds completely and show all command lines in the build tree. Jared ___ geda-user mailing list geda-user@moria.seul.org http://www.seul.org/cgi-bin/mailman/listinfo/geda-user
