low-hanging SRFI fruit (SRFI-98)

low-hanging SRFI fruit (SRFI-98)

[Julian Graham]

[-- Attachment #1: Type: text/plain, Size: 313 bytes --]

Hi Guilers,

While looking over the prerequisites for the proposed SRFI-100, I came
across SRFI-98, which seemed like pretty easy win for Guile.  So I did
a quick implementation on the subway -- find it attached.  I'll write
up unit tests and docs if people think it's suitable for integration.


Regards,
Julian

[-- Attachment #2: srfi-98.scm --]
[-- Type: text/x-scheme, Size: 1614 bytes --]

;;; srfi-98.scm --- An interface to access environment variables

;; Copyright (C) 2009 Free Software Foundation, Inc.
;;
;; This library is free software; you can redistribute it and/or
;; modify it under the terms of the GNU Lesser General Public
;; License as published by the Free Software Foundation; either
;; version 2.1 of the License, or (at your option) any later version.
;; 
;; This library 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
;; Lesser General Public License for more details.
;; 
;; You should have received a copy of the GNU Lesser General Public
;; License along with this library; if not, write to the Free Software
;; Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA

;;; Author: Julian Graham <julian.graham@aya.yale.edu>
;;; Date: 2009-05-26

;;; Commentary:

;; This is an implementation of SRFI-98 (An interface to access environment 
;; variables).
;;
;; This module is fully documented in the Guile Reference Manual.

;;; Code:

(define-module (srfi srfi-98)
  :use-module (srfi srfi-1)
  :export (get-environment-variable
	   get-environment-variables))

(cond-expand-provide (current-module) '(srfi-98))

(define get-environment-variable getenv)
(define (get-environment-variables)
  (define (string->alist-entry str)
    (let ((pvt (string-index str #\=))
	  (len (string-length str)))
      (and pvt (cons (substring str 0 pvt) (substring str (+ pvt 1) len)))))
  (filter-map string->alist-entry (environ)))

Re: low-hanging SRFI fruit (SRFI-98)

[Neil Jerram]

Julian Graham <joolean@gmail.com> writes:

> Hi Guilers,
>
> While looking over the prerequisites for the proposed SRFI-100, I came
> across SRFI-98, which seemed like pretty easy win for Guile.  So I did
> a quick implementation on the subway -- find it attached.  I'll write
> up unit tests and docs if people think it's suitable for integration.

Looks good to me, please go ahead.

    Neil

Re: low-hanging SRFI fruit (SRFI-98)

[Julian Graham]

[-- Attachment #1: Type: text/plain, Size: 198 bytes --]

> Looks good to me, please go ahead.

Find attached a patch that includes the implementation previously
attached, plus unit tests and additions to the NEWS and modules.texi
files.


Regards,
Julian

[-- Attachment #2: 0001-Implementation-of-SRFI-98-An-interface-to-environme.patch --]
[-- Type: text/x-patch, Size: 5866 bytes --]

From a1e4bcb30b80199de6f2cd620d950d04e0734592 Mon Sep 17 00:00:00 2001
From: Julian Graham <julian@transmetropolitan.(none)>
Date: Thu, 28 May 2009 18:15:05 -0400
Subject: [PATCH] Implementation of SRFI-98 (An interface to environment variables)

---
 NEWS                          |    1 +
 doc/ref/srfi-modules.texi     |   19 +++++++++++++++++
 module/srfi/srfi-98.scm       |   44 +++++++++++++++++++++++++++++++++++++++++
 test-suite/tests/srfi-98.test |   38 +++++++++++++++++++++++++++++++++++
 4 files changed, 102 insertions(+), 0 deletions(-)
 create mode 100644 module/srfi/srfi-98.scm
 create mode 100644 test-suite/tests/srfi-98.test

diff --git a/NEWS b/NEWS
index 1785fe8..94d6be7 100644
--- a/NEWS
+++ b/NEWS
@@ -10,6 +10,7 @@ Changes in 1.9.0:
 * New modules (see the manual for details)
 
 ** `(srfi srfi-18)', multithreading support
+** `(srfi srfi-98)', an interface to environment variables
 ** The `(ice-9 i18n)' module provides internationalization support
 
