threads / rfc / 20212

RFC patchSunOS grep does not understand -C<n>

Subject: [PATCH/RFC] SunOS grep does not understand -C<n>

## tl;dr

8 messages between Jul 24, 2009 and Dec 26, 2009. Diffs are folded; open one to read it.

replies: 7people: 3as markdown or json

Junio C Hamano· Jul 24, 2009, 05:30 UTC · lore

I noticed that the first "grep -C1" test in t7002 does not pass on my SunOS-5.11-i86pc, and that is not because our way to spawn external grep is broken, but because the native grep does not understand -C<n>.

Is it just me and my installation (i.e. I might have failed to install saner grep from the distribution that everybody uses), or everybody on SunOS is using this option himself because our Makefile doesn't do that automatically for them?

Just in case it is the latter, here is a proposed patch.
---
 Makefile |    1 +
 1 files changed, 1 insertions(+), 0 deletions(-)
Show changes to Makefile +1 −0
diff --git a/Makefile b/Makefile
index bde27ed..daf4296 100644
--- a/Makefile
+++ b/Makefile
@@ -728,6 +728,7 @@ ifeq ($(uname_S),SunOS)
 	NO_MKDTEMP = YesPlease
 	NO_MKSTEMPS = YesPlease
 	NO_REGEX = YesPlease
+	NO_EXTERNAL_GREP = YesPlease
 	ifeq ($(uname_R),5.7)
 		NEEDS_RESOLV = YesPlease
 		NO_IPV6 = YesPlease
Jeff King· Jul 24, 2009, 05:41 UTC · re: Junio C Hamano · lore

Re: [PATCH/RFC] SunOS grep does not understand -C<n>

On Thu, Jul 23, 2009 at 10:30:07PM -0700, Junio C Hamano wrote:
Show 10 quoted lines
> I noticed that the first "grep -C1" test in t7002 does not pass on my 
> SunOS-5.11-i86pc, and that is not because our way to spawn external
> grep is broken, but because the native grep does not understand -C<n>.
> 
> Is it just me and my installation (i.e. I might have failed to install
> saner grep from the distribution that everybody uses), or everybody on
> SunOS is using this option himself because our Makefile doesn't do that
> automatically for them?
> 
> Just in case it is the latter, here is a proposed patch.

Yes, I've been building with NO_EXTERNAL_GREP for my test box. The grep from Solaris 8, at least, doesn't understand '-e' either, which causes it to fail many other tests.

-Peff
Junio C Hamano· Jul 24, 2009, 06:27 UTC · re: Jeff King · lore

Re: [PATCH/RFC] SunOS grep does not understand -C<n>

Jeff King <peff@peff.net> writes:
Show 16 quoted lines
> On Thu, Jul 23, 2009 at 10:30:07PM -0700, Junio C Hamano wrote:
>
>> I noticed that the first "grep -C1" test in t7002 does not pass on my 
>> SunOS-5.11-i86pc, and that is not because our way to spawn external
>> grep is broken, but because the native grep does not understand -C<n>.
>> 
>> Is it just me and my installation (i.e. I might have failed to install
>> saner grep from the distribution that everybody uses), or everybody on
>> SunOS is using this option himself because our Makefile doesn't do that
>> automatically for them?
>> 
>> Just in case it is the latter, here is a proposed patch.
>
> Yes, I've been building with NO_EXTERNAL_GREP for my test box. The grep
> from Solaris 8, at least, doesn't understand '-e' either, which causes
> it to fail many other tests.
Thanks.

Just to double check, do we have "grep -e" (not "git grep -e") that makes Solaris fail? I didn't notice. Or do you mean "git grep -e" fails without NO_EXTERNAL_GREP?

Jeff King· Jul 24, 2009, 07:09 UTC · re: Junio C Hamano · lore

Re: [PATCH/RFC] SunOS grep does not understand -C<n>

On Thu, Jul 23, 2009 at 11:27:00PM -0700, Junio C Hamano wrote:
Show 9 quoted lines
> > Yes, I've been building with NO_EXTERNAL_GREP for my test box. The grep
> > from Solaris 8, at least, doesn't understand '-e' either, which causes
> > it to fail many other tests.
> 
> Thanks.
> 
> Just to double check, do we have "grep -e" (not "git grep -e") that makes
> Solaris fail?  I didn't notice.  Or do you mean "git grep -e" fails
> without NO_EXTERNAL_GREP?

I meant the latter; t7002 tests "git grep -e", which fails without NO_EXTERNAL_GREP.

-Peff
Junio C Hamano· Jul 24, 2009, 06:29 UTC · re: Junio C Hamano · lore

[PATCH/RFC] FreeBSD iconv function signature is "old"

I noticed that my build on FreeBSD-7.2-RELEASE-i386 fails without OLD_ICONV=YesPlease.

Is it just me and my installation (i.e. I might have failed to install saner iconv from the port that everybody uses), or is everybody who runs on FreeBSD using this option himself because our Makefile doesn't do that automatically for them?

