Forum: Author Hangout

ams once against asked to login

Log out and log back in.

like paragraph indenting and line spacing

Those are styling in CSS, not html tags.

The builder allows you to have your own CSS file embedded into the EPUB if you wish, so you can change the style used.

Does it support svg images (there where svg images embeded in the docx)

It doesn't support any type of images embedded in .docx files.

And no, SVG files aren't supported (yet).

It returned with no error message at all?

I tried the files you emailed me. Worked first attempt:

https://bookapy.com/tmp/epub-builder/par-del-lalb-do-j.-p.-de-hortius.epub

This is frustrating. I can't replicate any bugs and yet so far it hasn't worked for anybody else.

Thank you for trying.

Maybe it's important, but what's your system's configuration? System version, browser and browser version?

Also tried edge and tried to leave out the dop.
Same problem remain.

Can you please give it another try. I added extensive input data logging to see what's being received. It could be the date. My own browser shows 2023-02-09 format and that's what the builder expects. Previous experience with dates show that this is always the case, but who knows.

Thank you for the help.

Well, I found the reason.

Somehow the same code that works perfectly in Safari, doesn't work in the latest version of Firefox.

I updated Firefox to the latest version and it's stopping the form from submitting. So the form wasn't submitting the form at all except on Safari (desktop and mobile) but nothing else.

I have to rework the javascript validation code.

If you do not see my test in your logs, that means that you are confusing GET and POST server side.

No, I'm not. It's not my first script you know. This has been weird. The javascript validation intercept is blocking Firefox and Chrome from submitting the form. Nothing was reaching the server, trust me.

Alright, new code is up. I tested it with Firefox and it worked. Did not test it with Chrome, but it should work too.

Please give it a try.

🤞

With firefox, I can see the multipart posted data, it is there. And the server do respond with a 200 to the posted data.

Maybe it seems to be working, but somehow the request isn't being delivered to the server.

I inserted code to log whatever data was being posted to the server at the start of the script. So the flow, once the upload button was clicked and the server received the data, before doing anything, it would log all the data received to a log file. Nothing is being logged at all with Firefox and Chrome. As if the request didn't reach the server.

I'm working on some new code. I'll post when there is something new to work with.

Not sure if it is the normal save target, but that should be protected by the file owner login. Currently, it is public no login required.

Not important. Once built, the system give you a button to download the file. Once the file is downloaded the file gets erased.

I was testing it, and after putting in a text file and image, it would just sit there and output nothing that I could see.

How big were the files?

If possible, can you email me the same file that resulted in nothing? (lazeez@storiesonline.net)

I downloaded the text file and your cover and tried. Worked as expected, first attempt (working from a particularly slow internet connection too):

https://bookapy.com/tmp/epub-builder/bohica-mushroom.epub

I don't know what you guys are doing differently to end up with different results (not working at all).

Thank you for trying.

Maybe it's important, but what's your system's configuration? System version, browser and browser version?

Post 1 with docx

Fatal issue 1: file lisence.xhtml is not valid xhtml (missing [/div] tag)

Fixed.

Warning Issue 2: use of & r s q u o ; in end.xhtml and & a s y m p ; in info.xhtml

Not really issues, but fixed.

I trust the command line epubcheck over calibre considering how many invalid EPUBs I receive that were generated by Calibre. epubcheck never flagged those as issues.

Issue 3:

CDATA around style in comment -> empty comment warning. I'm not sure this one is legit.

Previous issues remains, insertion of {x}, no TOC, etc.

Empty comment, not an issue.

No TOC is documented, for now, one has to code the TOC in the text. Later I'll add a checkbox that when checked will have the builder try to find 'chapter xx' in the text and 'part xx'.

insertion of {x}? I don't know what you mean by that. Please refresh my memory.

Handling of 200 response with error is incorrect.
As the response doesn't replace the submit page
response contained:
[h4 class="error"]Unsupported Book Text File format[/h4]
Along with a form copy.

Yeah, artifact of the change in code to submit the form. I'll fix.

How does it detect that the document submited is Markdown?

The code checks its mime type. What file extension did your file have?

.md is detected by Safari and Firefox as text/markdown.

MIME type would be the problem, Windows 10 doesn't know the MIME type for .md:

Updated builder to detect Markdown files regardless of mime type support of underlying system.

Also fixed the issue with error reports.

To-do: Table of contents detection.

but that is probably hard to implement.

Due to the various options that each submission format uses for header, it's tricky yes. I took the easy way out and made it up to the author for now.

Word, due to language support offers the biggest challenge.

I will add the check box. I would never make it the default for the software to assume that there is a TOC. If the author ticks the box, then the software will try to figure it out.