Re: Patches: Trivial (add div), Progressive Enhancement, Externalise CSS/JS

[Alex Sassmannshausen]

Hi,

Sooner or later we will get there! :-)

To make the process easier, I've split the patches into 2 series. For
now, the JavaScript Progressive Enhancement patches.

>>>>> "Ludovic" == Ludovic Courtès <address@hidden> writes:

 alex sassmannshausen <address@hidden>
 > The second patch re-introduces changes to the SXML and JavaScript
 > to fulfill the criteria of Progressive Enhancement: - All content
 > is shown when JavaScript is not enabled on GUI browsers.  In
 > addition the patch implements JS that carries out the changes to
 > the HTML document successively rather than all at the end, which,
 > with the current size of the page, would cause a 'flicker'.

 Ludovic> Nice.  (What’s Progressive Enhancement?)

Progressive Enhancement is a web design/development approach where you
make sure that all functionality is available using a very low baseline
— i.e. no JS, smallest files possible, attempt to take old browsers into
account — and then add functionality as 'icing on the cake'. In this
way, users are (almost) never left behind.

 alex sassmannshausen <address@hidden> skribis:
 > The second patch re-introduces changes to the SXML and JavaScript to

 [...]

 > +        (td (span (strong ,(package-synopsis package)))

 Ludovic> Shouldn’t we use CSS instead of <strong>?

You are right. I've now changed this (patch 1 no longer uses strong,
patch 3 implements CSS font-weight heaviness.

 > +(define show_hide-grouper

 Ludovic> I should have mentioned it: the name of a procedure should
 Ludovic> describe its result or computation (when it’s a pure function,
 Ludovic> like ‘expt’), or its effect (like ‘display’).  Otherwise it’s
 Ludovic> harder to tell what it does etc.

In an attempt to do this I have now renamed the JS function
prep_pkg_descs (for prepare_package_descriptions), and the Scheme
function you are discussing here is now called
insert-prep_pkg_descs. Hope that does the trick.

 > +  (let ((lid '())
 > +        (group-counter 15))
 > +    (lambda (id)

 [...]

 > +      (if id
 > +          (begin
 > +            ;; If ID is not false, then we add ID to LID.

 Ludovic> But this function still has a single call with #f as its
 Ludovic> argument.  Can you remove the parameter and the dead code?  Or
 Ludovic> am I missing something?

Actually that function (now called insert-prep_pkg_descs is called in 2
places in the Scheme code, once with #f as argument (in packages->sxml),
and once (at the end) of package->sxml, with each package-id as
argument. The code is not dead, as it is what accumulates up to 15
package-ids, which are then used when prep_pkg_descriptions is inserted
into the HTML table (do a search for insert-prep_pkg_descs, and you
should get three hits, one being the definition.

 [...]

 > +/* Take n element IDs, prepare them for javascript enhanced
 > +display and hide the IDs by default. */
 > +function bulk_show_hide()

 Ludovic> Align ‘display’ with ‘Take’.

Done.

 > From 3c3bfb6dea1b20447ba8bb48ec95a8cc4b556129 Mon Sep 17 00:00:00 2001
 > From: Alex Sassmannshausen <address@hidden>

 [...]

This part deals with the splitting of the CSS and JS into separate
source files; I'll address that with a second patch series later (or
tomorrow).

 [...]

 > diff --git a/build-aux/package-list.css b/build-aux/package-list.css
 > new file mode 100644
 > index 0000000..67d423a
 > --- /dev/null
 > +++ b/build-aux/package-list.css

 Ludovic> Can you indent this file in C-mode in Emacs or something
 Ludovic> similar?

This has been done in patch no. 2, if I understood you correctly.

 Ludovic> Thanks for following up!

No problem, thanks for the thorough feedback — let me know if these
patches cause any problems.

Best wishes,

Alex

0001-list-packages-Progressive-Enhancement-approach-to-JS.patch
Description: Text Data

0002-list-packages-Tidy-CSS-in-preparation-for-split-into.patch
Description: Text Data

0003-list-packages-Improve-CSS-for-legibility-remove-redu.patch
Description: Text Data