Skip to contentSkip to conversation

perf tools: Fix build error on read only source.

PATCH Arnaldo Carvalho de Melo 4 participants 14 messages Aug 17, 2010 – Aug 19, 2010 perf events

lore
Share & tools
lore
#0 lore raw

Em Tue, Aug 17, 2010 at 11:17:54AM -0300, Arnaldo Carvalho de Melo escreveu:

Quoted 14 lines

Em Tue, Aug 17, 2010 at 11:08:40PM +0900, Kusanagi Kouichi escreveu:

tools/perf/feature-tests.mak endef · line 119

Quoted 11 lines
tools/perf/feature-tests.mak +1 -1
@@ -113,7 +113,7 @@ endef
 # try-cc
 # Usage: option = $(call try-cc, source-to-build, cc-options)
 try-cc = $(shell sh -c						  \
-	'TMP="$(TMPOUT).$$$$";			 		  \
+	'TMP="$(OUTPUT)$(TMPOUT).$$$$";				  \
 	 echo "$(1)" |						  \
 	 $(CC) -x c - $(2) -o "$$TMP" > /dev/null 2>&1 && echo y; \
 	 rm -f "$$TMP"')

Much better! Will test and merge, thanks!

Still has issues:

[acme@emilia linux-2.6-tip]$ rm -rf ~/git/build/perf/
[acme@emilia linux-2.6-tip]$ make O=~/git/build/perf -s -j9 -C tools/perf install
PERF_VERSION = 2.6.36.rc1.2.g19bad
* new build flags or prefix
[acme@emilia linux-2.6-tip]$ patch -p1 < Kouichi.patch
patching file tools/perf/Makefile
patching file tools/perf/feature-tests.mak
[acme@emilia linux-2.6-tip]$ rm -rf ~/git/build/perf/
[acme@emilia linux-2.6-tip]$ make O=~/git/build/perf -s -j9 -C tools/perf install
Makefile:503: No libdw.h found or old libdw.h found or elfutils is older than 0.138, disables dwarf support. Please install new elfutils-devel/libdw-dev Makefile:534: *** No gnu/libc-version.h found, please install glibc-dev[el]/glibc-static. Stop.
[acme@emilia linux-2.6-tip]$

Can you please investigate?

Thanks,

- Arnaldo

#1 lore raw
Quoted 31 lines · Arnaldo Carvalho de Melo

Em Tue, Aug 17, 2010 at 11:17:54AM -0300, Arnaldo Carvalho de Melo escreveu:

Quoted 14 lines

Em Tue, Aug 17, 2010 at 11:08:40PM +0900, Kusanagi Kouichi escreveu:

tools/perf/feature-tests.mak endef · line 119

Quoted 11 lines
tools/perf/feature-tests.mak +1 -1
@@ -113,7 +113,7 @@ endef
 # try-cc
 # Usage: option = $(call try-cc, source-to-build, cc-options)
 try-cc = $(shell sh -c						  \
-	'TMP="$(TMPOUT).$$$$";			 		  \
+	'TMP="$(OUTPUT)$(TMPOUT).$$$$";				  \
 	 echo "$(1)" |						  \
 	 $(CC) -x c - $(2) -o "$$TMP" > /dev/null 2>&1 && echo y; \
 	 rm -f "$$TMP"')

Much better! Will test and merge, thanks!

Still has issues:

[acme@emilia linux-2.6-tip]$ rm -rf ~/git/build/perf/
[acme@emilia linux-2.6-tip]$ make O=~/git/build/perf -s -j9 -C tools/perf install
PERF_VERSION = 2.6.36.rc1.2.g19bad
* new build flags or prefix
[acme@emilia linux-2.6-tip]$ patch -p1 < Kouichi.patch
patching file tools/perf/Makefile
patching file tools/perf/feature-tests.mak
[acme@emilia linux-2.6-tip]$ rm -rf ~/git/build/perf/
[acme@emilia linux-2.6-tip]$ make O=~/git/build/perf -s -j9 -C tools/perf install
Makefile:503: No libdw.h found or old libdw.h found or elfutils is older than 0.138, disables dwarf support. Please install new elfutils-devel/libdw-dev Makefile:534: *** No gnu/libc-version.h found, please install glibc-dev[el]/glibc-static. Stop.
[acme@emilia linux-2.6-tip]$

Quoted 5 lines

Can you please investigate?

