[PATCH] gnu: Add ninja.

[PATCH] gnu: Add ninja.

[宋文武]

* gnu/packages/ninja.scm: New file.
* gnu-system.am (GNU_SYSTEM_MODULES): Add it.
---
 gnu-system.am          |  1 +
 gnu/packages/ninja.scm | 82 ++++++++++++++++++++++++++++++++++++++++++++++++++
 2 files changed, 83 insertions(+)
 create mode 100644 gnu/packages/ninja.scm

diff --git a/gnu-system.am b/gnu-system.am
index 33d870c..acea574 100644
--- a/gnu-system.am
+++ b/gnu-system.am
@@ -191,6 +191,7 @@ GNU_SYSTEM_MODULES =				\
   gnu/packages/ncurses.scm			\
   gnu/packages/netpbm.scm			\
   gnu/packages/nettle.scm			\
+  gnu/packages/ninja.scm			\
   gnu/packages/node.scm				\
   gnu/packages/noweb.scm			\
   gnu/packages/ntp.scm				\
diff --git a/gnu/packages/ninja.scm b/gnu/packages/ninja.scm
new file mode 100644
index 0000000..8939750
--- /dev/null
+++ b/gnu/packages/ninja.scm
@@ -0,0 +1,82 @@
+;;; GNU Guix --- Functional package management for GNU
+;;; Copyright © 2015 Sou Bunnbu <iyzsong@gmail.com>
+;;;
+;;; This file is part of GNU Guix.
+;;;
+;;; GNU Guix is free software; you can redistribute it and/or modify it
+;;; under the terms of the GNU General Public License as published by
+;;; the Free Software Foundation; either version 3 of the License, or (at
+;;; your option) any later version.
+;;;
+;;; GNU Guix is distributed in the hope that it will be useful, but
+;;; WITHOUT ANY WARRANTY; without even the implied warranty of
+;;; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+;;; GNU General Public License for more details.
+;;;
+;;; You should have received a copy of the GNU General Public License
+;;; along with GNU Guix.  If not, see <http://www.gnu.org/licenses/>.
+
+(define-module (gnu packages ninja)
+  #:use-module ((guix licenses) #:select (asl2.0))
+  #:use-module (guix packages)
+  #:use-module (guix download)
+  #:use-module (guix build-system gnu)
+  #:use-module (gnu packages)
+  #:use-module (gnu packages python))
+
+(define-public ninja
+  (package
+    (name "ninja")
+    (version "1.5.3")
+    (source (origin
+              (method url-fetch)
+              (uri (string-append "https://github.com/martine/ninja/"
+                                  "archive/v" version ".tar.gz"))
+              (sha256
+               (base32
+                "1h3yfwcfl61v493vna6jia2fizh8rpig7qw2504cvkr6gid3p5bw"))))
+    (build-system gnu-build-system)
+    (arguments
+     '(#:phases
+       (alist-replace
+        'configure
+        (lambda _
+          (substitute* "src/subprocess-posix.cc"
+            (("/bin/sh") (which "sh"))))
+        (alist-replace
+         'build
+         (lambda _
+           (zero? (system "./configure.py --bootstrap")))
+         (alist-replace
+          'check
+          (lambda _
+            (and (zero? (system "./configure.py"))
+                 (zero? (system "./ninja ninja_test"))
+                 ;; SubprocessTest.InterruptChild fail when using 'system*'.
+                 ;; SubprocessTest.SetWithLots was skipped.
+                 ;; XXX: Raise [ulimit -n] well above 1025 to make this test go.
+                 (zero? (system (string-append
+                                 "./ninja_test "
+                                 "--gtest_filter="
+                                 "-SubprocessTest.SetWithLots")))))
+          (alist-replace
+           'install
+           (lambda* (#:key outputs #:allow-other-keys)
+             (let* ((out (assoc-ref outputs "out"))
+                    (bin (string-append out "/bin"))
+                    (doc (string-append out "/share/doc/ninja")))
+               (mkdir-p bin)
+               (copy-file "ninja" (string-append bin "/ninja"))
+               (mkdir-p doc)
+               (copy-file "doc/manual.asciidoc"
+                          (string-append doc "/manual.asciidoc"))))
+           %standard-phases))))))
+    (native-inputs `(("python" ,python-2)))
+    (home-page "http://martine.github.io/ninja/")
+    (synopsis "Small build system")
+    (description
+     "Ninja is a small build system with a focus on speed.  It differs from
+other build systems in two major respects: it is designed to have its input
+files generated by a higher-level build system, and it is designed to run
+builds as fast as possible.")
+    (license asl2.0)))
-- 
2.1.2