Just in case it is the latter, here is a proposed patch.
---
 Makefile |    1 +
 1 files changed, 1 insertions(+), 0 deletions(-)
Show changes to Makefile +1 −0
diff --git a/Makefile b/Makefile
index bde27ed..9a25c08 100644
--- a/Makefile
+++ b/Makefile
@@ -777,6 +777,7 @@ ifeq ($(uname_O),Cygwin)
 endif
 ifeq ($(uname_S),FreeBSD)
 	NEEDS_LIBICONV = YesPlease
+	OLD_ICONV = YesPlease
 	NO_MEMMEM = YesPlease
 	BASIC_CFLAGS += -I/usr/local/include
 	BASIC_LDFLAGS += -L/usr/local/lib
Junio C Hamano· Dec 26, 2009, 09:09 UTC · re: Junio C Hamano · lore

[PATCH/RFC] FreeBSD iconv function signature is "old"

Junio C Hamano <gitster@pobox.com> writes:
Show 9 quoted lines
> I noticed that my build on FreeBSD-7.2-RELEASE-i386 fails without
> OLD_ICONV=YesPlease.
>
> Is it just me and my installation (i.e. I might have failed to install
> saner iconv from the port that everybody uses), or is everybody who runs
> on FreeBSD using this option himself because our Makefile doesn't do that
> automatically for them?
>
> Just in case it is the latter, here is a proposed patch.

By the way, FreeBSD-8.0-RELEASE-i386 has the same issue. I'll queue it to 'maint' unless anybody objects...

Show 15 quoted lines
>  Makefile |    1 +
>  1 files changed, 1 insertions(+), 0 deletions(-)
>
> diff --git a/Makefile b/Makefile
> index bde27ed..9a25c08 100644
> --- a/Makefile
> +++ b/Makefile
> @@ -777,6 +777,7 @@ ifeq ($(uname_O),Cygwin)
>  endif
>  ifeq ($(uname_S),FreeBSD)
>  	NEEDS_LIBICONV = YesPlease
> +	OLD_ICONV = YesPlease
>  	NO_MEMMEM = YesPlease
>  	BASIC_CFLAGS += -I/usr/local/include
>  	BASIC_LDFLAGS += -L/usr/local/lib
Brandon Casey· Jul 24, 2009, 15:43 UTC · re: Junio C Hamano · lore

Re: [PATCH/RFC] SunOS grep does not understand -C<n>

Junio C Hamano wrote:
Show 8 quoted lines
> I noticed that the first "grep -C1" test in t7002 does not pass on my 
> SunOS-5.11-i86pc, and that is not because our way to spawn external
> grep is broken, but because the native grep does not understand -C<n>.
> 
> Is it just me and my installation (i.e. I might have failed to install
> saner grep from the distribution that everybody uses), or everybody on
> SunOS is using this option himself because our Makefile doesn't do that
> automatically for them?

Hmm. I have _not_ been setting NO_EXTERNAL_GREP and the tests have _not_ been failing.

The system grep does not seem to have the -C option:
   # /usr/xpg4/bin/grep -C1 include ws.c
   /usr/xpg4/bin/grep: illegal option -- C
   Usage:  grep [-E|-F] [-c|-l|-q] [-bhinsvwx] [file ...]
   grep [-E|-F] [-c|-l|-q] [-bhinsvwx] -e pattern... [-f pattern_file]...[file...]
   grep [-E|-F] [-c|-l|-q] [-bhinsvwx] [-e pattern]... -f pattern_file [file...]
but git grep -C1 works fine
   # git grep -C1 include ws.c
   ws.c-
   ws.c:#include "cache.h"
   ws.c:#include "attr.h"
   ws.c-

I have been compiling with the SUNWspro compiler suite though, and I see these statements in builtin-grep.c:

   #ifndef NO_EXTERNAL_GREP
   #ifdef __unix__
   #define NO_EXTERNAL_GREP 0
   #else
   #define NO_EXTERNAL_GREP 1
   #endif
   #endif

So possibly, the SUNWspro compiler does not set the __unix__ macro. A quick compile of a test program confirms it.

   #include <stdio.h>
   int main (int argc, char* argv[])
   {
   #ifdef __unix__
   	puts("__unix__ is set");
   #else
   	puts("__unix__ is not set");
   #endif
   	return 0;
   }
   # /opt/SUNWspro/bin/cc -o test.out test.c
   ./test.out
   __unix__ is not set
   # gcc -o test.out test.c
   __unix__ is set
So, I have been getting NO_EXTERNAL_GREP all along without knowing it.
-brandon
Junio C Hamano· Jul 24, 2009, 15:55 UTC · re: Brandon Casey · lore

Re: [PATCH/RFC] SunOS grep does not understand -C<n>

Brandon Casey <brandon.casey.ctr@nrlssc.navy.mil> writes:
> So, I have been getting NO_EXTERNAL_GREP all along without knowing it.

Thanks for a detailed analysis. I use gcc, and I suspect Peff also does. So it seems that I should apply my "in case it is the latter" patch.

Thanks.

← back to recent threads