Diego Biurrun <[email protected]> writes: > On Thu, May 19, 2011 at 06:45:11PM +0100, Måns Rullgård wrote: >> Diego Biurrun <[email protected]> writes: >> >> > On Thu, May 19, 2011 at 06:33:46PM +0100, Måns Rullgård wrote: >> >> Diego Biurrun <[email protected]> writes: >> >> >> >> > On Thu, May 19, 2011 at 05:39:00PM +0100, Måns Rullgård wrote: >> >> >> Diego Biurrun <[email protected]> writes: >> >> >> >> >> >> > --- a/libavcodec/Makefile >> >> >> > +++ b/libavcodec/Makefile >> >> >> > @@ -676,14 +676,10 @@ include $(SUBDIR)../subdir.mak >> >> >> > >> >> >> > -$(SUBDIR)cos_tables.c: $(SUBDIR)cos_tablegen$(HOSTEXESUF) >> >> >> > - $(M)./$< > $@ >> >> >> > - >> >> >> > -$(SUBDIR)cos_fixed_tables.c: $(SUBDIR)cos_tablegen$(HOSTEXESUF) >> >> >> > - $(M)./$< cos fixed > $@ >> >> >> > +$(SUBDIR)%_tables.c: $(SUBDIR)cos_tablegen$(HOSTEXESUF) >> >> >> > + $(M)./$< $* > $@ >> >> >> > >> >> >> > -$(SUBDIR)sin_tables.c: $(SUBDIR)cos_tablegen$(HOSTEXESUF) >> >> >> > - $(M)./$< sin > $@ >> >> >> > +.SECONDARY: $(SUBDIR)cos_tables.c $(SUBDIR)cos_fixed_tables.c >> >> >> > $(SUBDIR)sin_tables.c >> >> >> >> >> >> Using .SECONDARY is fragile and almost always wrong. In fact, I have >> >> >> never yet seen a case where using it did the right thing. Thus I am >> >> >> highly sceptical to this patch. >> >> > >> >> > It has always worked fine for me >> >> >> >> But you don't know why. Never do things you do not understand why they >> >> appear to work. >> > >> > I have read and understood the GNU Make manual, thanks for asking. >> >> It doesn't seem that way. > > Well, you are wrong - I read it cover to cover.
I'm not disputing that you read it. >> For example, you were unaware of the changes in GNU make 3.82. > > I read the manual for 3.81. 3.82 is recent and I had no time to study it > sufficiently. No distro I use regularly has picked it up yet, so I saw > no need for hurry. You know that GNU Make releases happen once or twice > per decade and take some time to trickle down into standard toolchains. It is showing up here and there already, and with releases being so infrequent, there is really no excuse for not reading the release notes when they do happen. >> >> > - what do you suggest instead .PRECIOUS? >> >> >> >> That one has the same issues. I've never managed to get it behaving >> >> consistently. >> > >> > So what were the problems you encountered? I never had issues. >> >> "Intermediate" files being randomly deleted, same files not always being >> rebuilt when they should etc, etc. It is much better to write makefiles >> such that no files are considered "intermediate" and subjected to >> bizarre rules. > > I'm not a fan of .SECONDARY or .PRECIOUS myself (there's no way to > specify patterns for them for example), but I found them to work > as advertised. The probably do work as advertised, the problem being that this is rarely what you really want. Both seem like dirty hacks added because someone was too lazy to do things properly. -- Måns Rullgård [email protected] _______________________________________________ libav-devel mailing list [email protected] https://lists.libav.org/mailman/listinfo/libav-devel