 * Changes to the distribution
diff --git a/doc/ref/srfi-modules.texi b/doc/ref/srfi-modules.texi
index 1fa50b2..638aa46 100644
--- a/doc/ref/srfi-modules.texi
+++ b/doc/ref/srfi-modules.texi
@@ -47,6 +47,7 @@ get the relevant SRFI documents from the SRFI home page
 * SRFI-61::                     A more general `cond' clause
 * SRFI-69::                     Basic hash tables.
 * SRFI-88::                     Keyword objects.
+* SRFI-98::                     An interface to access environment variables.
 @end menu
 
 
@@ -3608,6 +3609,24 @@ Return the keyword object whose name is @var{str}.
 @end example
 @end deffn
 
+@node SRFI-98
+@subsection SRFI-98 An interface to access environment variables.
+@cindex SRFI-98
+@cindex environment variables
+
+This is a portable wrapper around Guile's built-in support for 
+interacting with the current environment, @xref{Runtime Environment}.
+
+@deffn {Scheme Procedure} get-environment-variable name
+Returns the value of the named environment variable as a string, or 
+@code{#f} if the named environment variable is not found.  This is 
+equivalent to @code{(getenv name)}.
+@end deffn
+
+@deffn {Scheme Procedure} get-environment-variables
+Returns names and values of all the environment variables as an
+association list.
+@end deffn
 
 @c srfi-modules.texi ends here
 
diff --git a/module/srfi/srfi-98.scm b/module/srfi/srfi-98.scm
new file mode 100644
index 0000000..924a205
--- /dev/null
+++ b/module/srfi/srfi-98.scm
@@ -0,0 +1,44 @@
+;;; srfi-98.scm --- An interface to access environment variables
+
+;; Copyright (C) 2009 Free Software Foundation, Inc.
+;;
+;; This library is free software; you can redistribute it and/or
+;; modify it under the terms of the GNU Lesser General Public
+;; License as published by the Free Software Foundation; either
+;; version 2.1 of the License, or (at your option) any later version.
+;; 
+;; This library 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
+;; Lesser General Public License for more details.
+;; 
+;; You should have received a copy of the GNU Lesser General Public
+;; License along with this library; if not, write to the Free Software
+;; Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+
+;;; Author: Julian Graham <julian.graham@aya.yale.edu>
+;;; Date: 2009-05-26
+
+;;; Commentary:
+
+;; This is an implementation of SRFI-98 (An interface to access environment 
+;; variables).
+;;
+;; This module is fully documented in the Guile Reference Manual.
+
+;;; Code:
+
+(define-module (srfi srfi-98)
+  :use-module (srfi srfi-1)
+  :export (get-environment-variable
+	   get-environment-variables))
+
+(cond-expand-provide (current-module) '(srfi-98))
+
+(define get-environment-variable getenv)
+(define (get-environment-variables)
+  (define (string->alist-entry str)
+    (let ((pvt (string-index str #\=))
+	  (len (string-length str)))
+      (and pvt (cons (substring str 0 pvt) (substring str (+ pvt 1) len)))))
+  (filter-map string->alist-entry (environ)))
diff --git a/test-suite/tests/srfi-98.test b/test-suite/tests/srfi-98.test
new file mode 100644
index 0000000..3fbb1ef
--- /dev/null
+++ b/test-suite/tests/srfi-98.test
@@ -0,0 +1,38 @@
+;;;; srfi-98.test --- Test suite for Guile's SRFI-98 functions. -*- scheme -*-
+;;;;
+;;;; Copyright 2009 Free Software Foundation, Inc.
+;;;;
+;;;; This program 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 2, or (at your option)
+;;;; any later version.
+;;;;
+;;;; This program 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 this software; see the file COPYING.  If not, write to
+;;;; the Free Software Foundation, Inc., 51 Franklin Street, Fifth Floor,
+;;;; Boston, MA 02110-1301 USA
+
+(define-module (test-srfi-98)
+  #:use-module (srfi srfi-98)
+  #:use-module (test-suite lib))
+
+(with-test-prefix "get-environment-variable"
+  (pass-if "get-environment-variable retrieves binding"
+    (putenv "foo=bar")
+    (equal? (get-environment-variable "foo") "bar"))
+
+  (pass-if "get-environment-variable #f on unbound name"
+    (unsetenv "foo")
+    (not (get-environment-variable "foo"))))      
+
+(with-test-prefix "get-environment-variables"
+
+  (pass-if "get-environment-variables contains binding"
+    (putenv "foo=bar")
+    (equal? (assoc-ref (get-environment-variables) "foo") "bar")))
+
-- 
1.6.0.4

