There are definitely some major improvements to this Alpha version 0.4 of your epub reader - the biggest step I have noticed, but there are some pieces of text being 'lost' in the process of displaying them.

The Woman in White seems to have most if not all of the text displayed, but the paragraph formatting seems to have been compromised, and there are 'blocks' - [] that seem to indicate that the LF/CR character code is being displayed instead of a character 'space', but this is more of a niggle than anything else - just something to 'tidy up'.

Images seem to display though some text is being shown over some of the images, making it difficult to read some of the text, which is another 'concern' - but the images are shown in both colour and black and white and do not seem to show any image corruption/break-up, so that is another good improvement. This seems to happen when you change the font size after the 'epub' has been already loaded, if I am not mistaken

I downloaded the Classic English novel Alice in Wonderland by Lewis Carroll, but from Chapter 2 - The Pool of Tears - the section of text from the last part of the chapter has got 'lost'

" cats: nasty, low, vulgar things! Don't let me hear the name again!'

`I won't indeed!' said Alice, in a great hurry to change the subject of conversation. `Are you--are you fond--of--of dogs?' The Mouse did not answer, so Alice went on eagerly: `There is such a nice little dog near our house I should like to show you! A little bright-eyed terrier, you know, with oh, such long curly brown hair! And it'll fetch things when you throw them, and it'll sit up and beg for its dinner, and all sorts of things--I can't remember half of them--and it belongs to a farmer, you know, and he says it's so useful, it's worth a hundred pounds! He says it kills all the rats and--oh dear!' cried Alice in a sorrowful tone, `I'm afraid I've offended it again!' For the Mouse was swimming away from her as hard as it could go, and making quite a commotion in the pool as it went.

So she called softly after it, `Mouse dear! Do come back again, and we won't talk about cats or dogs either, if you don't like them!' When the Mouse heard this, it turned round and swam slowly back to her: its face was quite pale (with passion, Alice thought), and it said in a low trembling voice, `Let us get to the shore, and then I'll tell you my history, and you'll understand why it is I hate cats and dogs.'

It was high time to go, for the pool was getting quite crowded with the birds and animals that had fallen into it: there were a Duck and a Dodo, a Lory and an Eaglet, and several other curious creatures. Alice led the way, and the whole party swam to the shore."

This problems is likely to happen with other 'epub books' but I have not had time to check more pages on this subject as yet - just noriced it with this book, so thought I would mention it now.

The scrolling issue that has also been mentioned by defender I also noticed, as it would be great to be able to use the page down/up keys to move within that field/area of text.

The TOP and BOTTOM of the page still seem to be cut short, as the top (at the top of the page) and bottom (at the bottom of the page) of the letters/characters are being 'chopped off' and you cannot scroll to the very last line of text in some instances, which needs fixing.

Great improvement over the last version, and looking forward to the next one.

The "blocks" are not LF/CR. They are UTF-8 Chars with the hexnumber "0xE28093"
They are unprintable by cairo. They are special versions of the char "-".
I must filter them.

Image/Text mixing between pages are known. Its a bug .
But ia have minimal time to develope.

The cuting of top/bottom on pages are known bug.
The scrolling (The reason) is on my toDo List.






Quote:

Quote: OK, as long as you know about the problem, and have a solution that is great.
Quote: OK, but you are doing really well so far, and you seem to have progressed really quickly so far - well done.
Quote: OK, I hope you manage to fix/squash that bug soon

@pegasos-sigi2: your reader could support PDF file with hyperlink ?

APDF don't support hyperlink of WArMUp document PDF :(

Quote:

VPDF is the most recent PDF Viewer in MorphOS. APDF is outdated.

Geit

sorry, i would like write :

VPDF don't support hyperlink of WArMUp document PDF :(

(APDF support them but have bug graphic)



[ Edité par Papiosaur 15.09.2014 - 14:01 ]

Quote:

The developer will not develope... LibPoppler (The engine of VPDF) is very huge.
Lot of fuctions of libPoppler are not used in VPDF.
Why? I don't know.

Quote:

File a feature request then.

Geit

Quote:

The developer will not develope... LibPoppler (The engine of VPDF) is very huge.
Lot of fuctions of libPoppler are not used in VPDF.
Why? I don't know.

Quote:

I learned this the hard way my self, recently. It is probably (never used it myself) not Cairo having problems.

The problem gets triggered when you try to convert the UTF-8 to lets say ISO-8859-1. This "-" (0xE28093) is not a minus. It is a dash to bind words ("en dash").

There is no proper char available in ISO-8859 table and you would corrupt this information during conversion. The right way would be to render the text as UTF-8 or turn Cairo into UTF rendering, so it takes the char and displays it using the ttf font without the need to use any charset conversion in between.

Geit



[ Edited by geit 15.09.2014 - 20:12 ]

My weekly News: Epub-0.5

News:

+ A SVG (Salable vector grafic) - Engine is now includet
+ Display of all versions of svg (extern file or embed in the html source)
+ Better recognition of cover images
+ support of more HTML 4.01 Styles
+ Change slider to scrollbar in the mui gui
+ More default styles in the settings window
+ Support of this keys: (shift) cursor up /cursor down
+ The scrollbar is enabled to use with the scrollwheel on the mouse.

The ebooks from Ebookmaps are now
supported. There are "Maps in ebooks".
To display these "ebookmaps", i need a support of svg.

I coudn't work on it on this weekend: The myterius "blocks".
The "absolutly mad" structure of ebookmap-files eat lot of my free time.

The Screenshots of "ebookmaps":
ebookmaps.jpg
ebookmaps-1.jpg

[ Edited by pegasos-sigi2 21.09.2014 - 14:18 ]

Quote:

There is nothing mad about these ebookmap files. They are just normal epub files following the standards.

