The 505 in libprs500 - Page 6

kovidgoyal · 11-16-2007, 01:08 PM

The most flexible way to control what goes into the table of contents is to use an OPF file to specify a separate html file that contains only those links that you want in the TOC

shoggot · 11-20-2007, 07:43 AM

Side note - I've noticed that a file with closing tags () take on the order of 12x time to process via html2lrf.exe - 3.2MB file, has ~14k tags. About 15 seconds without the close tags, about 4 minutes with.

kovidgoyal · 11-20-2007, 11:46 AM

Thats an interesting observation. One way to figure out why that may be is to run html2lrf with the --verbose switch. It will then write the result of the pre-processed HTML to a file. It may be that the preprocessed HTML is more complex in the second case. In any case, I would appreciate it if you could send me this file.

shoggot · 11-20-2007, 02:09 PM

Mmm, the file itself contains material covered under copyright... but as I've replicated the results with three separate files now, it should be easy enough to replicate.

The files were each over 3MB (as .htm files), and contained only the bare essential tags, so could be replicated like so, under bash:

Quote:

echo "<html><head></head><body></body>" > foo.htm
num=1
while [[ $num -lt 50000 ]];do
echo "abcdefghijklmnopqrstuvwxyz ABCDEFGHIJKLMNOPQRSTUVWXYZ 0123456789 " >> foo.htm
num=$(($num+1))
done
echo "</body></html>" >> foo.htm

And this is the commandline used against the result:
./html2lrf.exe --title="barfoo" --author="foobar" --font-delta=-2 --left-margin=5 --right-margin=5 --top-margin=5 foo.htm

Followed by a

Quote:

sed -e s/"<\/p>"//g foo.htm > foo2.htm

And run the commandline against that. Output seems to be identical, or within a few bytes either way. On-reader appearance is identical, anyway - just processing times.

I'm running a truncated version of that test file with verbose options, to see the results:

Well, the result's interesting (temp files). Snipping for space, but you'll see the diff (and I have no idea why it should have any effect.)
With tags:

Quote:

<html><head></head><body></body>
abcdefghijklmnopqrstuvwxyz ABCDEFGHIJKLMNOPQRSTUVWXYZ 0123456789 
abcdefghijklmnopqrstuvwxyz ABCDEFGHIJKLMNOPQRSTUVWXYZ 0123456789 
...
abcdefghijklmnopqrstuvwxyz ABCDEFGHIJKLMNOPQRSTUVWXYZ 0123456789 
</html>

Without tags:

Quote:

<html><head></head><body></body>
abcdefghijklmnopqrstuvwxyz ABCDEFGHIJKLMNOPQRSTUVWXYZ 0123456789
abcdefghijklmnopqrstuvwxyz ABCDEFGHIJKLMNOPQRSTUVWXYZ 0123456789
abcdefghijklmnopqrstuvwxyz ABCDEFGHIJKLMNOPQRSTUVWXYZ 0123456789
abcdefghijklmnopqrstuvwxyz ABCDEFGHIJKLMNOPQRSTUVWXYZ 0123456789
</html>

The closed </body> tag is due to my malforming the input html, but there you go.

If you really have to have the original file, I guess we'll make it happen in the name of progress, but I'd prefer to bypass that step.

kovidgoyal · 11-20-2007, 02:18 PM

That test case is fine, I don't need the original file. I'll look into it.

kovidgoyal · 11-20-2007, 02:40 PM

The basic reason is that the HTML "phrase \n" becomes a P object containing a single text object, then a text object and then another object while "phrase \n" becomes two P objects, each containing a single text object. Not much I can do about that I'm afraid, as that is the correct object representation of the HTML

EDIT: Incidentally you should use the --base-font-size option instead of --font-delta

MSWallack · 11-23-2007, 11:51 PM

I'm not sure if I've found a problem or if I'm getting the expected result. I finally connected my 505 to libprs500 (v0.4.24). In the table of contents of the 505, it appears that every book that I've added by using libprs has the author listed as Author Author (e.g., Arthur C. Clarke Arthur C. Clarke). Is this simply showing the Author field followed by the Author sort field or is something odd happening?

kovidgoyal · 11-24-2007, 12:17 PM

Hmm not sure why that should be happening. Can you save one these files to the disk and run lrf-meta on them.

MSWallack · 11-24-2007, 12:31 PM

Author (kana) is Clarke, Arthur C.
Author is Arthur C. Clarke

kovidgoyal · 11-24-2007, 01:21 PM

OK I see you've opened a bug report, I'll look at it when I return.

spooky69 · 11-25-2007, 06:23 AM

Hey, I plan on purchasing the PRS-505 within the next couple of months and I'd just like to say that I'm glad that there is somebody so dedicated to making the device as good as possible. I feel a lot more confident about my purchase because of this. Keep up the good work!

pcboy70 · 12-03-2007, 02:49 PM

I just joined this forum (just got a PRS-505). Looks like a great source of info.

I just installed the current version of libprs500 on my MacBook Pro running OS X.5.1 and it seems to mostly work (haven't tried everything yet but I did try the RSS feed from the BBC and it worked well) but - I can see my library, reader and my SD card but not my MS card (both are installed in the reader). Am I missing a step or does this need to be addressed?

Looking forward to using the software since Sony doesn't seem to care about me and my Mac!

pcboy70

kovidgoyal · 12-03-2007, 02:55 PM

libprs500 only supports one card at a time. That's not going to change.

kovidgoyal · 12-11-2007, 02:07 PM

Just to let everyone know, as far as I know, all outstanding issues with the 505 have been resolved.

