[hunchentoot-devel] edicl & Quicklisp - formal release tarballs or production-ready master branches?

[Edi Weitz]

Content preview: On Fri, Sep 9, 2011 at 7:49 PM, Sebastian Tennant <***@smolny.plus.com>
wrote: > Incidentally, the library I'd actually like to see updated soonest
is cl-who as > the version in Quicklisp doesn't include support for HTML5.

Perhaps I could assist with 'releasing' an updated version of cl-who

as a > warm-up to assisting you with a hunchentoot release? [...]

Content analysis details: (-0.7 points, 5.0 required)

pts rule name description
---- ---------------------- --------------------------------------------------
0.0 FREEMAIL_FROM Sender email is commonly abused enduser mail provider
(nhabedi[at]gmail.com)
-0.7 RCVD_IN_DNSWL_LOW RBL: Sender listed at http://www.dnswl.org/, low
trust
[74.125.82.179 listed in list.dnswl.org]
-0.0 SPF_PASS SPF: sender matches SPF record
0.0 T_DKIM_INVALID DKIM-Signature header exists but is not valid
Archived-At: <http://permalink.gmane.org/gmane.lisp.who.general/274>

On Fri, Sep 9, 2011 at 7:49 PM, Sebastian Tennant

Incidentally, the library I'd actually like to see updated soonest is cl-who as
the version in Quicklisp doesn't include support for HTML5.
Perhaps I could assist with 'releasing' an updated version of cl-who as a
warm-up to assisting you with a hunchentoot release?

It's been a while (maybe two years ago), so I'm not totally sure, but
I /think/ the unreleased version of CL-WHO contains architectural
changes of mine which so far haven't been sufficiently documented
and/or tested. You might want to take a look at this before you make
a release.

Just a recommendation. Of course, you can proceed as you like... :)

Cheers,
Edi.

PS: cl-who-devel on Cc.

[Sebastian Tennant]

[...] I /think/ the unreleased version of CL-WHO contains architectural
changes of mine which so far haven't been sufficiently documented and/or
tested.

Not surprisingly, the bulk of the changes are in who.lisp and a lot of these
arise from fifteen function definitions having been moved to the newly
introduced file; util.lisp. Disregarding these, along with changes reuslting
from the introduction of HTML5 support and the introduction of the test suite,
and the only 'architectural' changes we need to consider (vis-a-vis testing and
documentation) can be summarised thus:



Multiple changes to function CONVERT-ATTRIBUTES.

Function REPLACE-HTM & TREE-TO-COMMANDS-AUX removed, affecting functions
TREE-TO-TEMPLATE & TREE-TO-COMMANDS.

util.lisp introduces two new (unexported) macros; WITH-UNIQUE-NAMES (used in
WITH-HTML-OUTPUT) & WITH-REBINDING (currently unused).

Further changes (unrelated to WITH-UNIQUE-NAMES) to macro WITH-HTML-OUTPUT.

Functions ESCAPE-STRING-ISO-8859 & SHOW-HTML-EXPANSION removed, not affecting
any other functions.



As far as documentation goes (inline and in index.html), there's not that much
that still needs doing, except perhaps:

1. TREE-TO-COMMANDS lacks a documentation string. Not part of the API.

2. The lambda list of WITH-HTML-OUTPUT includes a new REST parameter which is
not documented in the docstring. Very much part of the API so needs
documenting in index.html too.

3. If WITH-REBINDING is not going to be used, should it not be removed, or a
note added explaining why it may be used in the future. Not part of the
API.

4. Canonical home of CL-WHO moved from BKNR to github - needs updating in
index.html.

It'd be great if you could find time to look at 1, 2 & 3 Edi, (assuming these
are the architectural changes you're referring to). I'm simply not qualified
(yet).

I can start testing however, and that is something I hope to get around to
soon.


Sebastian

P.S. Tarball (attached) may come in handy - it contains old and new versions of
the some of the functions in question which I used to produce simplified
diffs.

--
Emacs' AlsaPlayer - Music Without Jolts
Lightweight, full-featured and mindful of your idyllic happiness.
http://home.gna.org/eap

[Edi Weitz]

Content preview: Sebastian, thanks for taking the time to go through this.
Details below. I /think/ the main thing I wanted was an easier way for users
of CL-WHO to extend it (without having to deal with undocumented and/or unexported
internal stuff). Judging from your description of the changes, it seems that
I stopped somewhere in the middle. Maybe I even have uncommitted changes
somewhere at home. (But then they must be on an older laptop that I currently
don't have access to.) [...]

Content analysis details: (-0.7 points, 5.0 required)

pts rule name description
---- ---------------------- --------------------------------------------------
0.0 FREEMAIL_FROM Sender email is commonly abused enduser mail provider
(nhabedi[at]gmail.com)
-0.7 RCVD_IN_DNSWL_LOW RBL: Sender listed at http://www.dnswl.org/, low
trust
[209.85.213.179 listed in list.dnswl.org]
-0.0 SPF_PASS SPF: sender matches SPF record
0.0 T_DKIM_INVALID DKIM-Signature header exists but is not valid
Archived-At: <http://permalink.gmane.org/gmane.lisp.who.general/278>

Sebastian, thanks for taking the time to go through this. Details below.

I /think/ the main thing I wanted was an easier way for users of
CL-WHO to extend it (without having to deal with undocumented and/or
unexported internal stuff). Judging from your description of the
changes, it seems that I stopped somewhere in the middle. Maybe I
even have uncommitted changes somewhere at home. (But then they must
be on an older laptop that I currently don't have access to.)

Some comments inline:

On Fri, Sep 23, 2011 at 10:42 AM, Sebastian Tennant

 1. TREE-TO-COMMANDS lacks a documentation string.  Not part of the API.

Did it have a documentation string once? (I guess so.) If it had,
then I probably removed it because its behaviour changed. I probably
(?) also wanted to export it...

 3. If WITH-REBINDING is not going to be used, should it not be removed, or a
   note added explaining why it may be used in the future.  Not part of the
   API.

Should certainly be removed. I'm sure I just copied it from some of
my other libs.

Thanks again,
Edi.