Re: [PATCH] gnu: Add ninja.

[Ludovic Courtès]

宋文武 <iyzsong@gmail.com> skribis:

> * gnu/packages/ninja.scm: New file.
> * gnu-system.am (GNU_SYSTEM_MODULES): Add it.

[...]

> +          'check
> +          (lambda _
> +            (and (zero? (system "./configure.py"))
> +                 (zero? (system "./ninja ninja_test"))
> +                 ;; SubprocessTest.InterruptChild fail when using 'system*'.
> +                 ;; SubprocessTest.SetWithLots was skipped.
> +                 ;; XXX: Raise [ulimit -n] well above 1025 to make this test go.

Does it mean that the test is currently failing?

> +                 (zero? (system (string-append
> +                                 "./ninja_test "
> +                                 "--gtest_filter="
> +                                 "-SubprocessTest.SetWithLots")))))

Please use ‘system*’ (with separate arguments) rather than ‘system’.
The latter runs “/bin/sh -c ...” whereas the former runs the program
directly.

Thanks,
Ludo’.

Re: [PATCH] gnu: Add ninja.

[宋文武]

Ludovic Courtès <ludo@gnu.org> writes:

> 宋文武 <iyzsong@gmail.com> skribis:
>
>> * gnu/packages/ninja.scm: New file.
>> * gnu-system.am (GNU_SYSTEM_MODULES): Add it.
>
> [...]
>
>> +          'check
>> +          (lambda _
>> +            (and (zero? (system "./configure.py"))
>> +                 (zero? (system "./ninja ninja_test"))
>> +                 ;; SubprocessTest.InterruptChild fail when using 'system*'.
>> +                 ;; SubprocessTest.SetWithLots was skipped.
>> +                 ;; XXX: Raise [ulimit -n] well above 1025 to make this test go.
>
> Does it mean that the test is currently failing?
Yes, SetWithLots fail with the 'Raise ...' line.
>
>> +                 (zero? (system (string-append
>> +                                 "./ninja_test "
>> +                                 "--gtest_filter="
>> +                                 "-SubprocessTest.SetWithLots")))))
>
> Please use ‘system*’ (with separate arguments) rather than ‘system’.
> The latter runs “/bin/sh -c ...” whereas the former runs the program
> directly.
Use 'system*' to run "./ninja_test" will cause InterruptChild to fail :(
(as I mentioned in the comment)
>
> Thanks,
> Ludo’.

Re: [PATCH] gnu: Add ninja.

[Ludovic Courtès]

宋文武 <iyzsong@gmail.com> skribis:

> Ludovic Courtès <ludo@gnu.org> writes:
>
>> 宋文武 <iyzsong@gmail.com> skribis:
>>
>>> * gnu/packages/ninja.scm: New file.
>>> * gnu-system.am (GNU_SYSTEM_MODULES): Add it.
>>
>> [...]
>>
>>> +          'check
>>> +          (lambda _
>>> +            (and (zero? (system "./configure.py"))
>>> +                 (zero? (system "./ninja ninja_test"))
>>> +                 ;; SubprocessTest.InterruptChild fail when using 'system*'.
>>> +                 ;; SubprocessTest.SetWithLots was skipped.
>>> +                 ;; XXX: Raise [ulimit -n] well above 1025 to make this test go.
>>
>> Does it mean that the test is currently failing?
> Yes, SetWithLots fail with the 'Raise ...' line.