Re: low-hanging SRFI fruit (SRFI-98)

[Neil Jerram]

Julian Graham <joolean@gmail.com> writes:

>> Looks good to me, please go ahead.
>
> Find attached a patch that includes the implementation previously
> attached, plus unit tests and additions to the NEWS and modules.texi
> files.
>
>
> Regards,
> Julian

Hi Julian,

Thanks for doing this.  The code and test all look fine to me; I just
have a few comments on the commit message and the documentation.

Am I correct in thinking that you can commit this yourself? - subject
to Savannah coming back up, of course.  If so, please do that after
incorporating the following points.

> From a1e4bcb30b80199de6f2cd620d950d04e0734592 Mon Sep 17 00:00:00 2001
> From: Julian Graham <julian@transmetropolitan.(none)>
> Date: Thu, 28 May 2009 18:15:05 -0400
> Subject: [PATCH] Implementation of SRFI-98 (An interface to environment variables)

When you commit, the commit message should include ChangeLog-style
entries for the changes.

> ---
>  NEWS                          |    1 +
>  doc/ref/srfi-modules.texi     |   19 +++++++++++++++++
>  module/srfi/srfi-98.scm       |   44 +++++++++++++++++++++++++++++++++++++++++
>  test-suite/tests/srfi-98.test |   38 +++++++++++++++++++++++++++++++++++
>  4 files changed, 102 insertions(+), 0 deletions(-)
>  create mode 100644 module/srfi/srfi-98.scm
>  create mode 100644 test-suite/tests/srfi-98.test
>
> diff --git a/NEWS b/NEWS
> index 1785fe8..94d6be7 100644
> --- a/NEWS
> +++ b/NEWS
> @@ -10,6 +10,7 @@ Changes in 1.9.0:
>  * New modules (see the manual for details)
>  
>  ** `(srfi srfi-18)', multithreading support
> +** `(srfi srfi-98)', an interface to environment variables
>  ** The `(ice-9 i18n)' module provides internationalization support
>  
>  * Changes to the distribution
> diff --git a/doc/ref/srfi-modules.texi b/doc/ref/srfi-modules.texi
> index 1fa50b2..638aa46 100644
> --- a/doc/ref/srfi-modules.texi
> +++ b/doc/ref/srfi-modules.texi
> @@ -47,6 +47,7 @@ get the relevant SRFI documents from the SRFI home page
>  * SRFI-61::                     A more general `cond' clause
>  * SRFI-69::                     Basic hash tables.
>  * SRFI-88::                     Keyword objects.
> +* SRFI-98::                     An interface to access environment variables.

That's not quite idiomatic English, is it?  I know it's the title of
the SRFI, but I think that `Accessing environment variables' would be
simpler and better.

>  @end menu
>  
>  
> @@ -3608,6 +3609,24 @@ Return the keyword object whose name is @var{str}.
>  @end example
>  @end deffn
>  
> +@node SRFI-98
> +@subsection SRFI-98 An interface to access environment variables.

Same again here.

> +@cindex SRFI-98
> +@cindex environment variables
> +
> +This is a portable wrapper around Guile's built-in support for 
> +interacting with the current environment, @xref{Runtime Environment}.
> +
> +@deffn {Scheme Procedure} get-environment-variable name
> +Returns the value of the named environment variable as a string, or 
> +@code{#f} if the named environment variable is not found.  This is 
> +equivalent to @code{(getenv name)}.

Looking at this from the point of view of a developer wanting to use
this procedure...  Should name be a symbol, or a string; or can it be
either?

> +@end deffn
> +
> +@deffn {Scheme Procedure} get-environment-variables
> +Returns names and values of all the environment variables as an
> +association list.

In which the keys and values are strings?  Or are the keys symbols?

Regards,
      Neil

Re: low-hanging SRFI fruit (SRFI-98)

[Julian Graham]

Hi Neil,

> Am I correct in thinking that you can commit this yourself? - subject
> to Savannah coming back up, of course.  If so, please do that after
> incorporating the following points.

