[PATCH] gnu: Add di.

[PATCH] gnu: Add di.

[John J Foerch]

[-- Attachment #1: Type: text/plain, Size: 122 bytes --]

Hello,

Here is a package for the program 'di' (df alternative).  Let me know if
anything is amiss, cheers!

John Foerch


[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #2: package definition for di --]
[-- Type: text/x-diff, Size: 1846 bytes --]

From 4ef5a6edba2e10a59865a6aba134d2f0532c7503 Mon Sep 17 00:00:00 2001
From: "John J. Foerch" <jjfoerch@earthlink.net>
Date: Sat, 2 Jul 2016 15:08:30 -0400
Subject: [PATCH] gnu: Add di.

* gnu/packages/admin.scm (di): New variable.
---
 gnu/packages/admin.scm | 31 +++++++++++++++++++++++++++++++
 1 file changed, 31 insertions(+)

diff --git a/gnu/packages/admin.scm b/gnu/packages/admin.scm
index 415a35a..f128f68 100644
--- a/gnu/packages/admin.scm
+++ b/gnu/packages/admin.scm
@@ -1708,3 +1708,34 @@ throughput (in the same interval).")
      "The Fuck tries to match a rule for a previous, mistyped command, creates
 a new command using the matched rule, and runs it.")
     (license license:x11)))
+
+(define-public di
+  (package
+    (name "di")
+    (version "4.42")
+    (source
+     (origin
+       (method url-fetch)
+       (uri (string-append "https://gentoo.com/di/di-" version ".tar.gz"))
+       (sha256
+        (base32 "1i6m9zdnidn8268q1lz9fd8payk7s4pgwh5zlam9rr4dy6h6a67n"))))
+    (build-system gnu-build-system)
+    (arguments
+     `(#:phases
+       (modify-phases %standard-phases
+         (delete 'configure)
+         (delete 'check)
+         (add-before 'build 'setup-environment
+           (lambda* (#:key outputs #:allow-other-keys)
+             (setenv "CC" "gcc")
+             (setenv "prefix" (assoc-ref outputs "out"))
+             #t)))
+       #:make-flags (list "-e")))
+    (home-page "https://www.gentoo.com/di/")
+    (synopsis "Advanced df like disk information utility")
+    (description
+     "'di' is a disk information utility, displaying everything
+(and more) that your 'df' command does.  It features the ability to display
+your disk usage in whatever format you prefer.  It is designed to be highly
+portable.  Great for heterogenous networks.")
+    (license license:zlib)))
-- 
2.9.0

Re: [PATCH] gnu: Add di.

[Leo Famulari]

On Sat, Jul 02, 2016 at 03:13:52PM -0400, John J Foerch wrote:
> * gnu/packages/admin.scm (di): New variable.

Thanks!

> +    (arguments
> +     `(#:phases
> +       (modify-phases %standard-phases
> +         (delete 'configure)
> +         (delete 'check)

Did you try setting `#:test-target "test"` in the arguments? The
Makefile has a test target.

> +         (add-before 'build 'setup-environment
> +           (lambda* (#:key outputs #:allow-other-keys)
> +             (setenv "CC" "gcc")
> +             (setenv "prefix" (assoc-ref outputs "out"))
> +             #t)))

Can you move these variables into #:make-flags?

> +       #:make-flags (list "-e")))

> +    (description
> +     "'di' is a disk information utility, displaying everything
> +(and more) that your 'df' command does.  It features the ability to display

How about wrapping 'df' in texinfo like this? @code{df}

Re: [PATCH] gnu: Add di.

[John J Foerch]

Leo Famulari <leo@famulari.name> writes:

> On Sat, Jul 02, 2016 at 03:13:52PM -0400, John J Foerch wrote:
>> * gnu/packages/admin.scm (di): New variable.
>
> Thanks!
>
>> +    (arguments
>> +     `(#:phases
>> +       (modify-phases %standard-phases
>> +         (delete 'configure)
>> +         (delete 'check)
>
> Did you try setting `#:test-target "test"` in the arguments? The
> Makefile has a test target.
>

I didn't know about that.  I'll try it out.