Thanks,

- Arnaldo

With POSIX shell patch, $(OUTPUT) has not been created yet when it is used by try-cc. This fixes the issue anyway.

tools/perf/Makefile +1 -0
index dcb9700..5b1c12b 100644
@@ -281,6 +281,7 @@ endif
 
 -include feature-tests.mak
 
+$(shell mkdir -p $(OUTPUT) 2> /dev/null)
 ifeq ($(call try-cc,$(SOURCE_HELLO),-Werror -fstack-protector-all),y)
 	CFLAGS := $(CFLAGS) -fstack-protector-all
 endif
maintainer
#2 lore raw
Quoted 8 lines · Kusanagi Kouichi

With POSIX shell patch, $(OUTPUT) has not been created yet when it is used by try-cc. This fixes the issue anyway.

tools/perf/Makefile +0 -0
index dcb9700..5b1c12b 100644

tools/perf/Makefile endif · line 287

Quoted 7 lines
tools/perf/Makefile +1 -0
@@ -281,6 +281,7 @@ endif
 
 -include feature-tests.mak
 
+$(shell mkdir -p $(OUTPUT) 2> /dev/null)
 ifeq ($(call try-cc,$(SOURCE_HELLO),-Werror -fstack-protector-all),y)
 	CFLAGS := $(CFLAGS) -fstack-protector-all
 endif

So you're wanting to provide a non-existing O=foo ?

I don't think we should, simply bail when it doesn't exist, kernel builds do the same:

# make O=plop-build bzImage
/bin/sh: line 0: cd: plop-build: No such file or directory
Makefile:117: *** output directory "plop-build" does not exist.  Stop.
#3 lore raw

Em Wed, Aug 18, 2010 at 02:40:24PM +0200, Peter Zijlstra escreveu:

Quoted 20 lines · Kusanagi Kouichi
Quoted 9 lines

With POSIX shell patch, $(OUTPUT) has not been created yet when it is used by try-cc. This fixes the issue anyway.

tools/perf/Makefile +0 -0
index dcb9700..5b1c12b 100644

tools/perf/Makefile endif · line 287

Quoted 7 lines
tools/perf/Makefile +1 -0
@@ -281,6 +281,7 @@ endif
 
 -include feature-tests.mak
 
+$(shell mkdir -p $(OUTPUT) 2> /dev/null)
 ifeq ($(call try-cc,$(SOURCE_HELLO),-Werror -fstack-protector-all),y)
 	CFLAGS := $(CFLAGS) -fstack-protector-all
 endif

So you're wanting to provide a non-existing O=foo ?

Quoted 6 lines

I don't think we should, simply bail when it doesn't exist, kernel builds do the same:

# make O=plop-build bzImage
/bin/sh: line 0: cd: plop-build: No such file or directory
Makefile:117: *** output directory "plop-build" does not exist.  Stop.

Agreed that making it work like the kernel brings less surprises, but isn't it convenient to be able to just do:

rm -rf ~/build
make O=~/build

and have everything work? Saving some keystrokes 8)

- Arnaldo

maintainer
#4 lore raw
Quoted 4 lines · Arnaldo Carvalho de Melo

rm -rf ~/build
make O=~/build

and have everything work? Saving some keystrokes 8)

The turn side is that if you typo the output dir you'll probably not ever find it again.. and you could have typed: rm -rf ~/build/* instead.

#5 lore raw

Em Wed, Aug 18, 2010 at 04:27:45PM +0200, Peter Zijlstra escreveu:

Quoted 8 lines · Arnaldo Carvalho de Melo
Quoted 4 lines

rm -rf ~/build
make O=~/build

and have everything work? Saving some keystrokes 8)

The turn side is that if you typo the output dir you'll probably not ever find it again.. and you could have typed: rm -rf ~/build/* instead.

Not something I'll argue too much :)

I'll apply the patch as it fixes a bug and later make the perf Makefile complain about the output dir not existing, i.e. providing a message similar to what the kernel proper does instead of the cryptic message we get today.