Can I?  As far as I know I haven't been granted the appropriate
permissions, but I'll look into it.


> That's not quite idiomatic English, is it?  I know it's the title of
> the SRFI, but I think that `Accessing environment variables' would be
> simpler and better.

No, I suppose it isn't.  But my assumption was that for documentation
nodes for SRFI modules, we were going with the title of the SRFI --
and that's what SRFI-98 is called.  I'll update it.


Regards,
Julian

Re: low-hanging SRFI fruit (SRFI-98)

[Ludovic Courtès]

Hi Julian,

I concur with Neil's review.  One more thing:

Julian Graham <joolean@gmail.com> writes:

> --- a/NEWS
> +++ b/NEWS
> @@ -10,6 +10,7 @@ Changes in 1.9.0:
>  * New modules (see the manual for details)
>  
>  ** `(srfi srfi-18)', multithreading support
> +** `(srfi srfi-98)', an interface to environment variables
>  ** The `(ice-9 i18n)' module provides internationalization support

This change can go in 1.8, so you might want to modify the patch with
that line moved under "Changes in 1.8.7".  :-)

Thanks,
Ludo'.

Re: low-hanging SRFI fruit (SRFI-98)

[Neil Jerram]

Julian Graham <joolean@gmail.com> writes:

> Hi Neil,
>
>> Am I correct in thinking that you can commit this yourself? - subject
>> to Savannah coming back up, of course.  If so, please do that after
>> incorporating the following points.
>
> Can I?  As far as I know I haven't been granted the appropriate
> permissions, but I'll look into it.

I thought we organised that towards the end of your SRFI-18 work...
Anyway, once Savannah is alive again I'll check and `make it so' if it
isn't already the case.

>> That's not quite idiomatic English, is it?  I know it's the title of
>> the SRFI, but I think that `Accessing environment variables' would be
>> simpler and better.
>
> No, I suppose it isn't.  But my assumption was that for documentation
> nodes for SRFI modules, we were going with the title of the SRFI --
> and that's what SRFI-98 is called.  I'll update it.

Thanks.  I agree that there's an argument for keeping the original
title, but I think it's more important for us to make sure that our
manual reads well.

(Especially in the light of the next email that I'm about to write...)

Regards,
       Neil

Re: low-hanging SRFI fruit (SRFI-98)

[Neil Jerram]

Neil Jerram <neil@ossau.uklinux.net> writes:

> Julian Graham <joolean@gmail.com> writes:
>
>> Hi Ludovic,
>>
>>> You can't undo it.  :-)
>>
>> But surely the "owner" of the Savannah repository can remove an
>> offending commit from the history, no?
>
> I thought so, yes.  But I hesitate to contradict Ludo...
>
> Don't worry too much, this looks to me like something we can fix.
> I'll have time to investigate further later this evening, if no one
> gets there before me.

OK, I'm about to try sorting this out, which will change the last 2
commits of the master branch.  Please hold off doing any pulling or
pushing until I've done that and everything looks OK again.

Thanks,
    Neil

Re: low-hanging SRFI fruit (SRFI-98)

[Neil Jerram]

Neil Jerram <neil@ossau.uklinux.net> writes:

> OK, I'm about to try sorting this out, which will change the last 2
> commits of the master branch.  Please hold off doing any pulling or
> pushing until I've done that and everything looks OK again.

OK, I reconstructed what I think master should be (as of right now,
including Andy's commits this evening), and tried to `git push
--force' that, but it was rejected by the server:

neil@arudy:~/SW/Guile/git$ git push --force
Counting objects: 103, done.
Compressing objects: 100% (74/74), done.
Writing objects: 100% (76/76), 16.88 KiB, done.
Total 76 (delta 58), reused 4 (delta 2)
error: denying non-fast forward refs/heads/master (you should pull first)
To ssh://ossau@git.sv.gnu.org/srv/git/guile.git
 ! [remote rejected] master -> master (non-fast forward)
error: failed to push some refs to 'ssh://ossau@git.sv.gnu.org/srv/git/guile.git'

So to show my reconstruction, I pushed it as a new `njmaster' branch
instead.

