Any chance of Baen Library in lrf format?

[Xenophon]

Originally Posted by JSWolf

But do your scripts take into account the need for some page break(s) and the fact that the paragraph indents are too big. I can go with l;eaving the quotes as is, but thge indents are too large and the only way to fix that is to go in and change the css code. Also, pages breaking needs to be fixed in at least one spot in a book I was testing last night. Also, the chapter text size is too large. These are things no script can fix unless the script edits the css code.

Nope. Those are things that would need to be addressed through pre-processing or some other command-line switch. A reasonably straight-forward thing to do... once someone figures out what the "right" fixes are.

[JSWolf]

Originally Posted by Xenophon

Nope. Those are things that would need to be addressed through pre-processing or some other command-line switch. A reasonably straight-forward thing to do... once someone figures out what the "right" fixes are.

Perhaps they could be made part of the --bean switch. That would do.

[jasonkchapman]

Originally Posted by DrMoze

The free beer offer is still open to anyone in the NYC area. Right now I have Heartland Brewery's Cornhusker Lager on tap.

Which totally rocks, by the way. (In case anyone needs a second opinion.)

[WayneD]

Originally Posted by Xenophon

The only remaining problem that I've seen is occasional missing scene breaks.

I've noticed some missing scene breaks in a non-Baen book I converted with libprs500, and it turned out to be a bug (whether in the reader or in libprs500, I haven't determined yet): a paragraph with a non-breaking space as the sole content is not rendering as an empty line. I tweaked my version of libprs500 to change all such paragraphs into (centered) "* * *" scene breaks (which I prefer anyway, since the break might appear at the edge of the page).

I wonder if this is the same problem you noticed?

[Xenophon]

Originally Posted by WayneD

I've noticed some missing scene breaks in a non-Baen book I converted with libprs500, and it turned out to be a bug (whether in the reader or in libprs500, I haven't determined yet): a paragraph with a non-breaking space as the sole content is not rendering as an empty line. I tweaked my version of libprs500 to change all such paragraphs into (centered) "* * *" scene breaks (which I prefer anyway, since the break might appear at the edge of the page).

I wonder if this is the same problem you noticed?

That's probably it. Would you please enter a ticket at libprs500.kovidgoyal.net with the necessary info so that kovid can fix this? You can certainly request the "* * *" option as part of the fix. In fact, you can even submit your patch (hint hint).

[bkilian]

Originally Posted by JSWolf

But do your scripts take into account the need for some page break(s) and the fact that the paragraph indents are too big. I can go with l;eaving the quotes as is, but thge indents are too large and the only way to fix that is to go in and change the css code. Also, pages breaking needs to be fixed in at least one spot in a book I was testing last night. Also, the chapter text size is too large. These are things no script can fix unless the script edits the css code.

I have no idea what you're on about this. The Baen CSS has "page-break-before" on H1, so adding --force-page-break-before-tag=h1 fixes the few that html2lrf misses, and comparing paragraph indents, the LIT format version actually has larger paragraph indents than the LRF converted version. If anything, the LRF versions are too small. (I'm assuming you're talking about the first line indent).
I use the LIT for conversions now, since it gives the best results from my extensive testing. I only have a few issues with it still:
a) lit2lrf does not use the defined <spine> in the OPF, and I really hate having the TOC actually in the book. The TOC should only be used to define TOC entries, the book should only contain the files defined by <spine>, in the order that <spine> defines them. For now I'm living with the ugly TOC pages, but I'd really prefer for them to go away.
b) Some files require individual settings, for instance "Wind Rider's Oath" by David Weber requires an embedded font (I use Times New Roman) because it has crazy characters (The second o in "Sothōii") that the default fonts don't support.