If Kouichi-san or Bernd could provide that patch for perf/core, even better! ;-)

- Arnaldo

#6 lore raw

Em Wed, Aug 18, 2010 at 01:56:42PM -0300, Arnaldo Carvalho de Melo escreveu:

Quoted 10 lines

Em Wed, Aug 18, 2010 at 04:27:45PM +0200, Peter Zijlstra escreveu:

Quoted 8 lines · Arnaldo Carvalho de Melo
Quoted 4 lines

rm -rf ~/build
make O=~/build

and have everything work? Saving some keystrokes 8)

The turn side is that if you typo the output dir you'll probably not ever find it again.. and you could have typed: rm -rf ~/build/* instead.

Quoted 6 lines

Not something I'll argue too much :)

I'll apply the patch as it fixes a bug and later make the perf Makefile complain about the output dir not existing, i.e. providing a message similar to what the kernel proper does instead of the cryptic message we get today.

[acme@emilia linux-2.6-tip]$ rm -rf ~/git/build/perf/
[acme@emilia linux-2.6-tip]$ make O=~/git/build/perf -C tools/perf
make: Entering directory `/home/acme/git/linux-2.6-tip/tools/perf'
Makefile:503: No libdw.h found or old libdw.h found or elfutils is older than 0.138, disables dwarf support. Please install new elfutils-devel/libdw-dev Makefile:534: *** No gnu/libc-version.h found, please install glibc-dev[el]/glibc-static. Stop.
make: Leaving directory `/home/acme/git/linux-2.6-tip/tools/perf'
[acme@emilia linux-2.6-tip]$

- Arnaldo

#7 lore raw

Em Wed, Aug 18, 2010 at 02:01:58PM -0300, Arnaldo Carvalho de Melo escreveu:

Quoted 12 lines

Em Wed, Aug 18, 2010 at 01:56:42PM -0300, Arnaldo Carvalho de Melo escreveu:

Quoted 11 lines

Em Wed, Aug 18, 2010 at 04:27:45PM +0200, Peter Zijlstra escreveu:

Quoted 8 lines · Arnaldo Carvalho de Melo
Quoted 4 lines

rm -rf ~/build
make O=~/build

and have everything work? Saving some keystrokes 8)

The turn side is that if you typo the output dir you'll probably not ever find it again.. and you could have typed: rm -rf ~/build/* instead.

Not something I'll argue too much :)

Ok, here it is, stolen directly from the toplevel kernel Makefile, will push to Ingo today:

[acme@emilia linux-2.6-tip]$ rm -rf ~/git/build/perf
[acme@emilia linux-2.6-tip]$ make -j9 O=~/git/build/perf -C tools/perf /bin/sh: line 0: cd: /home/acme/git/build/perf/: No such file or directory
make: Entering directory `/home/acme/git/linux-2.6-tip/tools/perf' Makefile:12: *** output directory "/home/acme/git/build/perf/" does not exist. Stop.
make: Leaving directory `/home/acme/git/linux-2.6-tip/tools/perf' [acme@emilia linux-2.6-tip]$ mkdir ~/git/build/perf
[acme@emilia linux-2.6-tip]$ make -j9 O=~/git/build/perf -C tools/perf
make: Entering directory `/home/acme/git/linux-2.6-tip/tools/perf'
MKDIR /home/acme/git/build/perf/bench
<SNIP>
PERF_VERSION = 2.6.36.rc1.2.g033a273.dirty
make: Leaving directory `/home/acme/git/linux-2.6-tip/tools/perf'
make: Entering directory `/home/acme/git/linux-2.6-tip/tools/perf'
GEN /home/acme/git/build/perf/common-cmds.h
* new build flags or prefix
GEN perf-archive
CC /home/acme/git/build/perf/builtin-annotate.o
<SNIP>
AR /home/acme/git/build/perf/libperf.a
LINK /home/acme/git/build/perf/perf
make: Leaving directory `/home/acme/git/linux-2.6-tip/tools/perf' [acme@emilia linux-2.6-tip]$

- Arnaldo

tools/perf/Makefile +10 -4
index dcb9700..4f1fa77 100644
@@ -5,6 +5,12 @@ endif
 # The default target of this Makefile is...
 all::
 