>> +         (add-before 'build 'setup-environment
>> +           (lambda* (#:key outputs #:allow-other-keys)
>> +             (setenv "CC" "gcc")
>> +             (setenv "prefix" (assoc-ref outputs "out"))
>> +             #t)))
>
> Can you move these variables into #:make-flags?
>
>> +       #:make-flags (list "-e")))

I don't think so.  The main Makefile calls make in the 'C' directory,
which has its own Makefile, so for the variables to be visible in that
second make process, they need to be environment variables.

>
>> +    (description
>> +     "'di' is a disk information utility, displaying everything
>> +(and more) that your 'df' command does.  It features the ability to display
>
> How about wrapping 'df' in texinfo like this? @code{df}

Okay, will do.

Thank you,

--
John Foerch

Re: [PATCH] gnu: Add di.

[John J Foerch]

John J Foerch <jjfoerch@earthlink.net> writes:

> Leo Famulari <leo@famulari.name> writes:
>
>> On Sat, Jul 02, 2016 at 03:13:52PM -0400, John J Foerch wrote:
>>> * gnu/packages/admin.scm (di): New variable.
>>
>> Thanks!
>>
>>> +    (arguments
>>> +     `(#:phases
>>> +       (modify-phases %standard-phases
>>> +         (delete 'configure)
>>> +         (delete 'check)
>>
>> Did you try setting `#:test-target "test"` in the arguments? The
>> Makefile has a test target.
>>
>
> I didn't know about that.  I'll try it out.
>

A test relating to an included perl module Filesys::di fails.  I hadn't
intended to include that module in the package, as all I'm interested in
is the 'di' binary.  Given that, I would suggest that we bypass the
tests.

--
John Foerch

Re: [PATCH] gnu: Add di.

[Alex Kost]

John J Foerch (2016-07-04 06:47 +0300) wrote:

> John J Foerch <jjfoerch@earthlink.net> writes:
>
>> Leo Famulari <leo@famulari.name> writes:
>>
>>> On Sat, Jul 02, 2016 at 03:13:52PM -0400, John J Foerch wrote:
>>>> * gnu/packages/admin.scm (di): New variable.
>>>
>>> Thanks!
>>>
>>>> +    (arguments
>>>> +     `(#:phases
>>>> +       (modify-phases %standard-phases
>>>> +         (delete 'configure)
>>>> +         (delete 'check)
>>>
>>> Did you try setting `#:test-target "test"` in the arguments? The
>>> Makefile has a test target.
>>>
>>
>> I didn't know about that.  I'll try it out.
>>
> A test relating to an included perl module Filesys::di fails.  I hadn't
> intended to include that module in the package, as all I'm interested in
> is the 'di' binary.  Given that, I would suggest that we bypass the
> tests.

As for me, I would also disable the tests.  All these handmade Makefiles
in all sub-directories are a nightmare.  I think trying to figure out
what is needed to run the tests doesn't worth it.

Also don't forget to add a copyright line for yourself in the beginning
of "admin.scm".

-- 
Alex

Re: [PATCH] gnu: Add di.

[John J Foerch]

[-- Attachment #1: Type: text/plain, Size: 1216 bytes --]

Alex Kost <alezost@gmail.com> writes:

> John J Foerch (2016-07-04 06:47 +0300) wrote:
>
>> John J Foerch <jjfoerch@earthlink.net> writes:
>>
>>> Leo Famulari <leo@famulari.name> writes:
>>>
>>>> On Sat, Jul 02, 2016 at 03:13:52PM -0400, John J Foerch wrote:
>>>>> * gnu/packages/admin.scm (di): New variable.
>>>>
>>>> Thanks!
>>>>
>>>>> +    (arguments
>>>>> +     `(#:phases
>>>>> +       (modify-phases %standard-phases
>>>>> +         (delete 'configure)
>>>>> +         (delete 'check)
>>>>
>>>> Did you try setting `#:test-target "test"` in the arguments? The
>>>> Makefile has a test target.
>>>>
>>>
>>> I didn't know about that.  I'll try it out.
>>>
>> A test relating to an included perl module Filesys::di fails.  I hadn't
>> intended to include that module in the package, as all I'm interested in
>> is the 'di' binary.  Given that, I would suggest that we bypass the
>> tests.
>
> As for me, I would also disable the tests.  All these handmade Makefiles
> in all sub-directories are a nightmare.  I think trying to figure out
> what is needed to run the tests doesn't worth it.
>
> Also don't forget to add a copyright line for yourself in the beginning
> of "admin.scm".

Here is an updated patch:


[-- Attachment #2: gnu: Add di. --]
[-- Type: text/x-diff, Size: 2160 bytes --]

From 05098e58a85149c96597bbacec5f0f9ef200d299 Mon Sep 17 00:00:00 2001
From: "John J. Foerch" <jjfoerch@earthlink.net>
Date: Sat, 2 Jul 2016 15:08:30 -0400
Subject: [PATCH] gnu: Add di.

* gnu/packages/admin.scm (di): New variable.
---
 gnu/packages/admin.scm | 32 ++++++++++++++++++++++++++++++++
 1 file changed, 32 insertions(+)

diff --git a/gnu/packages/admin.scm b/gnu/packages/admin.scm
index 415a35a..0fd2c87 100644
--- a/gnu/packages/admin.scm
+++ b/gnu/packages/admin.scm
@@ -11,6 +11,7 @@
 ;;; Copyright © 2016 Ricardo Wurmus <rekado@elephly.net>
 ;;; Copyright © 2016 Efraim Flashner <efraim@flashner.co.il>
 ;;; Copyright © 2016 Peter Feigl <peter.feigl@nexoid.at>
+;;; Copyright © 2016 John J. Foerch <jjfoerch@earthlink.net>
 ;;;
 ;;; This file is part of GNU Guix.
 ;;;
@@ -1708,3 +1709,34 @@ throughput (in the same interval).")
      "The Fuck tries to match a rule for a previous, mistyped command, creates
 a new command using the matched rule, and runs it.")
     (license license:x11)))
+
+(define-public di
+  (package
+    (name "di")
+    (version "4.42")
+    (source
+     (origin
+       (method url-fetch)
+       (uri (string-append "https://gentoo.com/di/di-" version ".tar.gz"))
+       (sha256
+        (base32 "1i6m9zdnidn8268q1lz9fd8payk7s4pgwh5zlam9rr4dy6h6a67n"))))
+    (build-system gnu-build-system)
+    (arguments
+     `(#:phases
+       (modify-phases %standard-phases
+         (delete 'configure)
+         (delete 'check)
+         (add-before 'build 'setup-environment
+           (lambda* (#:key outputs #:allow-other-keys)
+             (setenv "CC" "gcc")
+             (setenv "prefix" (assoc-ref outputs "out"))
+             #t)))
+       #:make-flags (list "-e")))
+    (home-page "https://www.gentoo.com/di/")
+    (synopsis "Advanced df like disk information utility")
+    (description
+     "'di' is a disk information utility, displaying everything
+(and more) that your @code{df} command does.  It features the ability to
+display your disk usage in whatever format you prefer.  It is designed to be
+highly portable.  Great for heterogenous networks.")
+    (license license:zlib)))
-- 
2.9.0

Re: [PATCH] gnu: Add di.

[Leo Famulari]

On Mon, Jul 04, 2016 at 11:18:50AM -0400, John J Foerch wrote:
> * gnu/packages/admin.scm (di): New variable.