I solved (b) in my case by allowing individual files to define extra settings, and the ability to create a defaults file for an entire directory tree.
Here are the DOS scripts I use
Main Converter script - It will convert all files in all subdirectories of your current directory, unless a LRF file already exists, so it works for updates too:

@echo off
REM Set Defaults for all books.
set VARS=--disable-autorotation --header --left-margin=30 --right-margin=30
REM Set the Root for your book library
set ROOT=c:\EBooks\NewSrc

REM This variable can be changed in any directory with defaults.cmd
set DEFAULTVARS=

REM Add in any passed in options
set VARS=%VARS% %*
goto :main

:getdefaults
set curDir=%1
set curDir=%curDir:~1,-1%
:rinse
for %%j in ("%curDir%") DO set curDir=%%~dpj
set curDir=%curDir:~0,-1%
if exist "%curDir%\defaults.cmd" goto :found
if /i "%curDir%"=="%ROOT%" goto :notfound
goto :rinse
:found
call "%curDir%\defaults.cmd"
:notfound
goto :eof

:convert
set LIT=%1
set LRF=%LIT:~0,-4%lrf"
if exist %LRF% goto :eof
set OPTIONS=%LIT:~0,-4%cmd"
set DEFAULTVARS=
set EXTRAVARS=
set USEVARS=

call :getdefaults %LIT%

if exist %OPTIONS% (
   call %OPTIONS%
)

set USEVARS=%VARS% %DEFAULTVARS% %EXTRAVARS%

@echo Filename: %LIT% USEVARS : %USEVARS%

lit2lrf %USEVARS% -o %LRF% %LIT%

goto :eof


:main
for /R %%i in (*.lit) do (
   call :convert "%%i"
)

Defaults.cmd in my "root" of c:\ebooks\NewSRC, sets defaults for an entire tree (for instance all baen books):

SET DEFAULTVARS=--baen --force-page-break-before-tag=h1

And Wind_Riders_Oath.cmd (named the same as the .lit, and in the same directory)

SET EXTRAVARS=--serif-family="c:\windows\fonts, Times New Roman"

[kovidgoyal]

The --use-spine option is going to be implemented when I get a large enough block of time. Unfortunately it requires refactoring some of the internals of html2lrf, so it needs more time than the typical feature.

[WayneD]

Originally Posted by Xenophon

Would you please enter a ticket at libprs500.kovidgoyal.net with the necessary info so that kovid can fix this?

I had filed some bug info prior to my post here, and indeed, the bug that was causing vanishing scene breaks is now fixed in the svn version (and QA tested too ). Kovid has been very responsive to bug reports, which is great. I've only had my 505 for about a day now, but it's been really nice to be able to convert some of my already-purchased books, get them looking good, and really enjoy reading them.

[bkilian]

Originally Posted by kovidgoyal

The --use-spine option is going to be implemented when I get a large enough block of time. Unfortunately it requires refactoring some of the internals of html2lrf, so it needs more time than the typical feature.

Will --use-spine still generate the Table of Contents?
Sorry if I sounded grumpy about it, I'm not really, it just offends my sense of aesthetics. I'd like the generated LRF to look as similar as possible to the original LIT, and the TOC being shoehorned in at the beginning of the book violates that.
Your tool does an amazing job otherwise.

[kovidgoyal]

If the spine doesn't specify the TOC file as an element of the book then no it wont be generated.

[bkilian]

Originally Posted by kovidgoyal

If the spine doesn't specify the TOC file as an element of the book then no it wont be generated.

It isn't specified in the spine, it's specified as a reference:
<reference type="toc" title="Table of Contents" href="0671876422_top.htm" />
You could also use the <tour> section to generate a TOC.

[kovidgoyal]

Yeah then it wont be included.

[JSWolf]

But will it create an external ToC with --use-spine? Also, for LIT files where I have no control over the paragraph indent, I would like a switch for that so I can control how the paragraph indents look. Also, I'd like to be able to control the size of the chapter headings since they do come out rather large. This ia ll because Baen books do some odd things.

[kovidgoyal]

yeah it will. open a ticket requesting merging of an external CSS file. That way you can override BAEN css.

[JSWolf]

Originally Posted by kovidgoyal

yeah <this is to satisfy mobileread's message length criteria>

Is that a yeah for the ToC only of for the ToC plus all the other requests I just made?