+ifneq ($(OUTPUT),)
+# check that the output directory actually exists
+OUTDIR := $(shell cd $(OUTPUT) && /bin/pwd)
+$(if $(OUTDIR),, $(error output directory "$(OUTPUT)" does not exist))
+endif
+
 # Define V=1 to have a more verbose compile.
 # Define V=2 to have an even more verbose compile.
 #
@@ -931,15 +937,15 @@ $(OUTPUT)common-cmds.h: $(wildcard Documentation/perf-*.txt)
 	$(QUIET_GEN). util/generate-cmdlist.sh > $@+ && mv $@+ $@
 
 $(patsubst %.sh,%,$(SCRIPT_SH)) : % : %.sh
-	$(QUIET_GEN)$(RM) $@ $@+ && \
+	$(QUIET_GEN)$(RM) $(OUTPUT)$@ $(OUTPUT)$@+ && \
 	sed -e '1s|#!.*/sh|#!$(SHELL_PATH_SQ)|' \
 	    -e 's|@SHELL_PATH@|$(SHELL_PATH_SQ)|' \
 	    -e 's|@@PERL@@|$(PERL_PATH_SQ)|g' \
 	    -e 's/@@PERF_VERSION@@/$(PERF_VERSION)/g' \
 	    -e 's/@@NO_CURL@@/$(NO_CURL)/g' \
-	    [email protected] >$@+ && \
-	chmod +x $@+ && \
-	mv $@+ $(OUTPUT)$@
+	    [email protected] > $(OUTPUT)$@+ && \
+	chmod +x $(OUTPUT)$@+ && \
+	mv $(OUTPUT)$@+ $(OUTPUT)$@
 
 configure: configure.ac
 	$(QUIET_GEN)$(RM) $@ $<+ && \
tools/perf/feature-tests.mak +1 -1
index ddb68e6..7a7b608 100644
@@ -113,7 +113,7 @@ endef
 # try-cc
 # Usage: option = $(call try-cc, source-to-build, cc-options)
 try-cc = $(shell sh -c						  \
-	'TMP="$(TMPOUT).$$$$";			 		  \
+	'TMP="$(OUTPUT)$(TMPOUT).$$$$";				  \
 	 echo "$(1)" |						  \
 	 $(CC) -x c - $(2) -o "$$TMP" > /dev/null 2>&1 && echo y; \
 	 rm -f "$$TMP"')
#8 lore raw
Quoted 11 lines · Arnaldo Carvalho de Melo

Em Wed, Aug 18, 2010 at 02:01:58PM -0300, Arnaldo Carvalho de Melo escreveu:

Quoted 10 lines

Em Wed, Aug 18, 2010 at 01:56:42PM -0300, Arnaldo Carvalho de Melo escreveu:

Quoted 9 lines

Em Wed, Aug 18, 2010 at 04:27:45PM +0200, Peter Zijlstra escreveu:

Quoted 8 lines · Arnaldo Carvalho de Melo
Quoted 4 lines

rm -rf ~/build
make O=~/build

and have everything work? Saving some keystrokes 8)

The turn side is that if you typo the output dir you'll probably not ever find it again.. and you could have typed: rm -rf ~/build/* instead.

Well, usually interactive shells have a history. But since the other Makefiles also expect existing destination top directory ...

Quoted 3 lines
Quoted 1 line
Quoted 1 line

Not something I'll argue too much :)

Ok, here it is, stolen directly from the toplevel kernel Makefile, will push to

ACK, it's the same problem so the same solution is the best.

Quoted 5 lines

Ingo today:

[acme@emilia linux-2.6-tip]$ rm -rf ~/git/build/perf [acme@emilia linux-2.6-tip]$ make -j9 O=~/git/build/perf -C tools/perf /bin/sh: line 0: cd: /home/acme/git/build/perf/: No such file or directory

ACK, that should make the "problem" pretty clear.

