[9fans] new links binary+source

[9fans] new links binary+source

[mirtchov]

New links binary + source -- Geoff Collyer's changes are in, which
means that links consumes much less cpu time.  Also rendering using
Plan 9 fonts is implemented -- right now Links renders only one size
using /lib/font/bit/lucm/unicode.9.font, which has an almost complete
unicode set.

There are provisions made for rendering in multiple size fonts,
however we need more complete font descriptions to be able to cover
the full gamma of possibilities.  Alternatively one can finally sit
down and write an autogenerating font file server using, for example,
ttf2subf.

If you have cyberbit you may want to experiment with it, but know that
html_gr.c sets the default font height to 17, which is what
unicode.9.font has.

You will find that rendering with Plan 9 fonts makes the browser much,
much faster and more responsive.  Scrolling/panning also become
useful, all of a sudden.

Beware -- the new event handling code causes the browser to crash if
resized while new york times is loading.  It came with Geoff's
changes, but I haven't been able to pin it down.

The source size has been significantly decreased because we don't
carry the precompiled fonts Links used to have (font_data.c and
fontdata).  The binary is only 700K gzipped and 3.2MB gunzipped.  The
code weighs in at ~2MB gzipped.

New screenshots also available, including rendering of japanese web
pages with kanji fonts.

Overall a significant improvement. It even seems usable sometimes.

andrey

ps: http://pages.cpsc.ucalgary.ca/~mirtchov/p9/links/

Re: [9fans] new links binary+source

[Axel Belinfante]

Just FYI:
the new links binary is now also much more usable
over a (cable modem, or even just a local net)
drawterm connection (before the (re)drawing of the pages
was very slow)

Thanks to all who contributed!

Axel.

Re: [9fans] new links binary+source

[okamoto]

> New links binary + source -- Geoff Collyer's changes are in, which
> means that links consumes much less cpu time.  Also rendering using
> Plan 9 fonts is implemented -- right now Links renders only one size
> using /lib/font/bit/lucm/unicode.9.font, which has an almost complete
> unicode set.

Thank you very much Geoff!
However, I still have to resize the window to see www.cnn.com...

Fonts becomes very nice, and surprizingly, I can read Japanese
home page now (http://basalt.cias.osakafu-u.ac.jp/plan9), who
did this enable?

Kenji

Re: [9fans] new links binary+source

[Geoff Collyer]

Links is still quite slow on some pages, notably nytimes.com and
cnn.com.  I haven't looked into why, but I suspect it's not good at
retrieving lots of small images.  There are no doubt lots of bugs left
in links.  One workaround is to select "Kill all connections" from the
File menu and then go to that URL again.  Whatever has been cached
already should be displayed while it fetches the rest.

I believe Andrey did all the font work and the faster rendering.

The fundamental structure of links is just wrong: it tries to simulate
threads or processes sharing memory by using select to trigger
functions when input arrives, and using a convention that upon
completion of some activity, there's a always a `completion routine'
to be called next, possibly by select.  It would be better on Plan 9
to just rfork a process for each http request and let it run to
completion, but doing that may require doing considerable violence to
links (not that I'm opposed to that ☺).

A few of us are also looking at using webfs, at Russ's suggestion.
The more we can move out of links, the better.

Re: [9fans] new links binary+source

[okamoto]

> The fundamental structure of links is just wrong: it tries to simulate
> threads or processes sharing memory by using select to trigger
> functions when input arrives, and using a convention that upon
> completion of some activity, there's a always a `completion routine'
> to be called next, possibly by select.  It would be better on Plan 9
> to just rfork a process for each http request and let it run to
> completion, but doing that may require doing considerable violence to
> links (not that I'm opposed to that ☺).

This statment is very specific, and constructive. ☺

Links is the first web browser which displays UTF-8 pages good
except of older Panel library based one which I personally made it.
I hope many Japanese web writer will begin to write their pages 
by UTF-8.☺  To make it possible for JIS, S-JIS looks like ridiculous
to me, because only Japanese use those codes...

> A few of us are also looking at using webfs, at Russ's suggestion.
> The more we can move out of links, the better.

Sure, if we have better web browse, it better for all of us.

Kenji

Re: [9fans] new links binary+source

[okamoto]

> Fonts becomes very nice, 

Links seems to use only one size of fonts for text?
People demand much when one becomes satisfied.☺

Kenji

Re: [9fans] new links binary+source

[andrey mirtchovski]

On Fri, 19 Dec 2003 okamoto@granite.cias.osakafu-u.ac.jp wrote:

> > Fonts becomes very nice, 
> 
> Links seems to use only one size of fonts for text?
> People demand much when one becomes satisfied.☺

http://pages.cpsc.ucalgary.ca/~mirtchov/screenshots/varfont.png

you'll have to wait a bit more for this one -- most of the fonts with good
variable sizes don't have very good utf glyph coverage, unless you want to
use non-standard fonts (for plan9 at least) such as efont or 10646.

in my estimation we need about 10-12 different heights from each different
font that people may be thinking of using to cover a reasonable set of
people's desires..

i believe work is being done to remedy that problem. 

Re: [9fans] new links binary+source

[okamoto]

> http://pages.cpsc.ucalgary.ca/~mirtchov/screenshots/varfont.png

Fum, is this a snapshot of links?
If so, it's very good.

Kenji

Re: [9fans] new links binary+source

[mirtchov]

>> http://pages.cpsc.ucalgary.ca/~mirtchov/screenshots/varfont.png
>
> Fum, is this a snapshot of links?
> If so, it's very good.

yes, it is :) this also has italic fonts, which the original links doesn't have :P

Re: [9fans] new links binary+source

[okamoto]

> in my estimation we need about 10-12 different heights from each different
> font that people may be thinking of using to cover a reasonable set of
> people's desires..

How about to use lucida and lucidasans, where I have many sizes.

Kenji