Oh, I see.  Then can you make it clearer in the comment:

  ;; SubprocessTest.SetWithLots fails with:
  ;;
  ;;   Raise [ulimit -n] well above 1025 to make this test go.
  ;;
  ;; Skip it.

>>> +                 (zero? (system (string-append
>>> +                                 "./ninja_test "
>>> +                                 "--gtest_filter="
>>> +                                 "-SubprocessTest.SetWithLots")))))
>>
>> Please use ‘system*’ (with separate arguments) rather than ‘system’.
>> The latter runs “/bin/sh -c ...” whereas the former runs the program
>> directly.
> Use 'system*' to run "./ninja_test" will cause InterruptChild to fail :(
> (as I mentioned in the comment)

Ah right, that’s weird, but could you mention in the comment (1) how it
fails, and (2) that because of this, we use ‘system’ instead of
‘system*’?

It’s important to explain both the problem and the solution/consequence,
otherwise it can be hard by reading the comment to understand if we’re
talking about an unfixed issue, a fixed issue, or a fix.

OK to push with these changes.

Thanks,
Ludo’.

Re: [PATCH] gnu: Add ninja.

[宋文武]

[-- Attachment #1: Type: text/plain, Size: 1996 bytes --]

Ludovic Courtès <ludo@gnu.org> writes:

> 宋文武 <iyzsong@gmail.com> skribis:
>
>> Ludovic Courtès <ludo@gnu.org> writes:
>>
>>> 宋文武 <iyzsong@gmail.com> skribis:
>>>
>>>> * gnu/packages/ninja.scm: New file.
>>>> * gnu-system.am (GNU_SYSTEM_MODULES): Add it.
>>>
>>> [...]
>>>
>>>> +          'check
>>>> +          (lambda _
>>>> +            (and (zero? (system "./configure.py"))
>>>> +                 (zero? (system "./ninja ninja_test"))
>>>> +                 ;; SubprocessTest.InterruptChild fail when using 'system*'.
>>>> +                 ;; SubprocessTest.SetWithLots was skipped.
>>>> +                 ;; XXX: Raise [ulimit -n] well above 1025 to make this test go.
>>>
>>> Does it mean that the test is currently failing?
>> Yes, SetWithLots fail with the 'Raise ...' line.
>
> Oh, I see.  Then can you make it clearer in the comment:
>
>   ;; SubprocessTest.SetWithLots fails with:
>   ;;
>   ;;   Raise [ulimit -n] well above 1025 to make this test go.
>   ;;
>   ;; Skip it.
>
>>>> +                 (zero? (system (string-append
>>>> +                                 "./ninja_test "
>>>> +                                 "--gtest_filter="
>>>> +                                 "-SubprocessTest.SetWithLots")))))
>>>
>>> Please use ‘system*’ (with separate arguments) rather than ‘system’.
>>> The latter runs “/bin/sh -c ...” whereas the former runs the program
>>> directly.
>> Use 'system*' to run "./ninja_test" will cause InterruptChild to fail :(
>> (as I mentioned in the comment)
>
> Ah right, that’s weird, but could you mention in the comment (1) how it
> fails, and (2) that because of this, we use ‘system’ instead of
> ‘system*’?
>
> It’s important to explain both the problem and the solution/consequence,
> otherwise it can be hard by reading the comment to understand if we’re
> talking about an unfixed issue, a fixed issue, or a fix.
>
> OK to push with these changes.

[-- Attachment #2: 0001-gnu-Add-ninja.patch --]
[-- Type: text/x-patch, Size: 4821 bytes --]

From 712f161a358a4bbcf005c5875e6d91eff21bdef4 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?=E5=AE=8B=E6=96=87=E6=AD=A6?= <iyzsong@gmail.com>
Date: Fri, 9 Jan 2015 21:37:30 +0800
Subject: [PATCH] gnu: Add ninja.