Bernd

Signature · 2 lines
mobile: +43 664 4416156              http://www.sysprog.at/
    Linux Software Development, Consulting and Services
#9 lore raw
In reply to Kusanagi Kouichi

Em Wed, Aug 18, 2010 at 09:37:12PM +0900, Kusanagi Kouichi escreveu:

Quoted 29 lines · Arnaldo Carvalho de Melo
Quoted 11 lines

Em Tue, Aug 17, 2010 at 11:17:54AM -0300, Arnaldo Carvalho de Melo escreveu:

Quoted 9 lines

Em Tue, Aug 17, 2010 at 11:08:40PM +0900, Kusanagi Kouichi escreveu:

Review endef · line 115

Quoted 6 lines

+++ b/tools/perf/feature-tests.mak

(quoted diff) +1 -1
@@ -113,7 +113,7 @@ endef
 try-cc = $(shell sh -c						  \
-	'TMP="$(TMPOUT).$$$$";			 		  \
+	'TMP="$(OUTPUT)$(TMPOUT).$$$$";				  \
 	 echo "$(1)" |						  \

Much better! Will test and merge, thanks!

Quoted 8 lines

Still has issues:
[acme@emilia linux-2.6-tip]$ rm -rf ~/git/build/perf/
[acme@emilia linux-2.6-tip]$ make O=~/git/build/perf -s -j9 -C tools/perf install
Makefile:503: No libdw.h found or old libdw.h found or elfutils is older than 0.138, disables dwarf support. Please install new elfutils-devel/libdw-dev Makefile:534: *** No gnu/libc-version.h found, please install glibc-dev[el]/glibc-static. Stop.
[acme@emilia linux-2.6-tip]$

Can you please investigate?

With POSIX shell patch, $(OUTPUT) has not been created yet when it is used by try-cc. This fixes the issue anyway.

tools/perf/Makefile +0 -0
index dcb9700..5b1c12b 100644

tools/perf/Makefile endif · line 284

Quoted 4 lines
tools/perf/Makefile +1 -0
@@ -281,6 +281,7 @@ endif
 
 -include feature-tests.mak
 
+$(shell mkdir -p $(OUTPUT) 2> /dev/null)

But with this we're back using $(shell mkdir), Bernd, ideas on how to properly fix this, probably the feature tests have to be triggered by the first rule and probably be something like:

$(phony feature-tests):
-include feature-tests.mak

DIRECTORY_DEPS = $(LIB_OBJS) $(BUILTIN_OBJS) $(OUTPUT)PERF-VERSION-FILE $(OUTPUT)common-cmds.h $(DIRECTORY_DEPS): $(sort $(dir $(DIRECTORY_DEPS))) feature-tests

completely untested! Unsure if this would trigger the mkdir first tho.

Right?

- Arnaldo

#10 lore raw
Quoted 33 lines · Arnaldo Carvalho de Melo

Em Wed, Aug 18, 2010 at 09:37:12PM +0900, Kusanagi Kouichi escreveu:

Quoted 32 lines · Arnaldo Carvalho de Melo
Quoted 11 lines

Em Tue, Aug 17, 2010 at 11:17:54AM -0300, Arnaldo Carvalho de Melo escreveu:

Quoted 9 lines

Em Tue, Aug 17, 2010 at 11:08:40PM +0900, Kusanagi Kouichi escreveu:

Review endef · line 115

Quoted 6 lines

+++ b/tools/perf/feature-tests.mak

(quoted diff) +1 -1
@@ -113,7 +113,7 @@ endef
 try-cc = $(shell sh -c						  \
-	'TMP="$(TMPOUT).$$$$";			 		  \
+	'TMP="$(OUTPUT)$(TMPOUT).$$$$";				  \
 	 echo "$(1)" |						  \

Much better! Will test and merge, thanks!

Quoted 8 lines

Still has issues:
[acme@emilia linux-2.6-tip]$ rm -rf ~/git/build/perf/
[acme@emilia linux-2.6-tip]$ make O=~/git/build/perf -s -j9 -C tools/perf install
Makefile:503: No libdw.h found or old libdw.h found or elfutils is older than 0.138, disables dwarf support. Please install new elfutils-devel/libdw-dev Makefile:534: *** No gnu/libc-version.h found, please install glibc-dev[el]/glibc-static. Stop.
[acme@emilia linux-2.6-tip]$

Can you please investigate?

