[racket] Best way to propose changes to Scribble CSS files
This is amazing! Thank you very much!
Given all the little rendering things that people are pointing out, I
should say that the docs are currently using a transitional DOCTYPE,
which could be part of the problem. You probably have good reasons for
it, but I thought I'd mention it just in case.
Thanks again!
Dave
On 11/10/2013 12:16 AM, Matthew Butterick wrote:
> Thank you for the comments so far. As for the gentleman who was
> "reluctant to say anthing negative," please don't be --- the point of
> making a prototype is to find out what doesn't work.
>
> I won't cover every suggestion here on the mailing list because it
> will become unwieldy. But I'll move the trickier topics onto the
> github repo as issues, and those interested can comment further there.
>
>
>
>
> On Sat, Nov 9, 2013 at 5:52 PM, Greg Hendershott
> <greghendershott at gmail.com <mailto:greghendershott at gmail.com>> wrote:
>
> This is so wonderful and I'm so happy you did this. It is awesomely
> awesome. Truly.
>
> Thank you, thank you, thank you!
>
>
> My only gripe so far is the floating search box that insists on being
> in the top-left corner and won't scroll away. On a desktop browser,
> some people find that sort of thing mildly annoying. On a phone
> browser, it's worse. For example if you double-tap the main column to
> zoom in, the search box is still stuck up there floating on top of
> text you want to read. tl;dr I suggest letting it scroll off normally
> along with "on-this-page", preferably IMO always, but at least on
> small screens.
>
> One other mobile issue: At least on Chrome for Android, the main text
> is a comfortable size, but the code examples are very tiny. (Whereas
> on the desktop, the sizes seem relatively equal.)
>
>
>
> On Sat, Nov 9, 2013 at 2:12 AM, Matthew Butterick
> <mb.list.acct at gmail.com <mailto:mb.list.acct at gmail.com>> wrote:
> > Rather than edit the Scribble CSS files in the main Racket repo,
> I decided
> > it would be more efficient to make a new repo to act as a
> prototype (and
> > serve it via github pages). That way, it's easier for others to
> try the new
> > CSS and report problems & suggestions. Once the changes look
> good, I can put
> > them into a pull request for the main repo.
> >
> > What I did is grab part of the docs from the built version of
> 5.90.0.9 and
> > dropped that in the prototype repo. Then I edited the CSS files.
> >
> > The prototype repo is here:
> >
> > https://github.com/mbutterick/racket-doc-redo/tree/gh-pages
> >
> > To preview the pages in a web browser, start here:
> >
> > http://mbutterick.github.io/racket-doc-redo/doc/index.html
> >
> > This is my first attempt at using github pages, so if there's
> breakage,
> > blame me.
> >
> > As for the design changes, there's more refinement and
> nitpickery to come,
> > but the basic idea is intact, and good enough to criticize.
> Mostly I've
> > aimed to simplify and update the layout, while keeping the
> character of the
> > documentation intact. I've also tried to address one key functional
> > shortcoming of the current CSS: its fixed width.
> >
> > And yes, it is somewhat less colorful overall, though not
> because I oppose
> > "the colors of the rainbow" ;) Rather, I just think the color
> has better
> > effect when it's used sparingly. Like a day spa for the mind.
> >
> > Matthew Butterick
> >
> >
> >
> >
> > On Wed, Oct 16, 2013 at 3:41 PM, Robby Findler
> <robby at eecs.northwestern.edu <mailto:robby at eecs.northwestern.edu>>
> > wrote:
> >>
> >> I've just finished reading your (beautiful!) book and am
> excitedly looking
> >> forward to what you come up with. Do let us know if you get
> stuck anywhere.
> >> (And yes: we apparently like all the colors of the rainbow more
> than you
> >> seem to; hopefully you won't hold that against us :).
> >>
> >> Robby
> >>
> >>
> >>
> >> On Fri, Oct 11, 2013 at 5:35 PM, Matthew Butterick
> >> <mb.list.acct at gmail.com <mailto:mb.list.acct at gmail.com>> wrote:
> >>>
> >>> Consistent with my pledge at RacketCon, I've been working on some
> >>> potential improvements to the default CSS files used by
> Scribble for Racket
> >>> documentation. Before I get too far I just want to make sure
> I'm going about
> >>> it the right way.
> >>>
> >>> I've read the tutorials by Joe Politz and Greg Hendershott
> about how to
> >>> contribute to Racket via Github. I made a fresh fork of
> plt/racket yesterday
> >>> and built it from source. But the Scribble CSS files are
> handled a little
> >>> differently than others.
> >>>
> >>> I see that the documentation gets built into racket/racket/doc/,
> >>> including the CSS files. So if I edit the files in that
> directory, I can see
> >>> the CSS changes reflected in the docs. However, the whole doc
> directory is
> >>> ignored in the git repo. And I need to edit files that git can
> see.
> >>>
> >>> So I found the original home of the CSS files in
> >>> racket/pkgs/scribble-pkgs/scribble-lib/scribble/. If I update
> these files,
> >>> then git sees them. But the changes aren't reflected in the live
> >>> documentation.
> >>>
> >>> My workaround has just been to replace the copies in
> racket/racket/doc
> >>> with symlinks to the files in scribble-pkgs. That way, as I
> update the CSS
> >>> in scribble-pkgs, git can see the updates, but they're also
> reflected in the
> >>> live docs. (These symlinks will get wiped out next time I
> rebuild from
> >>> source, but that's the price of progress.)
> >>>
> >>>
> >>> 1) What's the best way to propose Scribble CSS updates? Should
> I assemble
> >>> a pull request for
> racket/pkgs/scribble-pkgs/scribble-lib/scribble/ ?
> >>>
> >>> 2) Is there a better way of connecting the CSS file in
> scribble-pkgs to
> >>> the actual CSS file used by the documentation? (i.e., other
> than my symlink
> >>> technique).
> >>>
> >>> 3) Anyone who wants to try out the new Scribble CSS files or
> contribute
> >>> to the update will have the same problem, however. I'm not
> sure how to avoid
> >>> this given that Scribble's HTML rendering policy is to bring
> the CSS files
> >>> along for the ride and eliminate dependency on the source
> directory. OTOH,
> >>> it's a drag to have to rebuild the docs just to see the effect
> of a few CSS
> >>> files.
> >>>
> >>> 4) For now I'm just working with the CSS, and not delving into the
> >>> Scribble HTML renderer, on the idea that changing fewer files
> is better, and
> >>> maintaining compatibility with existing doc sources is
> essential. That said,
> >>> there are some occasional defects in the Scribble HTML output
> that puts
> >>> things out of reach of CSS (e.g., I've found styling
> hard-coded into the
> >>> HTML in places).
> >>>
> >>>
> >>> Matthew Butterick
> >>>
> >>> ____________________
> >>> Racket Users list:
> >>> http://lists.racket-lang.org/users
> >>>
> >>
> >
> >
> > ____________________
> > Racket Users list:
> > http://lists.racket-lang.org/users
> >
>
>
>
>
> ____________________
> Racket Users list:
> http://lists.racket-lang.org/users
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.racket-lang.org/users/archive/attachments/20131110/9d069032/attachment-0001.html>