calibre 5: Beta testers needed

[kovidgoyal]

Originally Posted by jony08

The shortcut Ctrl+Enter used to apply the Add notes window is used by most browsers to select the URL field. Could the shortcut be changed to something else in the Content server?

Ctrl+enter does not select the url field in either chrome or firefox for me.

[jony08]

Originally Posted by kovidgoyal

Ctrl+enter does not select the url field in either chrome or firefox for me.

I don't know how that happened to me earlier.

[nixsee]

I like the new chapter sections for highlights. But I wonder if it is possible to prepend the name of the parent sections to the name, or (even better) have the chapters nested within collapsible parent sections - as is the case with the Table of Contents.

Here's a few highlights that I have from one book that is organized into 4 Parts, each with dozens of chapters. Other than the order there's no indication to know which Part the chapter is from.

[kovidgoyal]

@nixsee yes there is, hover your mouse over the chapter name.

[nixsee]

Ah, great! I didn't realize that was there. Thanks very much for all the hard work and openness to feedback from the peanut gallery!

[nixsee]

I found a bug - highlights aren't always being attributed to the correct chapter. It seems to have to do with the viewer thinking that as soon as the next chapter's title is within view that you are in that chapter. I assume this only applies to Flow Mode, and even then only applies to certain books (as most books seem to only flow until the end of a chapter, whereas this one doesn't do that - I assume there is something off with the chapter tags, but don't know how to (or even want to) fix this for every book that might have this problem.

Here are highlights when the next chapter is not in view.




And when you scroll down so that it is in view (note Chpater 12 being bolded in the TOC).



Even stranger: when I add a new highlight above the previous ones, it registers in chapter 12 because the viewer thinks I'm in that chapter now.

[chaley]

Originally Posted by davidfor

Good to know. What about "raw_field()"? I probably use that as much as "field()".

I rewrote the formatter to build and save an interpretable tree instead of always interpreting while parsing.

Other changes: it

• inlines raw_field
• automatically converts test() to if-then-else, avoiding interpreting the non-used expression (branch).
• improves syntax checking. For example, now that errant semicolon will be signaled regardless of the value of the conditional expression

These changes dramatically improve performance of composite columns and icon/color rules.

[davidfor]

Originally Posted by chaley

I rewrote the formatter to build and save an interpretable tree instead of always interpreting while parsing.

Other changes: it

• inlines raw_field
• automatically converts test() to if-then-else, avoiding interpreting the non-used expression (branch).
• improves syntax checking. For example, now that errant semicolon will be signaled regardless of the value of the conditional expression

These changes dramatically improve performance of composite columns and icon/color rules.

That sounds very good. Unfortunately, once I updated my source, I got:

Traceback (most recent call last):
  File "E:\Development\GitHub\calibre\src\calibre\utils\formatter.py", line 633, in safe_format
  File "E:\Development\GitHub\calibre\src\calibre\utils\formatter.py", line 598, in evaluate
  File "E:\Development\GitHub\calibre\src\calibre\utils\formatter.py", line 522, in _eval_program
  File "E:\Development\GitHub\calibre\src\calibre\utils\formatter.py", line 319, in program
  File "E:\Development\GitHub\calibre\src\calibre\utils\formatter.py", line 324, in expression_list
  File "E:\Development\GitHub\calibre\src\calibre\utils\formatter.py", line 424, in expr
  File "E:\Development\GitHub\calibre\src\calibre\utils\formatter.py", line 422, in expr
  File "E:\Development\GitHub\calibre\src\calibre\utils\formatter.py", line 348, in do_node_if
  File "E:\Development\GitHub\calibre\src\calibre\utils\formatter.py", line 424, in expr
  File "E:\Development\GitHub\calibre\src\calibre\utils\formatter.py", line 422, in expr
  File "E:\Development\GitHub\calibre\src\calibre\utils\formatter.py", line 365, in do_node_func
  File "E:\Development\GitHub\calibre\src\calibre\utils\formatter.py", line 424, in expr
  File "E:\Development\GitHub\calibre\src\calibre\utils\formatter.py", line 422, in expr
  File "E:\Development\GitHub\calibre\src\calibre\utils\formatter.py", line 400, in do_node_raw_field
  File "E:\Development\GitHub\calibre\src\calibre\utils\formatter.py", line 392, in do_node_raw_field
  File "E:\Development\GitHub\calibre\src\calibre\utils\formatter.py", line 311, in error
ValueError: Interpreter: Unknown field #pages
Error evaluating column named: #pages1777625035779470211

This appears to be for some column colouring rules using my "Pages" column which has the lookup name "#pages". These re just comparisons to a value. And I have other rules using the column 'Words'/#words. The errors are identical except that the number after the lookup name changes.