* gnu/packages/ninja.scm: New file.
* gnu-system.am (GNU_SYSTEM_MODULES): Add it.
---
 gnu-system.am          |  1 +
 gnu/packages/ninja.scm | 87 ++++++++++++++++++++++++++++++++++++++++++++++++++
 2 files changed, 88 insertions(+)
 create mode 100644 gnu/packages/ninja.scm

diff --git a/gnu-system.am b/gnu-system.am
index 4086067..39459b8 100644
--- a/gnu-system.am
+++ b/gnu-system.am
@@ -191,6 +191,7 @@ GNU_SYSTEM_MODULES =				\
   gnu/packages/ncurses.scm			\
   gnu/packages/netpbm.scm			\
   gnu/packages/nettle.scm			\
+  gnu/packages/ninja.scm			\
   gnu/packages/node.scm				\
   gnu/packages/noweb.scm			\
   gnu/packages/ntp.scm				\
diff --git a/gnu/packages/ninja.scm b/gnu/packages/ninja.scm
new file mode 100644
index 0000000..fe3f955
--- /dev/null
+++ b/gnu/packages/ninja.scm
@@ -0,0 +1,87 @@
+;;; GNU Guix --- Functional package management for GNU
+;;; Copyright © 2015 Sou Bunnbu <iyzsong@gmail.com>
+;;;
+;;; This file is part of GNU Guix.
+;;;
+;;; GNU Guix is free software; you can redistribute it and/or modify it
+;;; under the terms of the GNU General Public License as published by
+;;; the Free Software Foundation; either version 3 of the License, or (at
+;;; your option) any later version.
+;;;
+;;; GNU Guix is distributed in the hope that it will be useful, but
+;;; WITHOUT ANY WARRANTY; without even the implied warranty of
+;;; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+;;; GNU General Public License for more details.
+;;;
+;;; You should have received a copy of the GNU General Public License
+;;; along with GNU Guix.  If not, see <http://www.gnu.org/licenses/>.
+
+(define-module (gnu packages ninja)
+  #:use-module ((guix licenses) #:select (asl2.0))
+  #:use-module (guix packages)
+  #:use-module (guix download)
+  #:use-module (guix build-system gnu)
+  #:use-module (gnu packages)
+  #:use-module (gnu packages python))
+
+(define-public ninja
+  (package
+    (name "ninja")
+    (version "1.5.3")
+    (source (origin
+              (method url-fetch)
+              (uri (string-append "https://github.com/martine/ninja/"
+                                  "archive/v" version ".tar.gz"))
+              (sha256
+               (base32
+                "1h3yfwcfl61v493vna6jia2fizh8rpig7qw2504cvkr6gid3p5bw"))))
+    (build-system gnu-build-system)
+    (arguments
+     '(#:phases
+       (alist-replace
+        'configure
+        (lambda _
+          (substitute* "src/subprocess-posix.cc"
+            (("/bin/sh") (which "sh"))))
+        (alist-replace
+         'build
+         (lambda _
+           (zero? (system* "./configure.py" "--bootstrap")))
+         (alist-replace
+          'check
+          (lambda _
+            (and (zero? (system* "./configure.py"))
+                 (zero? (system* "./ninja" "ninja_test"))
+                 ;; SubprocessTest.SetWithLots fails with:
+                 ;;   Raise [ulimit -n] well above 1025 to make this test go.
+                 ;; Skip it.
+                 ;;
+                 ;; SubprocessTest.InterruptChild fails when using 'system*':
+                 ;;   *** Failure in src/subprocess_test.cc:83
+                 ;;   ExitInterrupted == subproc->Finish()
+                 ;; Pass it by using 'system' instead of 'system*'.
+                 (zero? (system (string-append
+                                 "./ninja_test "
+                                 "--gtest_filter="
+                                 "-SubprocessTest.SetWithLots")))))
+          (alist-replace
+           'install
+           (lambda* (#:key outputs #:allow-other-keys)
+             (let* ((out (assoc-ref outputs "out"))
+                    (bin (string-append out "/bin"))
+                    (doc (string-append out "/share/doc/ninja")))
+               (mkdir-p bin)
+               (copy-file "ninja" (string-append bin "/ninja"))
+               (mkdir-p doc)
+               (copy-file "doc/manual.asciidoc"
+                          (string-append doc "/manual.asciidoc"))))
+           %standard-phases))))))
+    (native-inputs `(("python" ,python-2)))
+    (home-page "http://martine.github.io/ninja/")
+    (synopsis "Small build system")
+    (description
+     "Ninja is a small build system with a focus on speed.  It differs from
+other build systems in two major respects: it is designed to have its input
+files generated by a higher-level build system, and it is designed to run
+builds as fast as possible.")
+    (license asl2.0)))
-- 
2.1.2