With POSIX shell patch, $(OUTPUT) has not been created yet when it is used by try-cc. This fixes the issue anyway.

tools/perf/Makefile +0 -0
index dcb9700..5b1c12b 100644
@@ -281,6 +281,7 @@ endif
 
 -include feature-tests.mak
 
Quoted 7 lines

tools/perf/Makefile endif · line 284

Quoted 1 line
tools/perf/Makefile +1 -0
+$(shell mkdir -p $(OUTPUT) 2> /dev/null)

But with this we're back using $(shell mkdir), Bernd, ideas on how to properly fix this, probably the feature tests have to be triggered by the first rule and probably be something like:

$(phony feature-tests):

First, this should probably have been:
.PHONY: feature-tests

Quoted 8 lines

-include feature-tests.mak

DIRECTORY_DEPS = $(LIB_OBJS) $(BUILTIN_OBJS) $(OUTPUT)PERF-VERSION-FILE $(OUTPUT)common-cmds.h $(DIRECTORY_DEPS): $(sort $(dir $(DIRECTORY_DEPS))) feature-tests

completely untested! Unsure if this would trigger the mkdir first tho.

Right?

This doesn't trigger it. I played around a little bit and didn't find any way to trigger any rule.
The "problem" is IMHO that "try-cc" is used in "ifeq" clauses and these are evaluated as make goes through the Makefile before the first (real) rule (like "all") is triggered.

So ATM I see only 3 possibilities:
1) The old (aka ugly) "solution" from above.
2) Add to try-cc something like

	[ -d "$(OUTPUT)" ] || mkdir -p "$(OUTPUT)" 2>/dev/null;

before the line with the "echo".
The '[ -d "$(OUTPUT)" ]' is not really necessary as mkdir -p is a no-op on existing directories.
3) Another possibility would be to move the temporary file into the /tmp directory as into
---- snip ----

try-cc = $(shell sh -c						  \
	'TMP="/tmp/$(TMPOUT).$$$$";		 		  \
	 echo "$(1)" |						  \
	 $(CC) -x c - $(2) -o "$$TMP" > /dev/null 2>&1 && echo y; \
	 rm -f "$$TMP"')

---- snip ----
If yes, we could use (and thus depend on) mktemp - which is usual shell script way to generate unique temporary filename - as in ---- snip ----

try-cc = $(shell sh -c   \
	'TMP="$$(mktemp)";   \
	 echo "$(1)" |   \
	 $(CC) -x c - $(2) -o "$$TMP" > /dev/null 2>&1 && echo y; \
	 rm -f "$$TMP"')

---- snip ----

Bernd

Signature · 2 lines
mobile: +43 664 4416156              http://www.sysprog.at/
    Linux Software Development, Consulting and Services
#11 lore raw

Em Wed, Aug 18, 2010 at 05:25:41PM +0200, Bernd Petrovitsch escreveu:

Quoted 8 lines · Arnaldo Carvalho de Melo
Quoted 5 lines

But with this we're back using $(shell mkdir), Bernd, ideas on how to properly fix this, probably the feature tests have to be triggered by the first rule and probably be something like:

$(phony feature-tests):

First, this should probably have been:
.PHONY: feature-tests

Humm, isn't $(phony feature-tests): variation valid and shorter, humm, yeah, seems to be for !GNU make :-\

(http://makepp.sourceforge.net/1.18/t_phony.html)

Quoted 15 lines
Quoted 8 lines

-include feature-tests.mak

DIRECTORY_DEPS = $(LIB_OBJS) $(BUILTIN_OBJS) $(OUTPUT)PERF-VERSION-FILE $(OUTPUT)common-cmds.h $(DIRECTORY_DEPS): $(sort $(dir $(DIRECTORY_DEPS))) feature-tests

completely untested! Unsure if this would trigger the mkdir first tho.

Right?

This doesn't trigger it. I played around a little bit and didn't find any way to trigger any rule.
The "problem" is IMHO that "try-cc" is used in "ifeq" clauses and these are evaluated as make goes through the Makefile before the first (real) rule (like "all") is triggered.

