regex question

[DrChiper]

The new regex-function works great, the section numbering was just something what I needed. However, I stumbled onto some expressions which were not handled properly:

<h1 class="center"><a href="chapter001.xhtml#start">Test</a></h1>
<h1 class="center" id="" ><span class="something">Test</span></h1>
<h1 class="center" id="" ><b><span class="something">Test</span></b></h1>

The default find construct is not enough:

(<h1[^<>]*>)([^<>]+</h1>)

I tried to enhance it but either got no match at all, or, very greedy, it selected almost everything

Can somebody teach me how to select "just enough"?

[kovidgoyal]

(<h1[^<>]*>)(.+?</h1>)

[DrChiper]

I was sooo close ...

Anyway, thanks Kovid

Humble suggestion: Update the manual with this version as I'm afraid the examples I provided will likely to be encountered by many more users.

[kovidgoyal]

Well, this expression will not match if the heading tag is split over multiple lines, for that you need:

(?s)(<h2[^<>]*>)(.+?</h2>)

[DrChiper]

Indeed! Although not very common, constructs like this do happen:

  <h1 class="center"><a href="Section003.xhtml#start">Test</a><br/>

  Some sub header</h1>

Nice improvement, Kovid!

[theducks]

Originally Posted by DrChiper

Indeed! Although not very common, constructs like this do happen:

  <h1 class="center"><a href="Section003.xhtml#start">Test</a><br/>

  Some sub header</h1>

Nice improvement, Kovid!

I think it is common
I use a similar one a Lot when I want the Chapter Title to appear in the TOC

<h2 class="CN">- 1 - <br /><span class="CT">A New Year</span></h2>

[Paulie_D]

Originally Posted by DrChiper

Indeed! Although not very common, constructs like this do happen:

  <h1 class="center"><a href="Section003.xhtml#start">Test</a><br/>

  Some sub header</h1>

As a 'coder'of sorts, (primarily web) I would tell anyone that using the break tag for spacing is poor practise.

I'm not saying it doesn't have it's place but using it for spacing is not recommended. That's what margins/padding are for.

This

  <h1 class="center"><a href="Section003.xhtml#start">Test</a><br/>

  Some sub header</h1>

Should be replaced with this

  <h1 class="center"><a href="Section003.xhtml#start">Test</a></h1>

  <h2 class="center">Some sub header</h2>

It's certainly more semantic.

Whether this has an specific relevance to e-publishing is perhaps questionable but 'poor' code is something to be noted even if we choose to use it for own own reasons.

Understand I'm not running anyone down for it...just expressing one man's opinion.

[theducks]

Originally Posted by Paulie_D

As a 'coder'of sorts, (primarily web) I would tell anyone that using the break tag for spacing is poor practise.

I'm not saying it doesn't have it's place but using it for spacing is not recommended. That's what margins/padding are for.

This

  <h1 class="center"><a href="Section003.xhtml#start">Test</a><br/>

  Some sub header</h1>

Should be replaced with this

  <h1 class="center"><a href="Section003.xhtml#start">Test</a></h1>

  <h2 class="center">Some sub header</h2>

It's certainly more semantic.

Whether this has an specific relevance to e-publishing is perhaps questionable but 'poor' code is something to be noted even if we choose to use it for own own reasons.

Understand I'm not running anyone down for it...just expressing one man's opinion.

It may be 'proper' coding, bet it fails the Auto-TOC job

Test

Some Sub header

(both are essentially the same location)

is not what is desired in a TOC, especially on a smaller screen

[Paulie_D]

Originally Posted by theducks

It may be 'proper' coding, bet it fails the Auto-TOC job

Test

Some Sub header

(both are essentially the same location)

is not what is desired in a TOC, especially on a smaller screen

Oh...I agree...if you know you're breaking the "rules" for a specific reason then it's not a biggie but I do think it's worth noting.

That said, couldn't a function be written to combine two distinct elements into a single TOC item?

Dunno, just asking..I'm still a novice at regex.

Regardless, the TOC 'link' would still be the same point (the H1) so you'd still arrive at the same place wouldn't you?

[DiapDealer]

In this case, the br tag is not being used to achieve any kind of spacing or padding. It's being used as it was intended--to simply make sure the text that follows starts on a new line.

[DrChiper]

@Paulie_D: It is what I see in epubs, not what I have created myself. Sometimes you have to deal with what you have, and that is almost never what you really want. But I mostly can live with it given it just appears acceptable enough on my trusted e-reader

(And if not, then there is always calibre ... )

[Paulie_D]

Please...I wasn't suggesting anyone here was doing anything wrong here just trying to make a general point.

Guess I did it badly.

[signum]

Regarding multi-line titles and auto TOC:

My preference is to have smaller-sized subtitles, to clearly point them out as subordinate to the main title. So, a typical set of headings would look like:

<h2>Main heading</h2>
<h3>Subtitle</h3>

where h1, h2, h3 and h4 are styled with at least a "text-align: center". They can all share the same css block. The magic comes when setting this up for an auto TOC generator. The key is to use the "title=" specification. Using the above example, the headings then look like:

<h2 title="Main heading: Subtitle">Main heading</h2>
<h3 class="sigil_not_in_toc">Subtitle</h3>

The auto TOC generator will use what's in the "title=" in preference to what is in between the ">stuff<" angle bars. BTW, it's been my experience that extra html such as "< br / >" in the "title=" gets stripped out by the auto TOC generator. H2 and h3 can be anything you like, as long as you get the "title=" right. The stuff between the angle bars gets displayed as desired and the "title=" only shows up in the TOC.

[DiapDealer]

Originally Posted by Paulie_D

Please...I wasn't suggesting anyone here was doing anything wrong here just trying to make a general point.

Guess I did it badly.

Not really. I didn't mean to sound harsh, I just thought your assumption that the br in the given example(s) was about creating space was inaccurate.

Yes, using br tags to create spacing between paragraphs or different sections is considered bad form. But surely, using the br tag to force a line break is why the br tag exists in the first place, no?

[Paulie_D]

Originally Posted by DiapDealer

But surely, using the br tag to force a line break is why the br tag exists in the first place, no?

Yes...provided the intention is that the whole element is intended to be read as a single item.

So in your example, it's correct because the number and title are intended to be a single item but for visual purposes it needs to look like two lines.

In the other example a major heading and sub-heading would not be intended to be a single item and so "should" be two items.