Re: Grayscale on Handspring Visor and handling of maxdepth?

2001-07-13 Thread David A. Desrosiers


 The Visor does not seem to accept 4-grayscale (2bpp) images, it only
 shows them in black-and-white (although I set colour depth to 2 in the
 Preferences) and stretched by a factor of 2 vertically (so I can only
 see the left half of the image).

I have a Visor Deluxe, and although the battery life on the Visor is
atroscious, it definately displays 2bpp greyscale properly. The stretched
image is a hint that you're probably using a --max{width|height} value
somewhere or an --alt{maxwidth|maxheight} value. I've never seen images
stretched on Plucker at all, unless I screwed up those two values.

As you probably figured out, PalmOS versions before 3.0 only
supported monochrome in APIs. Palm OS versions 3.0 through 3.3 somewhat
supported 2 bit gray. Palm OS 3.5 and above supports up to 8 bit for devices
that support it. Your Visor is running 3.1H, modified a bit for Handspring's
hardware.

That being said, let's go through a quick debug:

1. Are you running any hacks on your Visor? Any SysTrap() calls?

2. What version of Plucker?

   What version of the parser?

   What version of Python?

   What version of netpbm2/PIL/ImageMagick?

3. What is the URL you're gathering, and what are the parameters
   you're passing to it to gather it?

4. What OS on your host computer? Windows? Solaris? OS/2? Linux?

5. What does your ~/.pluckerrc look like? Are you overriding
   parameters in ~/.pluckerrc by passing the parser arguments when
   you run it?

With those answers, we can better diagnose what might be going on.

 I am certain that my Visor can display 4 grayscales, the Handspring
 About or Info windows display grayscales.

This may help:

http://www.palmos.com/dev/tech/hardware/compare.html

 I have tried various combinations of netpbm and ImageMagick and the
 --bpp option, but could not achieve a satisfying result.

You mean netpbm2, right? You don't see 'ppmtoTbmp' trying to run
anywhere, do you? Are your netpbm2 palm templates working? Are they in the
right place to be found by netpbm2? For example (in .pluckerrc)

palm1bit_graymap_file = /usr/lib/palmgray1.map
palm2bit_graymap_file = /usr/lib/palmgray2.map
palm4bit_graymap_file = /usr/lib/palmgray4.map
palm8bit_stdcolormap_file = /usr/lib/palmcolor8.map
palm16bit_stdcolormap_file = /usr/lib/palmcolor16.map

You'll only need the first two since your Visor will not support
more colors/levels of grey than that.

 The other problem is the handling of the maxdepth flag (-M or MAXDEPTH
 directive in HREFs). With a maxdepth of 1, it only seems to download the
 specified page but not to follow any links, but with maxdepth=2 it seems
 to craw an additional 2 levels down the tree. I would want it to only
 follow one additional level, but I cannot seem to manage that. What am I
 missing here?

Currently, a maxdepth value of 1 will download the page referenced.
A maxdepth of 2 will download that page, and the pages of the links
referenced on it (I still disagree with this numbering, but.. not my call
right now). When you run the parser, and use --maxdepth=1 or --maxdepth=2,
there should only be a 1-page difference in the levels it gathers. It
shouldn't skip any levels, unless there are overriding parameters in either
the URL construct you pass to Plucker, or the ~/.pluckerrc file again.  Try
using --maxdepth=2 on the commandline and see if that helps. I've never ever
used the -M at all, so I can't say that it works, but I know --maxdepth
definately does.

Hope that helps.


/d





Re: Grayscale on Handspring Visor and handling of maxdepth?

2001-07-13 Thread Bernd Sieker

On 13.07.01, 07:24:30, David A. Desrosiers wrote:

Hi,

First, thanks a lot for the quick reply.

 
   I have a Visor Deluxe, and although the battery life on the Visor is
 atroscious, it definately displays 2bpp greyscale properly. The stretched

Do you mean it lasts long or it doesn't? I think that alkaline
batteries last rather long in my visor, but then, I have nothing to
compare it to.

 image is a hint that you're probably using a --max{width|height} value
 somewhere or an --alt{maxwidth|maxheight} value. I've never seen images
 stretched on Plucker at all, unless I screwed up those two values.

I don't use any option like that at all. And if I choose 1bpp images,
all is fine.

 
   As you probably figured out, PalmOS versions before 3.0 only
 supported monochrome in APIs. Palm OS versions 3.0 through 3.3 somewhat
 supported 2 bit gray. Palm OS 3.5 and above supports up to 8 bit for devices
 that support it. Your Visor is running 3.1H, modified a bit for Handspring's
 hardware.

Version info says System PalmOS 3.1H2. I thought the H was for
modified by Handspring, but I might be wrong.

 
   That being said, let's go through a quick debug:
 
   1. Are you running any hacks on your Visor? Any SysTrap() calls?

None that I know of. I have the Battery Preferences Panel, but its
docs say it is a proper application, and not a hack. The next closest
thing to a hack is a timer program that keeps running when other
applications are used or the visor is turned off. It's called
ToastTimer.

Remember that I'm a PalmOS newbie, and might not be aware of anything
conflicting with proper operation of other programs.

It might be of interest that I have both a German PalmOS and use the
German-language version of the plucker PalmOS viewer. (Oh btw, there
are some spelling errors, the most prominent being that standard is
spelled with a trailing d in German, too, and not with t. (Just
like, to be picky, definitely is spelled without any a in
English. ;-) Many confuse that, but look it up ...)

 
   2. What version of Plucker?

1.1.5

 
  What version of the parser?

Parser.py   $Id: Parser.py,v 1.13 2001/02/04 15:50:23 nordstrom Exp $


 
  What version of Python?

1.5.2, installed via NetBSD pkgsrc

 
  What version of netpbm2/PIL/ImageMagick?

Netpbm 9.7, installed via NetBSD pkgsrc
ImageMagick 5.2.8, installed via NetBSD pkgsrc.

Both cause the same problem.

Plus hand-installed ppmtoTbmp.


 
   3. What is the URL you're gathering, and what are the parameters
  you're passing to it to gather it?

Spider.py -P . -M 1 --bpp=2 -H http://www.netbsd.org/ -f NetBSD

 
   4. What OS on your host computer? Windows? Solaris? OS/2? Linux?

NetBSD-1.5.1

 
   5. What does your ~/.pluckerrc look like? Are you overriding
  parameters in ~/.pluckerrc by passing the parser arguments when
  you run it?

I will attach a gzip'd version of my ~/.pluckerrc, I'm not qualified
to sort out what might and might not be important for my problems.

 
   With those answers, we can better diagnose what might be going on.

Thanks for your efforts. Still, plucker already is a great tool for
me, and I hope we can make it _perfect_, soon :)

 
  I am certain that my Visor can display 4 grayscales, the Handspring
  About or Info windows display grayscales.
 
   This may help:
 
   http://www.palmos.com/dev/tech/hardware/compare.html
 
  I have tried various combinations of netpbm and ImageMagick and the
  --bpp option, but could not achieve a satisfying result.
 
   You mean netpbm2, right? You don't see 'ppmtoTbmp' trying to run
 anywhere, do you? Are your netpbm2 palm templates working? Are they in the
 right place to be found by netpbm2? For example (in .pluckerrc)

netpbm 9.7 plus hand-installed ppmtoTbmp. Without ppmtoTbmp I get lots
of errors during gathering, and don't see any images at all in
plucker. After installing ppmtoTbmp I get b/w images fine, but
incorrect rendering of greyscale images.

What exactly is netpbm2? Neither on google nor on freshmeat I can
find a single reference to it.

I have found a stand-alone copy of ppmtoTbmp (version 1.1) and
isntalled that and its q2.map and q4.map. I tried explicitly
specifying these two maps in ~/.pluckerrc, but that did not help
either.

I guess that plucker correctly finds all needed tools during
conversion, since I cannot see any errors related to file
conversion. Whereas previously, when ppmtoTbmp was missing, I could
see many such errors.



 [maxdepth clarified] 

Maybe I'm confused here, I'll try to sort it out. But thanks for the
hints. It always gathers all images neccessary to display the page,
right? Maybe that's what was confusing me, seeing such an enourmous
increase in files fetched between levels 1 and 2. (from about 10 to
about 1000.)

 
   Hope that helps.
 
 
 /d
 

Cheers,

Bernd

-- 
Bernd Sieker



plucker multiple database problem

2001-07-13 Thread Anthony Schellenberg



Hi. I'm new to plucker and can't seem to get 
anything other than the default database to work. I don't think anything 
is missing from my html, since even an exact copy of my home.html from the 
default.db directory will not work. Inside the default.db directory everything 
works fine. When I start the hotsync for a different database, it grabs a 
gif image that I have put on my html, then finishes as if thats all there is in 
the file. It ends up with a 0k file for my palm. Any 
ideas?

palm iiic, dev. version of OS 
4


Server relocation complete

2001-07-13 Thread David A. Desrosiers


There, that was relatively painless. The server has now been
relocated from one coast to the other (3,051 miles away). Zero downtime.

New things so far:

- MUCH faster access and response time now. Connection is roughly
  four times faster than the previous network.

- plkr.{com|net|org} domains route to the same page (it's our
  CreatorID, plucker.{com|net|org} were taken, so I registered this
  one). The previous domain, plucker.gnu-designs.{com|net|org} still
  will work. Someone want to volunteer to go through the docs and
  code and change the references to point to plkr.org instead? Or
  should we leave it as-is?

- CVS is now directly accessible via cvs.plkr.{com|net|org},
  plucker.sourcefubar.{com|net|org} (.com and .org seem to be down
  for some reason, routing problem, but .net is up, I've filed an
  incident)

- bugs.plkr.{com|net|org} works.

  The author of Mantis and I were working on some bugs last night in
  the new code. At some point in the near future, I'll roll to that
  new codebase. There are a lot of new things waiting in the wings.
  The most important for us as developers is the email-only
  interface into the system, so we don't have to be in a browser to
  update/view/close bugs. There's also a bit of a bug where
  bugs.plkr.org brings you to the generic project page. I'm going to
  add some code to allow that to bring you right to the Plucker bugs
  page. All the data is in one database, so it's not as easy as it
  looks to select it. Rasmus, Prescience and myself came up with a
  really good idea last night to try on this. More on that later.

- ftp is currently disabled, until I can secure it properly with the
  new account system I'm implementing. Mike, I'll pull your releases
  from sslug.dk for the moment until that's back online. Once I get
  this set back up and locked down, I'll make the releases
  accessible via gopher and ftp as well as web.

  I'm trying to give the developers a 'Release' page where we can
  upload and add a news blurb for releases. Delegation! =)

- cvs-over-ssh for developers, pserver for anonymous users. Much
  more secure, and allows the developers to use the box to do builds
  and also to change their password. CVS itself doesn't include a
  facility to change cvs passwords. This will fix that. chrooting
  ssh into a quota'd compile farm has proven to be a bit
  interesting. A week or two more and it will be working for
  everyone.

  cvs for those who are already using it will work as normal, but if
  you want to begin to point your $CVSROOT to
  cvs.sourcefubar.net:/cvs/plucker, that will be static. You could
  also use cvs.plkr.org:/cvs/plucker if you wanted. They both point
  to the same thing. I've changed it on the Contact page as well.

I'm also getting the mailing lists aggregated into a search engine.
I've got every single email since the beginning here online, so I'll add
that. Should be a week or three, once all the other services prove to be
stable.

I had the chance to interview Mark Lillywhite the other night, and
will stick the contents of that online as well. Mark, thanks! You may
receive a surprise in the mail soon, once I can figure out how to ship it
down there.

If there's anything else anybody wants to add, comment on, critique,
please let me know. Once I'm done with this migration and domain juggling
mess, I'll be hopping back on the codebase. I need to get away from
pilot-link for a few weeks, my eyes are going blurry looking at it.

There's a lot of new things planned, and quite a few other people
are jumping onboard and offering to help with various bits and pieces of the
project. This is good! To those doing translations, code fixes,
documentation, everything else, on behalf of the entire team, we thank you.
Keep up the great work!

Now off to catch a standby flight to the other coast to celebrate my
30th birthday with my girlfriend. Everybody, have a great weekend!



/d


p.s. (I'll be email-accessible if there are problems)