Support EPUB generation #98
Comments
|
Having Pandoc support would also allow PDF output a la Pamflet. #14 |
|
If someone doesn't grab this one sooner, I'll take a look at this one on or around July 1. https://twitter.com/andyczerwonka/status/874814697752494080 |
|
I would suggest that this story is not about supporting Pandoc, it's about generating both PDF and ePub for offline reading. Pandoc is an implementation detail - does anyone disagree? Is there some other reason for the Pandoc dependency? |
|
@andyczerwonka any updates? |
|
@cleverdeveloper I pulled the repository a couple of days ago and started to look at the build process and where to inject the documentation steps. I should have more to share this weekend. |
|
Apologies to all, I won't have an update until I get back from vacation mid-August. :( |
|
All, shall we close this issues? At the moment this issue is highly dependent on #81 given there is discussion that takes PDF into account. |
I don't see a reason to close this issue, just because there's a related discussion happening elsewhere. We can close it when Paradox is able to generate a Pandoc-compatible single markdown file. In my opinion, it's more of a matter of String concatenation rather than markdown parser implementation. |
|
I think that's fair, but there is clearly a dependency. We may be able to solve the PDF generation depending on library that is chosen, which is really the ask. Pandoc is an implementation concern. |
|
ok. I can rename this issue to support EPUB generation. |
|
Just wondering if there's been any further thinking on an implementation approach to solve the use case of publishing paradox docs. I might attempt to solve this over the next couple of months for my own purposes. |
|
@andyczerwonka IMHO PDF and EPUB are different use cases. For Flexmark Java support of EPUB, see vsch/flexmark-java#192 |
|
Any updates on epub support? |
According to the [Java Language Specification](https://docs.oracle.com/javase/specs/jls/se8/html/index.html) member of packages, e.g. subpackages are identifiers. Identifiers itself is ...an unlimited-length sequence of Java letters and Java digits, the first of which must be a Java letter. As a `Java letter` includes more than what is expressed by the regex `\b[a-z][a-z0-9_]*)\.`, package names which include uppercase or non ASCII characters cannot be processed. I know that there are some _naming conventions_ flying around but something like `aBc.DE.fg` is still a valid package name in my opinion. I have changed the regex in `packageDotsToSlash()` to `(\b\p{javaJavaIdentifierStart}\p{javaJavaIdentifierPart}*)\.`. Unfortunately this breaks with the changes applied for issue lightbend#397, lightbend#395, lightbend#98, lightbend#86 to handle inner classes. In order to support this notation, the regex can be relaxed to `(\b\p{javaLowerCase}\p{javaJavaIdentifierPart}*)\.`. I have added the variable 'scaladoc.strictPackageIdent' (which defaults to `false`) by which you can switch between the two regexp. Once set to true, the package name can conform to an identifier now.
According to the [Java Language Specification](https://docs.oracle.com/javase/specs/jls/se8/html/index.html) member of packages, e.g. subpackages are identifiers. Identifiers itself is ...an unlimited-length sequence of Java letters and Java digits, the first of which must be a Java letter. As a `Java letter` includes more than what is expressed by the regex `\b[a-z][a-z0-9_]*)\.`, package names which include uppercase or non ASCII characters cannot be processed. I know that there are some _naming conventions_ flying around but something like `aBc.DE.fg` is still a valid package name in my opinion. I have changed the regex in `packageDotsToSlash()` to `(\b\p{javaJavaIdentifierStart}\p{javaJavaIdentifierPart}*)\.`. Unfortunately this breaks with the changes applied for issue lightbend#397, lightbend#395, lightbend#98, lightbend#86 to handle inner classes. In order to support this notation, the regex can be relaxed to `(\b\p{javaLowerCase}\p{javaJavaIdentifierPart}*)\.`. I have added the variable 'scaladoc.strictPackageIdent' (which defaults to `false`) by which you can switch between the two regexp. Once set to true, the package name can conform to an identifier now.
According to the [Java Language Specification](https://docs.oracle.com/javase/specs/jls/se8/html/index.html) member of packages, e.g. subpackages are identifiers. Identifiers itself is ...an unlimited-length sequence of Java letters and Java digits, the first of which must be a Java letter. As a `Java letter` includes more than what is expressed by the regex `\b[a-z][a-z0-9_]*)\.`, package names which include uppercase or non ASCII characters cannot be processed. I know that there are some _naming conventions_ flying around but something like `aBc.DE.fg` is still a valid package name in my opinion. I have changed the regex in `packageDotsToSlash()` to `(\b\p{javaJavaIdentifierStart}\p{javaJavaIdentifierPart}*)\.`. Unfortunately this breaks with the changes applied for issue #397, #395, #98, #86 to handle inner classes. In order to support this notation, the regex can be relaxed to `(\b\p{javaLowerCase}\p{javaJavaIdentifierPart}*)\.`. I have added the variable 'scaladoc.strictPackageIdent' (which defaults to `false`) by which you can switch between the two regexp. Once set to true, the package name can conform to an identifier now.
See http://pandoc.org/MANUAL.html
Why? Generate epub
The text was updated successfully, but these errors were encountered: