Fascinating stuff. Well worth reading.
But many of the tricks have notes indicating that they invoke
undefined behavior according to the C Standard, or that their
efficiency depends on the target system's assembly language
And in all cases, there's a more straightforward way to do these
Jens Arnold wrote:
For this purpose, file.c caches up to one sector worth of data.
headbytes is any new data that fits into the sector cache
when it's already dirty.
That implies thst the file is being /written/ to disk? What's being
written when I'm just listening to music?
I'd like CVS commit access, but I don't know that I'd be much more
proactive than the current devs -- a patch /should/ a certain amount
of time on the tracker, to ensure it's beta tested.
That said, tooting my own horn, I think a couple of my patches are
pretty much no-brainers and if I had
I dunno about the nano, but on the ipod video, the framebuffer's just
a/an (two dimensional) array of short.
The defines for all lcd types are in lcd.h, starting on line 203 or
there abouts. I reproduce them here:
/* Memory copy of display bitmap */
#if LCD_DEPTH == 1
extern fb_data
Oh.
Currently I don't think we have the requisite code speed to play
movies on the ipod. To get there, we'll need at least two things: a
faster memcpy and faster video routines.
I've been working on an assembly language memcpy to help with this;
currently it is faster on large copy lengths
I'm trying to include some ARM assembly from the linux kernel sources
to rockbox for ipod.
As I'm not real familiar with the linux sources, could you do three
things for me for me?
1) point me to the latest and best version of:
linux/arch/arm/lib/memcpy.S
(is it 2.6.15-23?)
2) And the
In my last email to you, Abigail, I announced a patch to optimize
gui_list scrolling when the first item on a page remains the same.
That worked by only updating the two lines that changes, the line the
cursor left and the line the cursor arrived at.
I've now implemented another optimization,
A patch to optimize apps/gui/list.c, function gui_list_draw.
Previously, all lists were always redrawn in their entirety. With this
patch, when the list's start position (the item at the top of the
list) hasn't changed, only the /lines/ changed are updated. In almost
all cases, that means
Bluechip wrote:
You may (or may not) also like to take a moment to review the
history of the developer in question here, he has an impeccable
reputation (far from that which I could claim for myself) ...I
can't believe that he is doing this for anything other than
honourable reasons.
Is there a way to get timings with a greater resolution than ticks, on
the ipod?
The ipod 4G button reading code is divided into two parts. Function
ipod_4g_button_int( void ) sets up some ports, tnen calls
ipod_4g_button_read() which reads some ports.
Can ipod_4g_button_read() be called without first calling
ipod_4g_button_int?
I'm having some trouble understanding the status bar code. It looks
like it mostly polls values.
I have some values it just isn't prudent to poll. How can I force the
statusbar to display some values, immediately when I call a function
passing those values.
Even better, I'd like the bar to
The patch tracker seems to give (or Firefox assumes) the mime type of
the link to a patch (e.g.,
http://www.rockbox.org/tracker/?getfile=11546) is
application/octet-stream. This means that Firefox insists that the
file be downloaded or a handler program be specified, and worse,
refuses to
Stephan Wezel wrote:
or we could switch to svn(subversion). Because svn has an option with
that the eol-style ist set depending on the OS on wich the checkout is
done.
I heartily second this. Having used both cvs and subversion,
subversion is much better in its features and more
For years I've used scite for development, but am thinking that I want
to try an IDE.
Has anyone anything to suggest about using kdevelop for rockbox, or
another linux ide to suggest?
Where should I post, to the forum or to the mailing list?
I've been posting to the rockbox-dev list with any general dev or
build questions.
Lately, I realized there's an online forum.
Where should we be posting?
I prefer the lists, mostly because I only need to monitor two: rockbox
and
not and
unfortunately I don't speak the language.
Andreas Gnau wrote:
[EMAIL PROTECTED] schrieb:
Can I beg a bigger copy from someone with a legal copy of the VMWare
workshop?
There was a full version of VMWare 4.5.2 in the German c't Magazin some
time ago. Maybe you can get one via ebay. The only
I've been using (and loving) the VMWare debian install created for
rockbox. My only problem is I've run out of space on the virtual disk.
Can I beg a bigger copy from someone with a legal copy of the VMWare
workshop?
Thanks.
Patch to increase lcd speed on the ipod video.
Decreases runtime of lcd_update_rect by an average of 0.1 ticks per
call, or 1 millisecond; and up to three milliseconds max. This
increases the maximum frame rate by about one to one-and-one-half
frames per second.
Using the test_fps plugin,
The ipod4G scrolling code doesn't check for scrolling if the button
queue is not empty. So far so good.
At line 159:
/* The queue should have no other events when scrolling */
if (queue_empty(button_queue) old_wheel_value = 0) {
If the queue is empty and the loop is entered, the change in
Enclosed please find source code to a simple state machine designed to
parse input. Your comments are requested.
The state machine can parse input from a refillable buffer of char,
maintaining state across buffer refills. The machine returns a pointer
into the buffer, indicating where a
1. This code in playlist.c look suspicious: playlist-amount is only
tested against playlist-max_playlist_size AFTER playlist_amount is
used as an array index. (Lines 510, 513, 515.)
2. And, if amount = max_playlist_size, the function returns without
calling queue_post. Shouldn't we be able to
when I compile the iPod video target with64MB memory, I get a usable
firmware that uses all 64MB of memory, but none of the codecs or
plugins work.
To compile to 64MB, I changed line 757 of tools/configure from
memory=32 # 30GB models have 32MB, 60GB have 64MB
to
memory=64 # 30GB models have
55a20e9416034badded309ccc0346dc6 */cygdrive/j/.rockbox/codecs/vorbis.codec
a0b9dd68bf748d8d32efa9a1ab99d243 */cygdrive/j/.rockbox/codecs/wav.codec
1bde859c58332b36620c8ec816e15c41
*/cygdrive/j/.rockbox/codecs/wavpack.codec
Dave Chapman wrote:
[EMAIL PROTECTED] wrote:
when I compile the iPod video
The whole problem with the world is that fools and fanatics are always so
certain of
themselves, and wiser people so full of doubts. --Bertrand Russell
I couldn't agree more.
I have so many doubts in my life that I prefer keep them to myself :)
Good move insulting all developers in one shot.
25 matches
Mail list logo