And for a test column, using the templates:

{:'raw_field('#pages')'}

program:
raw_field('#words')

The latter gives:

Traceback (most recent call last):
  File "E:\Development\GitHub\calibre\src\calibre\utils\formatter.py", line 633, in safe_format
  File "E:\Development\GitHub\calibre\src\calibre\utils\formatter.py", line 598, in evaluate
  File "E:\Development\GitHub\calibre\src\calibre\utils\formatter.py", line 522, in _eval_program
  File "E:\Development\GitHub\calibre\src\calibre\utils\formatter.py", line 319, in program
  File "E:\Development\GitHub\calibre\src\calibre\utils\formatter.py", line 324, in expression_list
  File "E:\Development\GitHub\calibre\src\calibre\utils\formatter.py", line 424, in expr
  File "E:\Development\GitHub\calibre\src\calibre\utils\formatter.py", line 422, in expr
  File "E:\Development\GitHub\calibre\src\calibre\utils\formatter.py", line 400, in do_node_raw_field
  File "E:\Development\GitHub\calibre\src\calibre\utils\formatter.py", line 392, in do_node_raw_field
  File "E:\Development\GitHub\calibre\src\calibre\utils\formatter.py", line 311, in error
ValueError: Interpreter: Unknown field #pages
Error evaluating column named: #story_length

[kovidgoyal]

Released version 4.99.16 with the following changes:

- Reduced indentation in tree views
- The highlights panel in the content server viewer now also groups by chapter and shows color swatches
- Fix momentum scrolling not working in the cover grid
- Workaround for Firefox bug that broke searching backwards across file boundaries
- More accurate detection of current toc section when creating highlights
- Improvements to the template parser as detailed by chaley

[chaley]

Originally Posted by davidfor

And for a test column, using the templates:

{:'raw_field('#pages')'}

program:
raw_field('#words')

The latter gives:

Traceback (most recent call last):
  File "E:\Development\GitHub\calibre\src\calibre\utils\formatter.py", line 633, in safe_format
  File "E:\Development\GitHub\calibre\src\calibre\utils\formatter.py", line 598, in evaluate
  File "E:\Development\GitHub\calibre\src\calibre\utils\formatter.py", line 522, in _eval_program
  File "E:\Development\GitHub\calibre\src\calibre\utils\formatter.py", line 319, in program
  File "E:\Development\GitHub\calibre\src\calibre\utils\formatter.py", line 324, in expression_list
  File "E:\Development\GitHub\calibre\src\calibre\utils\formatter.py", line 424, in expr
  File "E:\Development\GitHub\calibre\src\calibre\utils\formatter.py", line 422, in expr
  File "E:\Development\GitHub\calibre\src\calibre\utils\formatter.py", line 400, in do_node_raw_field
  File "E:\Development\GitHub\calibre\src\calibre\utils\formatter.py", line 392, in do_node_raw_field
  File "E:\Development\GitHub\calibre\src\calibre\utils\formatter.py", line 311, in error
ValueError: Interpreter: Unknown field #pages
Error evaluating column named: #story_length

The first one in TPM works?

Neither fail for me using an integer column. Would you be willing to send me your metadata.db so I can debug this? If so, send it to calibre at charleshaley dot org.

EDIT: never mind. I installed the newest version of the calibre binary and it fails for me too. Don't know why yet.

[jony08]

Originally Posted by jony08

IMHO it would be useful to be able to edit the styles in the Add notes window and also to be able to delete the styles that are available from the beginning.

@kovid: Would it be possible to add the ability to edit styles?

Another small thing would be that when you open the search page on the Content Server, the search field should be selected automatically. Is it possible?

[kovidgoyal]

Originally Posted by jony08

@kovid: Would it be possible to add the ability to edit styles?

Sorry, no.

[jony08]

The annotations are not synchronized with the main view from the server and vice versa. Am I doing something wrong? I tried clicking on the Sync button in the controls.

[davidfor]

Originally Posted by chaley

The first one in TPM works?

It failed as well. The traceback was common enough that I decided to be lazy. And was to lazy to mention that

Neither fail for me using an integer column. Would you be willing to send me your metadata.db so I can debug this? If so, send it to calibre at charleshaley dot org.

EDIT: never mind. I installed the newest version of the calibre binary and it fails for me too. Don't know why yet.

I did the same just in case it need a proper install.

But, it looks like it fails if there isn't a value in the column. This was with a test library and there were a lot of books without either the word or page counts. When I set them for a book, the value in the template column was displayed correctly for the book.

[chaley]

Originally Posted by davidfor

It failed as well. The traceback was common enough that I decided to be lazy. And was to lazy to mention that

Found it, I think. The inlined raw_field didn't handle None-valued fields properly.

If you are willing, try using this code:

    def do_node_raw_field(self, prog):
        try:
            name = self.expr(prog.expression)
            res = getattr(self.parent_book, name, None)
            if res is not None:
                if isinstance(res, list):
                    fm = self.parent_book.metadata_for_field(name)
                    if fm is None:
                        return ', '.join(res)
                    return fm['is_multiple']['list_to_ui'].join(res)
            return unicode_type(res)
        except ValueError as e:
            raise e
        except:
            self.error(_('Unknown field {0}').format('parse error'))

It goes in utils.formatter.py, replacing the existing method starting at line 387.

EDIT: Kovid accepted the change. It is in source now.