JSWolf · 12-11-2007, 02:42 PM

Quote:

Originally Posted by kovidgoyal

Just to let everyone know, as far as I know, all outstanding issues with the 505 have been resolved.

Very nice!

11-20-2007, 07:43 AM	#77
shoggot Junior Member Posts: 6 Karma: 10 Join Date: Nov 2007 Device: sony prs-505	Side note - I've noticed that a file with closing <p> tags (</p>) take on the order of 12x time to process via html2lrf.exe - 3.2MB file, has ~14k <p> tags. About 15 seconds without the close tags, about 4 minutes with.

11-20-2007, 02:40 PM	#81
kovidgoyal creator of calibre Posts: 43,849 Karma: 22666666 Join Date: Oct 2006 Location: Mumbai, India Device: Various	The basic reason is that the HTML "<p>phrase </p>\n<p>" becomes a P object containing a single text object, then a text object and then another <p> object while "<p>phrase \n<p>" becomes two P objects, each containing a single text object. Not much I can do about that I'm afraid, as that is the correct object representation of the HTML EDIT: Incidentally you should use the --base-font-size option instead of --font-delta

11-23-2007, 11:51 PM	#82
MSWallack Right, Except When Wrong Posts: 353 Karma: 3968525 Join Date: Aug 2007 Location: Indianapolis Device: Kindle Oasis 3 (sometimes iPad Mini).	Author Author I'm not sure if I've found a problem or if I'm getting the expected result. I finally connected my 505 to libprs500 (v0.4.24). In the table of contents of the 505, it appears that every book that I've added by using libprs has the author listed as Author Author (e.g., Arthur C. Clarke Arthur C. Clarke). Is this simply showing the Author field followed by the Author sort field or is something odd happening?

12-03-2007, 02:49 PM	#87
pcboy70 Junior Member Posts: 1 Karma: 10 Join Date: Dec 2007 Device: PRS-505	Just joined - problem I just joined this forum (just got a PRS-505). Looks like a great source of info. I just installed the current version of libprs500 on my MacBook Pro running OS X.5.1 and it seems to mostly work (haven't tried everything yet but I did try the RSS feed from the BBC and it worked well) but - I can see my library, reader and my SD card but not my MS card (both are installed in the reader). Am I missing a step or does this need to be addressed? Looking forward to using the software since Sony doesn't seem to care about me and my Mac! pcboy70

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
libprs500	tommy still	Calibre	2	02-09-2008 10:55 AM
Using libprs500	jerryleejr	Calibre	2	01-28-2008 12:29 AM
Using Libprs500 for 505	JeffASonyReader	Calibre	17	01-21-2008 09:34 AM
Help!!! libprs500	MountainMan	Calibre	11	11-13-2007 11:25 PM
libprs500 and 505	carchase	Sony Reader	6	10-28-2007 03:06 PM

11-16-2007, 01:08 PM	#76
kovidgoyal creator of calibre Posts: 43,849 Karma: 22666666 Join Date: Oct 2006 Location: Mumbai, India Device: Various	The most flexible way to control what goes into the table of contents is to use an OPF file to specify a separate html file that contains only those links that you want in the TOC

11-20-2007, 11:46 AM	#78
kovidgoyal creator of calibre Posts: 43,849 Karma: 22666666 Join Date: Oct 2006 Location: Mumbai, India Device: Various	Thats an interesting observation. One way to figure out why that may be is to run html2lrf with the --verbose switch. It will then write the result of the pre-processed HTML to a file. It may be that the preprocessed HTML is more complex in the second case. In any case, I would appreciate it if you could send me this file.

11-20-2007, 02:18 PM	#80
kovidgoyal creator of calibre Posts: 43,849 Karma: 22666666 Join Date: Oct 2006 Location: Mumbai, India Device: Various	That test case is fine, I don't need the original file. I'll look into it.

11-24-2007, 12:17 PM	#83
kovidgoyal creator of calibre Posts: 43,849 Karma: 22666666 Join Date: Oct 2006 Location: Mumbai, India Device: Various	Hmm not sure why that should be happening. Can you save one these files to the disk and run lrf-meta on them.

11-24-2007, 12:31 PM	#84
MSWallack Right, Except When Wrong Posts: 353 Karma: 3968525 Join Date: Aug 2007 Location: Indianapolis Device: Kindle Oasis 3 (sometimes iPad Mini).	Author (kana) is Clarke, Arthur C. Author is Arthur C. Clarke

11-24-2007, 01:21 PM	#85
kovidgoyal creator of calibre Posts: 43,849 Karma: 22666666 Join Date: Oct 2006 Location: Mumbai, India Device: Various	OK I see you've opened a bug report, I'll look at it when I return.

11-25-2007, 06:23 AM	#86
spooky69 Stats: Posts: 233 Karma: 1189 Join Date: Nov 2007 Device: Sony PRS-505	Hey, I plan on purchasing the PRS-505 within the next couple of months and I'd just like to say that I'm glad that there is somebody so dedicated to making the device as good as possible. I feel a lot more confident about my purchase because of this. Keep up the good work!

12-03-2007, 02:55 PM	#88
kovidgoyal creator of calibre Posts: 43,849 Karma: 22666666 Join Date: Oct 2006 Location: Mumbai, India Device: Various	libprs500 only supports one card at a time. That's not going to change.

12-11-2007, 02:07 PM	#89
kovidgoyal creator of calibre Posts: 43,849 Karma: 22666666 Join Date: Oct 2006 Location: Mumbai, India Device: Various	Just to let everyone know, as far as I know, all outstanding issues with the 505 have been resolved.

Advert

Advert