Hello,
I'm a little confused about how chained pattern rules work, I though I
understood it, but now I'm getting cases where make cannot figure out
how to make my files.
For example, I have a Makefile containing:
%: %.gz; gzip -cd $< > $@
%: %.tar; tar -xf $<
If I have the file ``bash-4.1.tar.g
On Mon, 2010-04-26 at 20:02 -0700, Philip Guenther wrote:
> You're hitting the special handling of rules whose target is '%'
> that's described in the info pages in "10.5.5 Match-Anything Pattern
> Rules". If you stare at that doc a bit as well as the output from
> running make with the -d option,
> Hi,
> is that possible that we can add an option to "make" that we can
> generatge dependency graph? This will help us to debug issues (I know
> we already have --debug and --print-data-base option)
>
>
> Thanks
> Lynn
Since the dependency graph is more complex than just a tree (lines mus
ependency graph
>
> On 8/19/10, Luke Shumaker wrote:
> >> Hi,
> >> is that possible that we can add an option to "make" that we can
> >> generatge dependency graph? This will help us to debug issues (I know
> >> we already have --debug and --p
Oh, about the graphs: dotted line means it's an order-only dependency,
not an actual dependency, and a blue bubble indicates a PHONY target.
~ LukeShu
On Thu, 2010-08-19 at 23:45 -0400, Luke Shumaker wrote:
> Thank you for showing me graphviz, it's pretty neat!
>
> The atta
> > > > The attached patch adds the ``-g, --graph'' options to make; it will
> > > > spit out a graphviz file to stdout. (The patch is against the 3.82
> > > > release)
> > >
> > > I test the patch,it's great although it has bug.
> >
> > What bug?
>
> I compile cgit project with your patch
>
>
On Tue, 2010-08-24 at 13:32 -0400, Aleksey Yakovlev wrote:
> What is the file timestamps comparison code, used by the make? (in C/C
> ++ I guess)
> Is there any way to redefine it?
>
> Thanks
> Alex
It's in C.
Download the latest release's source code from
ftp://ftp.gnu.org/gnu/make/
or grab the
On Sat, 2010-08-28 at 09:02 +0430, ali hagigat wrote:
> 3.9 How make Reads a Makefile
> GNU make does its work in two distinct phases. During the first phase
> it reads all the make-
> files, included makefiles, etc. and internalizes all the variables and
> their values, implicit and
>
On Mon, 2010-09-06 at 10:20 +0430, ali hagigat wrote:
> print: *.c
> @echo "example"
>
> Considering the above rule in a makefile, what strings will be
> replaced with *.c in the first phase and the second phase of make
> process?
This is equivalent to
print: $(wildcard *.c)
@echo "
$$(addsuffix /%.c,foo bar) foo.h
> @echo '$+'
>
> and the result is:
> make: Nothing to be done for `/tmp/foo.o'.
>
> it did not expand to what you said!
>
> On Mon, Sep 6, 2010 at 8:53 PM, Luke Shumaker wrote:
> > On Mon, 2010-09-06 at 08:21
On Tue, 2010-09-07 at 15:40 -0400, Polder, Matthew J wrote:
> ...
> $(OBJS): $(patsubst %.o,%.cpp,$(notdir $@))
> $(CC) -c $^ -o $@
> $(CC) -c $(SRC_DIR)/$(*F).cpp -o $@
There are a few problems here:
First
$@ is not yet set when the dependencies are being evaluated. To use $@
in the de
On Thu, 2010-09-16 at 15:10 -0600, Daryl Lee wrote:
> So I added "export OSTYPE=$OSTYPE to my .bashrc and that cleaned
> everything up nicely. THANKS!!
That only needs to be "export OSTYPE"; without specifying a value bash
exports the current value.
~ LukeShu
_
On Sat, 2010-09-18 at 21:17 -0300, Sergio Belkin wrote:
> Hi folks,
>
> I have a Makefile that even when issuing for the first time the PHONY
> target clean wants to recompile as you can see below:
...
> What am I doing wrong?
>
> You can read the Makefile on http://pastebin.com/8zKGBQBt
>
> Tha
Forwarded Message
From: Sergio Belkin
> 2010/9/18 Luke Shumaker :
> I'm getting
> > Unknown paste ID, it may have expired or been deleted!
> from pastebin.