[-- Attachment #3: Type: text/plain, Size: 41 bytes --]


Is this OK?
>
> Thanks,
> Ludo’.

Re: [PATCH] gnu: Add ninja.

[Ludovic Courtès]

宋文武 <iyzsong@gmail.com> skribis:

> From 712f161a358a4bbcf005c5875e6d91eff21bdef4 Mon Sep 17 00:00:00 2001
> From: =?UTF-8?q?=E5=AE=8B=E6=96=87=E6=AD=A6?= <iyzsong@gmail.com>
> Date: Fri, 9 Jan 2015 21:37:30 +0800
> Subject: [PATCH] gnu: Add ninja.
>
> * gnu/packages/ninja.scm: New file.
> * gnu-system.am (GNU_SYSTEM_MODULES): Add it.

Yes, OK to push, thanks!

Ludo’.

Re: [PATCH] gnu: Add ninja.

[宋文武]

Ludovic Courtès <ludo@gnu.org> writes:

> 宋文武 <iyzsong@gmail.com> skribis:
>
>> From 712f161a358a4bbcf005c5875e6d91eff21bdef4 Mon Sep 17 00:00:00 2001
>> From: =?UTF-8?q?=E5=AE=8B=E6=96=87=E6=AD=A6?= <iyzsong@gmail.com>
>> Date: Fri, 9 Jan 2015 21:37:30 +0800
>> Subject: [PATCH] gnu: Add ninja.
>>
>> * gnu/packages/ninja.scm: New file.
>> * gnu-system.am (GNU_SYSTEM_MODULES): Add it.
>
> Yes, OK to push, thanks!
Even with 'system', the test InterruptChild failed on hydra for x86_64,
but it passed for i686, mips64el and my x86_64 locally by 'guix build'.

Perhaps it's a random issue?
Anyway, pushed. If it still fail, I think we have to skip it.

Thanks for review!
>
> Ludo’.

Re: [PATCH] gnu: Add ninja.

[Ludovic Courtès]

宋文武 <iyzsong@gmail.com> skribis:

> Ludovic Courtès <ludo@gnu.org> writes:
>
>> 宋文武 <iyzsong@gmail.com> skribis:
>>
>>> From 712f161a358a4bbcf005c5875e6d91eff21bdef4 Mon Sep 17 00:00:00 2001
>>> From: =?UTF-8?q?=E5=AE=8B=E6=96=87=E6=AD=A6?= <iyzsong@gmail.com>
>>> Date: Fri, 9 Jan 2015 21:37:30 +0800
>>> Subject: [PATCH] gnu: Add ninja.
>>>
>>> * gnu/packages/ninja.scm: New file.
>>> * gnu-system.am (GNU_SYSTEM_MODULES): Add it.
>>
>> Yes, OK to push, thanks!
> Even with 'system', the test InterruptChild failed on hydra for x86_64,
> but it passed for i686, mips64el and my x86_64 locally by 'guix build'.

I don’t see any failures actually:

  http://hydra.gnu.org/job/gnu/master/ninja-1.5.3.i686-linux
  http://hydra.gnu.org/job/gnu/master/ninja-1.5.3.mips64el-linux
  http://hydra.gnu.org/job/gnu/master/ninja-1.5.3.x86_64-linux

Are you sure it failed?