You really should not try to fix any book, but implement the standard properly and then - once it is working - go for the broken books like "Die Zukunft ist schön!" you brought up earlier, which is a really broken epub file with loads of errors in it. It is a perfekt example on how to not create ebooks.

It even looks like someone edited the files from a template by hand, as there are tons of unused dummy entries and wrong formatted (speaking of xhtml) lines to make the stuff more human readable, which violates the file format.

Do not use those files as template to build code on. You will fail. Use at least some more trusty stuff like wikipedia to get it done.

Geit

1.) If i implement the standart, i can't find any think in the ebookmap-files.
In the toc.nzx exists only one entry.
The whole file with the maps, was not listed.
It exists only in the listening of archive contents.
If i follow the standart, i found only the cover, not the maps.
The file is EPUB 2.0 so it must exist a NZX-File with the table of contents.
Only if a reader have a fall back, and it search in the "content.opf" ( != Standart),
i find the loosing file.



Quote:

Very well! another page (contens) of bartaemius 2 comes to front. On some pages more words (but still some missing).
Thank you for this nice update

Quote:

There is only one entry required in nav ncx to be valid. It is ment to "nav"igate, not to properly build books. Thats what spine (hence the name) is for, it defines the relevant files required for the backbone of the book.

The real way is using the manifest to grab all relevant files and types.
The spine section in OPF contains all files with relevant content to be displayed. In this case title page and html id references. These IDs represent the files in manifest having the same ids.

This will lead you to exactly two files:
titlepage.xhtml and ebookmap45.html

The latter contains all required entries which src argument leads to all the images. It is surrounded by tons of <div> sections, but thats what you get when dealing with html, xml and css. There is no need to cheat around to get the images as they are listed in manifest and reappear as image links in the main html file.

Note:

Just a note as I got time to test your tool. It reports all my epubs (which I created from src html file using calibre to prove-read them on my ipad ibook reader) as DRMed, which is impossible, as I created them.

No clue how you detect DRM, but an archive which contains human readable utf-8 files is unlikly DRMed.

I test this files with the xadmaster.library. If the lib can't unpack the files. Epub-Reader presume
they are DRM'd.
If the file corupt the xadmaster .librqary can't unpack the files,too.

Can i get a epub-file ?
Then i can test it.

Quote:

Quote:

I tried today and it opened all of them. Seems to be a glitch of some kind, but I know that other people had the same problem.

Geit

Weekly news: Epub-Reader_0.6

New:

+ No "blocks": Epub-Reader filter all unprintable chars. Special UTF8-Chars was moved in UTF8-Chartable to print them correctly.
+ "Real" use of Mousewheel. Now you can scroll the Ebook whith the wheel.
+ The Scrollbar is fixed: Now it scrolls exactly till the end of a page. Also it did't scroll, if a page is smaller then the display.
+ Two speed scrolling: LSHIFT + Mousewheele = Double scrollspeed.
+ Textwrapping arround pictures (see the screenshot)
+ Correct picturelayout (see screenshot)
+ HTML 4.01 Style Underline

Textwrapping arround pictures: Epub-Reader_0.6-1.jpg
No blocks: Epub-Reader_0.6-2.jpg
Correct picturelayout: Epub-Reader_0.6-3.jpg

This block filter stuff is not required.

The problem is that you convert the UTF8 text to iso8859 and then back to utf8 for displaying. During the first convert you are loosing the characters as iso8859 cannot keep all of them.

Simply only convert characters required to be displayed outside in the mui gui and hand over utf8 to cairo or ttengine. This will work fine without any side effects.

Geit

[ Edited by geit 28.09.2014 - 14:34 ]

Quote:

I convert nothing!
I use linear UTF8 all the time, no ISO 8859 or other thinks.
But any of these "blocks" are 0x0A and 0x0D and any of these \xe2\x80\x93.
0x0A and 0x0D are white-space control codes: '\f','\v','\n','\r'. These are not printable by cairo!
I must filter them.
0xE28093 are UTF8-Code for "-". Also not printable, but i change the charstream to ascii "-".

Quote:

Well, I know for sure that 0xE28093 is printable. At least tt_engine converts it right, as I got the very same issue, when using Text() and converted on my own, when implementing Reader.

http://www.geit.de/tmp/Reader.jpg

Geit

Hi Geit,

Does this mean you are also working, well basically have finished working on your own version of an epub reader?

If so, when will you be releasing your own 'Reader' version 1.0?

Quote:

It is basically working with any ebook and recently also supports amiga guide files.

Here a (incomplete) list of features already build in and working:

° epub support (No DRM of course)
° amigaguide support
° book like reading experience (no GUI elements required for reading)
° automatic justification (optional)
° automatic hyphenation (if a proper pattern file for the books language is present)(optional)
° dual page view
° multi window and tab support.
° app icon drop, window drop support
° project icon and shell support
° language weight algorithm to detect the books language if meta information missing
° browse pages via wheel, keyboard and mouse click (area based like on touch readers)
° fully customizsable view
° fully localized

And as a bonus it will be released with a free book and it comes with a so called recognition reader, which is basically a second reader application, with a unique new way of reading texts very fast.

Quote:

I consider it a beta and will not release beyond betatesters until all the known issues are fixed.

However, it is planed to deliver it with next MorphOS release as it may require some MorphOS 3.8 specific features, but I may consider releasing it earlier.

Geit

So we have two guys developing differnt software for the same purpose. Maybe some more people want to duplicate the work? That will be much better than having software for differnt purposes.
BTW why do you use this old ugly topaz font? MOS comes with so many better readable fonts. I was so happy to get rid of this old piece.

Quote:

In a halfe of my free time i port a ftp-server,too :))