>
> ~ LukeShu
Oops... strike two. Here we go :)
http://
Don't forget to CC the mailing list!
~ LukeShu
On Sun, 2010-09-19 at 19:17 -0300, Sergio Belkin wrote:
> 2010/9/19 Luke Shumaker :
> > On Sun, 2010-09-19 at 00:51 -0300, Sergio Belkin wrote:
> >> 2010/9/18 Luke Shumaker :
> >> > On Sat, 2010-09-18 at 21:17 -03
Don't do that. Try this:
ifneq ($(filter avrlib avrtest,$(MAKECMDGOALS)),)
PLATFORM=avr
endif
ifneq ($(filter dylib,$(MAKECMDGOALS)),)
PLATFORM=osx
endif
### end ###
This does make the assumption that the only way to make the dylib,
avrlib, and avrtest targets is to specify them on the command
,
> Matt van de Werken.
> ____
> From: Luke Shumaker [luke...@sbcglobal.net]
> Sent: Saturday, October 02, 2010 8:13 AM
> To: Van De Werken, Matt (CESRE, Pullenvale)
> Cc: Help-make@gnu.org
> Subject: RE: variables-as-targets driving me insane
>
> Don&
On Thu, 2010-10-07 at 17:10 -0400, David Boyce wrote:
> On Thu, Oct 7, 2010 at 4:44 PM, David Boyce wrote:
> > I like to use full paths everywhere, using a
> > BASE_DIR variable which can be inherited from an env var or derived
> > from $(MAKEFILES), but that's just me.
>
> Sorry, I meant MAKEFIL
On Fri, 2010-10-22 at 00:16 +0200, Vincent De Groote wrote:
> Hello,
>
> I have a problem with a makefile (sic !). gmake is executed in a build
> directory, with several subdirectories.
> I want to build a library from several sources, and then build an executable
> from a test file and that libr
I would suggest trying UnionFS (supported either natively, or via FUSE
on most *NIXen). What you do is mount /release/src under /local/src,
with /local/src/ on top and /release/src on bottom. UnionFS first
checks if a file exists on at the top layer, and keeps checking the next
layer until it fin
n all files listed in the
file `md5sums' and make each of those depend on `md5sums'. (The sed
script breaks on filenames with spaces... but so does make, so it's OK)
I hope this helps.
~ Luke Shumaker
On Thu, 2010-10-28 at 17:37 +0200, erik.r...@--.-- wrote:
> Hi all,
On Wed, 2010-11-10 at 14:55 +0100, Warlich, Christof wrote:
> Hi,
>
> consider this makefile:
>
> .PHONY: all
> all: sorted.tgt header.tgt reverse.tgt footer.tgt
> # The lines below are supposed to do this:
> #sorted.tgt: list1.lst list2.lst
> # cat $^ | sort >$@
> #header.tgt: list1.lst list2.ls
On Sat, 2010-11-13 at 23:58 +0100, Erik Rull wrote:
> Well yes - I could write a rule like
> copy_object: object
> []
>
> But this would mean that object gets executed also if I did not give it as
> an argument to the make command.
>
> I don't need a general handling of the sequence of
On Sun, 2010-12-26 at 08:17 +0330, ali hagigat wrote:
> 5.3 Recipe Execution
> (In practice, make may take shortcuts that do not affect the results.)
> -
> What does it mean?
It means that GNU Make may
On Sun, 2010-12-26 at 08:37 +0330, ali hagigat wrote:
> makefile21:
>
> .ONESHELL:
> all:
> y1=bbeeggiinn;echo $$y1
> echo $$y1
> echo eenndd
>
> /root/makefiles_examples> make -s -f makefile21
> bbeeggiinn
>
> eenndd
>
--
On Wed, 2010-12-29 at 10:12 +0330, ali hagigat wrote:
> Would you please explain the use of +/- as a prefix character? How
> they are used and what make will do based on these characters?
> Why the manual does not explain them clearly and separately?
`-' is explained in section 5.5: Errors in Reci
On Wed, 2010-12-29 at 13:32 +0100, Warlich, Christof wrote:
> Hi,
>
> I need to ensure that a certain tool is generated before any reciepes are
> executed for a certain extension. To better understand what I need, consider
> the following makefile:
>
> xxx.ext:
> ./tool
>
> xxx.ext: tool #
On Thu, 2010-12-30 at 13:17 +0100, Warlich, Christof wrote:
> Hi Luke,
>
> Thanks for your ideas. You are right, pattern rules is what I need to look at.