> Perhaps it's a random issue?

It looks like it.  It builds fine on my machine (x86_64).

I’ve looked at the test in question and it’s not obvious what could go
wrong:

--8<---------------cut here---------------start------------->8---
TEST_F(SubprocessTest, InterruptChild) {
  Subprocess* subproc = subprocs_.Add("kill -INT $$");
  ASSERT_NE((Subprocess *) 0, subproc);

  while (!subproc->Done()) {
    subprocs_.DoWork();
  }

  EXPECT_EQ(ExitInterrupted, subproc->Finish());
}
--8<---------------cut here---------------end--------------->8---

I think the idea is that ‘subproc_.Add’ launches a shell that gets
interrupted by the ‘kill’ subprocess, which ‘Finish’ must reflect.

Thanks,
Ludo’.

Re: [PATCH] gnu: Add ninja.

[Mark H Weaver]

ludo@gnu.org (Ludovic Courtès) writes:

> 宋文武 <iyzsong@gmail.com> skribis:
>
>> * gnu/packages/ninja.scm: New file.
>> * gnu-system.am (GNU_SYSTEM_MODULES): Add it.
>
> [...]
>
>> +          'check
>> +          (lambda _
>> +            (and (zero? (system "./configure.py"))
>> +                 (zero? (system "./ninja ninja_test"))
>> +                 ;; SubprocessTest.InterruptChild fail when using 'system*'.
>> +                 ;; SubprocessTest.SetWithLots was skipped.
>> +                 ;; XXX: Raise [ulimit -n] well above 1025 to make this test go.
>
> Does it mean that the test is currently failing?

This test failed during the core-updates x86_64 build on Hydra.  We get:

  [166/183] SubprocessTest.InterruptChild
  *** Failure in src/subprocess_test.cc:83
  ExitInterrupted == subproc->Finish()

In the hopes that it might be a transient failure, I restarted the
build, but the second time it failed in the same place.  See:

  http://hydra.gnu.org/build/201077/nixlog/1/raw
  http://hydra.gnu.org/build/201077/nixlog/2/raw

It seems that using 'system' instead of 'system*' wasn't enough to make
it work reliably.  Can you take another look?

Thanks for your contributions!

      Mark

Re: [PATCH] gnu: Add ninja.

[宋文武]

Mark H Weaver <mhw@netris.org> writes:

> ludo@gnu.org (Ludovic Courtès) writes:
>
>> 宋文武 <iyzsong@gmail.com> skribis:
>>
>>> * gnu/packages/ninja.scm: New file.
>>> * gnu-system.am (GNU_SYSTEM_MODULES): Add it.
>>
>> [...]
>>
>>> +          'check
>>> +          (lambda _
>>> +            (and (zero? (system "./configure.py"))
>>> +                 (zero? (system "./ninja ninja_test"))
>>> +                 ;; SubprocessTest.InterruptChild fail when using 'system*'.
>>> +                 ;; SubprocessTest.SetWithLots was skipped.
>>> +                 ;; XXX: Raise [ulimit -n] well above 1025 to make this test go.
>>
>> Does it mean that the test is currently failing?
>
> This test failed during the core-updates x86_64 build on Hydra.  We get:
>
>   [166/183] SubprocessTest.InterruptChild
>   *** Failure in src/subprocess_test.cc:83
>   ExitInterrupted == subproc->Finish()
>
> In the hopes that it might be a transient failure, I restarted the
> build, but the second time it failed in the same place.  See:
>
>   http://hydra.gnu.org/build/201077/nixlog/1/raw
>   http://hydra.gnu.org/build/201077/nixlog/2/raw
>
> It seems that using 'system' instead of 'system*' wasn't enough to make
> it work reliably.  Can you take another look?
OK, I pushed a patch to disable the test.
>
> Thanks for your contributions!
>
>       Mark