Quoted 7 lines

So ATM I see only 3 possibilities:
1) The old (aka ugly) "solution" from above.
2) Add to try-cc something like

	[ -d "$(OUTPUT)" ] || mkdir -p "$(OUTPUT)" 2>/dev/null;

before the line with the "echo".
The '[ -d "$(OUTPUT)" ]' is not really necessary as mkdir -p is a no-op on existing directories.

Probably 2) is the best, since we'll end up creating that dir anyway. But as PeterZ pointed out, this is all moot since the kernel behaviour for O= is to expect the directory specified to already exist.

We're left with the task of checking that and providing a meaningful answer if not, like this:

[acme@emilia linux-2.6-tip]$ ls -la ~/git/build/foo
ls: cannot access /home/acme/git/build/foo: No such file or directory [acme@emilia linux-2.6-tip]$ make O=~/git/build/foo
/bin/sh: line 0: cd: /home/acme/git/build/foo: No such file or directory Makefile:117: *** output directory "/home/acme/git/build/foo" does not exist. Stop. [acme@emilia linux-2.6-tip]$

Quoted 11 lines

3) Another possibility would be to move the temporary file into the /tmp directory as into
---- snip ----

try-cc = $(shell sh -c						  \
	'TMP="/tmp/$(TMPOUT).$$$$";		 		  \
	 echo "$(1)" |						  \
	 $(CC) -x c - $(2) -o "$$TMP" > /dev/null 2>&1 && echo y; \
	 rm -f "$$TMP"')

---- snip ----
If yes, we could use (and thus depend on) mktemp - which is usual shell script way to generate unique temporary filename - as in

There is no use of mktemp in Kbuild, that I try hard to use as reference here.

- Arnaldo

#12 lore raw
Quoted 2 lines · Arnaldo Carvalho de Melo

Em Wed, Aug 18, 2010 at 05:25:41PM +0200, Bernd Petrovitsch escreveu:

Quoted 1 line · Arnaldo Carvalho de Melo

[...]

Quoted 8 lines
Quoted 3 lines
Quoted 1 line

$(phony feature-tests):

First, this should probably have been:
.PHONY: feature-tests

Humm, isn't $(phony feature-tests): variation valid and shorter, humm, yeah, seems to be for !GNU make :-\

(http://makepp.sourceforge.net/1.18/t_phony.html)

Funny, I didn't knew GNU-make's "$(phony target): " construct.

Hmm, what GNU-make versions and other makes are considered (or should be) supported?

[...]

Quoted 2 lines

There is no use of mktemp in Kbuild, that I try hard to use as reference here.

FWIW fine with me - I just threw an idea in.

Bernd

Signature · 2 lines
mobile: +43 664 4416156              http://www.sysprog.at/
    Linux Software Development, Consulting and Services
#13 lore raw

Em Thu, Aug 19, 2010 at 12:31:48PM +0200, Bernd Petrovitsch escreveu:

Quoted 8 lines · Arnaldo Carvalho de Melo
Quoted 2 lines

Em Wed, Aug 18, 2010 at 05:25:41PM +0200, Bernd Petrovitsch escreveu:

Quoted 1 line · Arnaldo Carvalho de Melo

[...]

Quoted 4 lines
Quoted 3 lines
Quoted 1 line

$(phony feature-tests):

First, this should probably have been:
.PHONY: feature-tests

Quoted 6 lines
Quoted 4 lines

Humm, isn't $(phony feature-tests): variation valid and shorter, humm, yeah, seems to be for !GNU make :-\

(http://makepp.sourceforge.net/1.18/t_phony.html)

Funny, I didn't knew GNU-make's "$(phony target): " construct.

it doesn't have AFAIK, its just this makepp, that I never had heard about before, that has it, probably it introduced the idea. Quite sensible, it seems.

Quoted 2 lines

Hmm, what GNU-make versions and other makes are considered (or should be) supported?

I think only GNU make is supported.

Quoted 5 lines

[...]

Quoted 2 lines

There is no use of mktemp in Kbuild, that I try hard to use as reference here.

FWIW fine with me - I just threw an idea in.

Right, and I considered it, using Kbuild usage as a validation gauge to say we probably shouldn't use it :)

- Arnaldo