threads / discuss / 32810

Re: Re: Re: Re: Segmentation fault with latest git (070c57df)

Subject: Re: Re: Re: Re: Segmentation fault with latest git (070c57df)

## tl;dr

4 messages between Feb 1, 2013 and Feb 2, 2013.

replies: 3people: 3as markdown or json

Jongman Heo· Feb 1, 2013, 10:59 UTC · lore
Show 10 quoted lines
>> [...]
>>  Finished prerequisites of target file `builtin/fetch.o'.
>>  Prerequisite `builtin/fetch.c' is older than target `builtin/fetch.o'.
>>  Prerequisite `GIT-CFLAGS' is older than target `builtin/fetch.o'.
>> No need to remake target `builtin/fetch.o'.
>
>But it doesn't stimulate any prerequisites in make, which is weird.
>What's in builtin/.depend/fetch.o.d?
>
>-Peff
Hi, 
please see below~.
$ cat builtin/.depend/fetch.o.d            
fetch.o: builtin/fetch.c cache.h git-compat-util.h compat/bswap.h \
  strbuf.h hash.h advice.h gettext.h convert.h refs.h commit.h object.h \
  tree.h decorate.h builtin.h cache.h commit.h notes.h string-list.h \
  string-list.h remote.h transport.h remote.h run-command.h \
  parse-options.h sigchain.h submodule.h connected.h argv-array.h
cache.h:
git-compat-util.h:
compat/bswap.h:
strbuf.h:
hash.h:
advice.h:
gettext.h:
convert.h:
refs.h:
commit.h:
object.h:
tree.h:
decorate.h:
builtin.h:
cache.h:
commit.h:
notes.h:
string-list.h:
string-list.h:
remote.h:
transport.h:
remote.h:
run-command.h:
parse-options.h:
sigchain.h:
submodule.h:
connected.h:
argv-array.h:
Jonathan Nieder· Feb 1, 2013, 22:38 UTC · re: Jongman Heo · lore
Jongman Heo wrote:
>> But it doesn't stimulate any prerequisites in make, which is weird.
>> What's in builtin/.depend/fetch.o.d?
[...]
> please see below~.
>
> $ cat builtin/.depend/fetch.o.d            
> fetch.o: builtin/fetch.c cache.h git-compat-util.h compat/bswap.h \
That's the problem.  See the following thread:
  http://thread.gmane.org/gmane.comp.version-control.git/185625/focus=185680

Currently when COMPUTE_HEADER_DEPENDENCIES=auto git tests for dependency generation support by checking the output and exit status from the following command:

	$(CC) $(ALL_CFLAGS) -c -MF /dev/null -MMD -MP \
		-x c /dev/null -o /dev/null 2>&1

Perhaps this can be improved? Even something as simple as a ccache version test could presumably help a lot.

Hope that helps, Jonathan

Jonathan Nieder· Feb 1, 2013, 22:45 UTC · re: Jongman Heo · lore

[PATCH resend] Makefile: explicitly set target name for autogenerated dependencies

Date: Fri, 18 Nov 2011 17:23:24 -0600

"gcc -MF depfile -MMD -MP -c -o path/to/file.o" produces a makefile snippet named "depfile" describing what files are needed to build the target given by "-o". When ccache versions before v3.0pre0~187 (Fix handling of the -MD and -MDD options, 2009-11-01) run, they execute

	gcc -MF depfile -MMD -MP -E

instead to get the final content for hashing. Notice that the "-c -o" combination is replaced by "-E". The result is a target name without a leading path.

Thus when building git with such versions of ccache with COMPUTE_HEADER_DEPENDENCIES enabled, the generated makefile snippets define dependencies for the wrong target:

	$ make builtin/add.o
	GIT_VERSION = 1.7.8.rc3
	    * new build flags or prefix
	    CC builtin/add.o
	$ head -1 builtin/.depend/add.o.d
	add.o: builtin/add.c cache.h git-compat-util.h compat/bswap.h strbuf.h \

After a change in a header file, object files in a subdirectory are not automatically rebuilt by "make":

	$ touch cache.h
	$ make builtin/add.o
	$

Luckily we can prevent trouble by explicitly supplying the name of the target to ccache and gcc, using the -MQ option. Do so.

Reported-and-tested-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
Reported-by: : 허종만 <jongman.heo@samsung.com>
Signed-off-by: Jonathan Nieder <jrnieder@gmail.com>
---
 Makefile | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)
diff --git a/Makefile b/Makefile
index 731b6a8..5a2e02d 100644
--- a/Makefile
+++ b/Makefile
@@ -973,7 +973,8 @@ endif
 
 ifeq ($(COMPUTE_HEADER_DEPENDENCIES),auto)
 dep_check = $(shell $(CC) $(ALL_CFLAGS) \
-	-c -MF /dev/null -MMD -MP -x c /dev/null -o /dev/null 2>&1; \
+	-c -MF /dev/null -MQ /dev/null -MMD -MP \
+	-x c /dev/null -o /dev/null 2>&1; \
 	echo $$?)
 ifeq ($(dep_check),0)
 override COMPUTE_HEADER_DEPENDENCIES = yes
@@ -1843,7 +1844,7 @@ $(dep_dirs):
 
 missing_dep_dirs := $(filter-out $(wildcard $(dep_dirs)),$(dep_dirs))
 dep_file = $(dir $@).depend/$(notdir $@).d
-dep_args = -MF $(dep_file) -MMD -MP
+dep_args = -MF $(dep_file) -MQ $@ -MMD -MP
 ifdef CHECK_HEADER_DEPENDENCIES
 $(error cannot compute header dependencies outside a normal build. \
 Please unset CHECK_HEADER_DEPENDENCIES and try again)

-- 
1.8.1
Jeff King· Feb 2, 2013, 00:03 UTC · re: Jonathan Nieder · lore

Re: [PATCH resend] Makefile: explicitly set target name for autogenerated dependencies

On Fri, Feb 01, 2013 at 02:45:19PM -0800, Jonathan Nieder wrote:
Show 9 quoted lines
> After a change in a header file, object files in a subdirectory are
> not automatically rebuilt by "make":
> 
> 	$ touch cache.h
> 	$ make builtin/add.o
> 	$
> 
> Luckily we can prevent trouble by explicitly supplying the name of the
> target to ccache and gcc, using the -MQ option.  Do so.

Thanks, I missed the original thread last year, but this does look like the same problem. The fixed version of ccache is a few years old, but the OP did mention RHEL5, so that makes sense.

>  missing_dep_dirs := $(filter-out $(wildcard $(dep_dirs)),$(dep_dirs))
>  dep_file = $(dir $@).depend/$(notdir $@).d
> -dep_args = -MF $(dep_file) -MMD -MP
> +dep_args = -MF $(dep_file) -MQ $@ -MMD -MP

This looks like a nice simple solution. The only downside would be if -MQ is not supported by some gcc versions (in which case we would do better to improve the "can we use computed dependencies" check, and punish people on old ccache, not people on old gcc who otherwise could work without -MQ).

As far as I can tell, though, -MQ came along with the other dependency generation options in 2001:

  http://gcc.gnu.org/news/dependencies.html

which means it's not a problem to rely on. So the patch looks good to me.

-Peff

← back to recent threads