I guess that the rejection above is caused either either by the repo's
denynonfastforward config being TRUE, or by a hook (such as the one
that sends changes to guile-commits) failing to handle the push.

I think it is worth sorting this out, so I'll ask the Savannah hackers
about doing whatever is needed so that the push can succeed.

(In the meantime, please feel free to add to either master or
njmaster.)

Regards,
        Neil

Re: low-hanging SRFI fruit (SRFI-98)

[Andy Wingo]

Hi Neil,

On Thu 04 Jun 2009 00:08, Neil Jerram <neil@ossau.uklinux.net> writes:

> Neil Jerram <neil@ossau.uklinux.net> writes:
>
>> Julian Graham <joolean@gmail.com> writes:
>>
>>> Hi Ludovic,
>>>
>>>> You can't undo it.  :-)
>>>
>>> But surely the "owner" of the Savannah repository can remove an
>>> offending commit from the history, no?
>>
>> I thought so, yes.  But I hesitate to contradict Ludo...
>>
>> Don't worry too much, this looks to me like something we can fix.
>> I'll have time to investigate further later this evening, if no one
>> gets there before me.
>
> OK, I'm about to try sorting this out, which will change the last 2
> commits of the master branch.  Please hold off doing any pulling or
> pushing until I've done that and everything looks OK again.

You should be able to remove the ref entirely with "git push :master"
then "git push master" -- as least the last time I mucked with these
things that's how it worked.

BUT... I wouldn't bother, really. In the meantime I committed a few
things to `master' this morning, so we'd need to rebase anew. The merge
commit has no practical implications, and indeed will happen from time
to time with distributed development. And now if we fudge the history,
it could be that people's git pull invocations stop running, like your
snapshot did.

So while if you really want to do this, we can, but my advice would be
to let the tangle drop off into history ;-)

Cheers,

Andy
-- 
http://wingolog.org/

Re: low-hanging SRFI fruit (SRFI-98)

[Neil Jerram]

Andy Wingo <wingo@pobox.com> writes:

> You should be able to remove the ref entirely with "git push :master"
> then "git push master" -- as least the last time I mucked with these
> things that's how it worked.

Ah cool, I'll try to remember that for the future.

> BUT... I wouldn't bother, really. In the meantime I committed a few
> things to `master' this morning, so we'd need to rebase anew. The merge
> commit has no practical implications, and indeed will happen from time
> to time with distributed development. And now if we fudge the history,
> it could be that people's git pull invocations stop running, like your
> snapshot did.
>
> So while if you really want to do this, we can, but my advice would be
> to let the tangle drop off into history ;-)

I'm very nearly persuaded by this... but I feel that I don't
understand why some of the ways of looking at the merge commit show so
many diffs.  For example:
http://git.savannah.gnu.org/cgit/guile.git/commit/?id=2f9ae9b1040e1b9339bb0bc8b0013a5346622c44

(On the other hand, gitweb -
http://git.savannah.gnu.org/gitweb/?p=guile.git;a=commitdiff;h=2f9ae9b1040e1b9339bb0bc8b0013a5346622c44
- just says "Trivial merge", which is more reassuring, and I think
reflects reality.)

I'm imagining looking back through the history at some future time
when I've forgotten what happened here.  When that happens, I don't
want to be misled into thinking that the merge commit made loads of
changes, when in fact it didn't.

Is it just that I'm misunderstanding what cgit shows?

Regards,
        Neil

Re: low-hanging SRFI fruit (SRFI-98)

[Neil Jerram]

Neil Jerram <neil@ossau.uklinux.net> writes:

> Is it just that I'm misunderstanding what cgit shows?

Yes, it seems that this is just what cgit chooses to do.  From
http://hjemli.net/git/cgit/commit/?id=8a3685bcf2612206fc24a2421acb53dd83aeab85:

"The diffstat is calculated against the leftmost parent of the
commit. This gives nice information for "normal" merges while octopus
merges are less than optimal, so the diffstat isn't calculated for
those merges."

This seems obviously less useful to me than what gitweb does - which
is primarily to show the conflicts, but also make the complete parent
diffs available just a click away.  So it's either a reason to prefer
gitweb, or just something that one has to be aware of with cgit.

Anyway, I'm happy enough to leave master as is now.  I'll delete
`njmaster'.

Regards,
        Neil