A few final nitpicks before others make a decision about the test suite
;)

> +    (arguments
> +     `(#:phases
> +       (modify-phases %standard-phases
> +         (delete 'configure)
> +         (delete 'check)

We prefer to use #:tests? #f instead of deleting the check phase.

> +         (add-before 'build 'setup-environment
> +           (lambda* (#:key outputs #:allow-other-keys)
> +             (setenv "CC" "gcc")
> +             (setenv "prefix" (assoc-ref outputs "out"))
> +             #t)))
> +       #:make-flags (list "-e")))

I would like for '-e' to be expanded to '--environment-overrides'. It
seems that Schemers abhor abbreviations.

Sorry for not mentioning these before. I can do them myself when pushing
if there are no other changes to be made.

Re: [PATCH] gnu: Add di.

[John J Foerch]

Leo Famulari <leo@famulari.name> writes:

> On Mon, Jul 04, 2016 at 11:18:50AM -0400, John J Foerch wrote:
>> * gnu/packages/admin.scm (di): New variable.
>
> A few final nitpicks before others make a decision about the test suite
> ;)
>
>> +    (arguments
>> +     `(#:phases
>> +       (modify-phases %standard-phases
>> +         (delete 'configure)
>> +         (delete 'check)
>
> We prefer to use #:tests? #f instead of deleting the check phase.
>
>> +         (add-before 'build 'setup-environment
>> +           (lambda* (#:key outputs #:allow-other-keys)
>> +             (setenv "CC" "gcc")
>> +             (setenv "prefix" (assoc-ref outputs "out"))
>> +             #t)))
>> +       #:make-flags (list "-e")))
>
> I would like for '-e' to be expanded to '--environment-overrides'. It
> seems that Schemers abhor abbreviations.
>
> Sorry for not mentioning these before. I can do them myself when pushing
> if there are no other changes to be made.

Sounds good.  Thanks.

--
John Foerch

Re: [PATCH] gnu: Add di.

[John J Foerch]

John J Foerch <jjfoerch@earthlink.net> writes:

> Leo Famulari <leo@famulari.name> writes:
>
>> On Mon, Jul 04, 2016 at 11:18:50AM -0400, John J Foerch wrote:
>>> * gnu/packages/admin.scm (di): New variable.
>>
>> A few final nitpicks before others make a decision about the test suite
>> ;)
>>
>>> +    (arguments
>>> +     `(#:phases
>>> +       (modify-phases %standard-phases
>>> +         (delete 'configure)
>>> +         (delete 'check)
>>
>> We prefer to use #:tests? #f instead of deleting the check phase.
>>
>>> +         (add-before 'build 'setup-environment
>>> +           (lambda* (#:key outputs #:allow-other-keys)
>>> +             (setenv "CC" "gcc")
>>> +             (setenv "prefix" (assoc-ref outputs "out"))
>>> +             #t)))
>>> +       #:make-flags (list "-e")))
>>
>> I would like for '-e' to be expanded to '--environment-overrides'. It
>> seems that Schemers abhor abbreviations.
>>
>> Sorry for not mentioning these before. I can do them myself when pushing
>> if there are no other changes to be made.
>
> Sounds good.  Thanks.

Any remaining blocking issues on this?

--
John Foerch

Re: [PATCH] gnu: Add di.

[Leo Famulari]

On Thu, Jul 07, 2016 at 12:30:11PM -0400, John J Foerch wrote:
> Any remaining blocking issues on this?

Not really... I've been struggling to write the comment that justifies
skipping the tests :)

Any suggestions?

Re: [PATCH] gnu: Add di.

[Alex Kost]

Leo Famulari (2016-07-07 20:37 +0300) wrote:

> On Thu, Jul 07, 2016 at 12:30:11PM -0400, John J Foerch wrote:
>> Any remaining blocking issues on this?
>
> Not really... I've been struggling to write the comment that justifies
> skipping the tests :)

Nice comment after all :-)   ("Obscure test failures")

-- 
Alex

Re: [PATCH] gnu: Add di.

[Leo Famulari]

On Thu, Jul 07, 2016 at 12:30:11PM -0400, John J Foerch wrote:
> Any remaining blocking issues on this?

Pushed as 614023981!

Re: [PATCH] gnu: Add di.

[John J Foerch]

Leo Famulari <leo@famulari.name> writes:

> On Thu, Jul 07, 2016 at 12:30:11PM -0400, John J Foerch wrote:
>> Any remaining blocking issues on this?
>
> Pushed as 614023981!

Many thanks.

--
John Foerch

Re: [PATCH] gnu: Add di.

[Leo Famulari]

On Mon, Jul 04, 2016 at 09:43:55AM +0300, Alex Kost wrote:
> John J Foerch (2016-07-04 06:47 +0300) wrote:
> > A test relating to an included perl module Filesys::di fails.  I hadn't
> > intended to include that module in the package, as all I'm interested in
> > is the 'di' binary.  Given that, I would suggest that we bypass the
> > tests.
> 
> As for me, I would also disable the tests.  All these handmade Makefiles
> in all sub-directories are a nightmare.  I think trying to figure out
> what is needed to run the tests doesn't worth it.

I agree that hand-carved Makefiles are difficult to work with.

The failing Perl-related test runs successfully if Perl is added to the
environment. However, there is another failure later on:

make[1]: Leaving directory '/tmp/guix-build-di-4.42.drv-0/di-4.42/C'
make -e tests.done
make[1]: Entering directory '/tmp/guix-build-di-4.42.drv-0/di-4.42'
## running tests
CC=gcc DC="skip"  \
	./mkconfig/runtests.sh ./tests.d
mkconfig version 1.27
## locating valid shells
   /gnu/store/b1yqjimbdh5bf9jnizd4h7yf110744j2-bash-4.3.42/bin/bash [bash4 4.3.42(1)-release] (ok)
   /gnu/store/b1yqjimbdh5bf9jnizd4h7yf110744j2-bash-4.3.42/bin/sh [sh-bash4 4.3.42(1)-release] (ok)
buildsh ... build w/mkconfig.sh ... bash4 C* sh-bash4 C* ... failed
## stopping tests due to failures in ./tests.d/ pass 1
1 tests 1 failures
Makefile:116: recipe for target 'tests.done' failed
make[1]: *** [tests.done] Error 1
make[1]: Leaving directory '/tmp/guix-build-di-4.42.drv-0/di-4.42'
Makefile:112: recipe for target 'test' failed
make: *** [test] Error 2
phase `check' failed after 64.5 seconds

I skimmed 'mkconfig/runtests.sh', but I don't understand it well enough
to debug it.

If the source code is still maintained, I bet the authors will be
willing to help. They express interest in testing and portability on the
web page.

But if everyone else prefers to skip the tests, I won't object :)

Re: [PATCH] gnu: Add di.

[John J Foerch]

Leo Famulari <leo@famulari.name> writes:

> On Mon, Jul 04, 2016 at 09:43:55AM +0300, Alex Kost wrote:
>> John J Foerch (2016-07-04 06:47 +0300) wrote:
>> > A test relating to an included perl module Filesys::di fails.  I hadn't
>> > intended to include that module in the package, as all I'm interested in
>> > is the 'di' binary.  Given that, I would suggest that we bypass the
>> > tests.
>> 
>> As for me, I would also disable the tests.  All these handmade Makefiles
>> in all sub-directories are a nightmare.  I think trying to figure out
>> what is needed to run the tests doesn't worth it.
>
> I agree that hand-carved Makefiles are difficult to work with.
>
> The failing Perl-related test runs successfully if Perl is added to the
> environment. However, there is another failure later on:
>

I would favor leaving perl out of the package for the di binary
entirely, and if people want the perl module, use another package for
that.

-- 
John Foerch