>
> Cheers,
>
> Christof
No problem! Don't forget to CC: the mailing list, reply-to-all usually
does this.
--
~ LukeShu
http://lukeshu.ath
On Mon, 2011-01-17 at 10:28 -0800, givemecode wrote:
> I saw this question was posted but never answered on this forum. Hopefully
> this can be done:
>
> I have code in a makefile that creates a new file called "allheaders.h"
> which includes ALL the headers, and is intended to verify they compile
On Tue, 2011-01-18 at 07:51 +0330, ali hagigat wrote:
> The ‘-l’ option with no following number removes the load limit, if
> one was given with a previous ‘-l’ option.
> -
> How -l option without number removes the limit? I us
On Tue, 2011-01-18 at 09:13 +0330, ali hagigat wrote:
> it does not make sense we specify -l 2.5 and then use -l on command
> line, if we write:
> make -l 2.5
> will sub-makes inherit "-l 2.5" from their parent?
It does seem a little silly, but there are cases where it is useful. If
you don't nee
On Thu, 2011-01-27 at 17:28 +0100, Tobias Burnus wrote:
> Hello,
>
> I have a Makefile-dependency question regarding Fortran: In Fortran one
> can have so-called modules. If one compiles such a file, a .mod file is
> generated which contains the public interface and a .o file which
> contains t
Like everyone else, I really don't know where your problem is coming
from. However, if I had to guess, I'd say that one shell didn't close
the file in time for the next to open it, and one of them lost out. I'd
try doing something like
$(foreach item, $(very_long_list), $(shell echo '$(item)'>>
The trick is to get the different commands on different lines. Try
this:
define _ExecuteShcmdArray
$(foreach c,$1,@$($c)
)
endef
shellcmd1 = cd .. ; pwd
shellcmd2 = cd .. ; pwd
all:
$(call _ExecuteShcmdArray,shellcmd1 shellcmd2)
--
~ LukeShu
http://lukeshu.ath.cx
On Sun, 2011-02-27
gt; On Sat, 26 Feb 2011 21:23:15 -0500, Luke Shumaker
> > wrote:
> >> The trick is to get the different commands on different lines. Try
> >> this:
> >>
> >> define _ExecuteShcmdArray
> >> $(foreach c,$1,@$($c)
> >> )
> >> ende
I'm willing to write that off as a fault belonging to his email editor;
otherwise I would expect make to bork about a syntax error.
The problem is that recursive make doesn't track dependencies between
instances of make. However, it is hard to diagnose, because you have
fabricated the error messa
This is an ElectricMake issue, and this is a GNU Make mailing list, so
we may not be able to help you much.
However, it sounds like ElectricMake is setting some environmental
variables that it passes to GNU Make. For example if the GNU Makefile
sets `CFLAGS+=-g' then when you run it from the comm
I've always used the example from the GNU Make manual:
%.d: %.c
@set -e; rm -f $@; \
$(CC) -M $(CPPFLAGS) $< > $@.; \
sed 's,\($*\)\.o[ :]*,\1.o $@ : ,g' < $@. > $@; \
rm -f $@.
sources = foo.c bar.c
include $(sources:.c=.d)
--
~ LukeShu
http
ou linked, if {a,b,c,d,e} and at
least one of {x,y,z} have a common stem, then pattern rules can be
used with GNU Make (this syntax is non-portable). However, if the
filenames have nothing to do with eachother, then GNU Make has no way
of expressing that.
--
Happy hacking,
~ Luke Shumaker
__
Paul.
Without looking at the source, this feels a lot like the
.DELETE_ON_ERROR: bug that was in either 3.81 or 3.82 (I can't
recall which).
--
Happy hacking,
~ Luke Shumaker
___
Help-make mailing list
Help-make@gnu.org
https://lists.gnu.org/mailman/listinfo/help-make
erstand it :) I just wanted there to be one fewer
thing for someone looking at it to have to think about.
--
Happy hacking,
~ Luke Shumaker
___
Help-make mailing list
Help-make@gnu.org
https://lists.gnu.org/mailman/listinfo/help-make
41 matches
Mail list logo
