Bug in index creation with Sigil

Patrick V · 02-10-2020, 12:09 PM

Hi,

I found a problem in indexing in Sigil. The terms I need to index are inside footnotes, more precisely in <a></a>

To reproduce the problem, consider a simple epub page with those lines:

<body>
<a id="here">This is not a pipe</a>

<a id="there">This is not a pipe (but) </a>

</body>

The lines differ by the presence of parenthesis in the second .

If you mark a term in the second line, it will not appear in the index because of the parenthesis.

Same problem with brackets.

I use Sigil 0.9.7 in Linux, but I reproduced the problem in latest version on Windows.

Regards,

Patrick

KevinH · 02-10-2020, 12:33 PM

Did you mark the parenthesis or brackets too or just the string "but"? Or does this happen with any word being selected?

If only with the bracketed word, have you tried inserting a single space between "but" and the surrounding parentheses or brackets?

I will look into it when I get back to my dev machine.

Patrick V · 02-10-2020, 01:07 PM

I marked a word outside the parenthesis. But if there is a parenthesis everywhere within the <a> </a>, the word is not indexed.

No error is reported when you mark a word in a sentence with parenthesis or when you create the index.

I did not try to mark the word inside the parenthesis.

KevinH · 02-10-2020, 01:30 PM

Okay, this is caused by the () or [] being interpreted as QRegularExpression special characters instead of normal characters. But since QRegularExpressions are also used to build indexes, we have to be careful how we fix this without breaking things.

Thank you for your bug report, and testcase.

KevinH · 02-10-2020, 02:07 PM

The fix for this will appear in the next release.

Patrick V · 02-11-2020, 03:08 AM

Another problem I discovered. If you want to mark a word in this block:

<a class="myclass">This is not a pipe</a>

The class "sigil_index_marker" will replace "myclass". It'd be better to add "sigil_index_marker" alongside the existing class.

Patrick

KevinH · 02-11-2020, 09:12 AM

Yes, this problem only happens if the text you mark for index inclusion's immediate parent is already an anchor and one that has an existing class. The more general case typically causes a new anchor to be inserted and so no class collision is possible.

This should now be fixed in master.

Again, thank you for your bug report.

02-10-2020, 12:09 PM	#1
Patrick V Junior Member Posts: 8 Karma: 10 Join Date: Sep 2019 Device: Kobo Aura H2O	Bug in index creation with Sigil Hi, I found a problem in indexing in Sigil. The terms I need to index are inside footnotes, more precisely in <a></a> To reproduce the problem, consider a simple epub page with those lines: <body> <p><a id="here">This is not a pipe</a></p> <p><a id="there">This is not a pipe (but) </a></p> </body> The lines differ by the presence of parenthesis in the second <p>. If you mark a term in the second line, it will not appear in the index because of the parenthesis. Same problem with brackets. I use Sigil 0.9.7 in Linux, but I reproduced the problem in latest version on Windows. Regards, Patrick

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Bug in the creation of the summary on Kindle	Pyanepsion	Conversion	0	08-29-2018 03:42 AM
[BUG] Index lists all words starting with string	1v4n0	Sigil	9	05-17-2018 01:28 PM
Change index split names after epub creation?	nws	ePub	2	04-08-2015 02:54 AM
New bug on table of contents creation?	saxondawg	Conversion	3	06-22-2011 12:34 AM
Automatic index links creation in mobipocket	ragdoll	Kindle Formats	1	02-08-2008 07:07 AM

02-10-2020, 12:33 PM	#2
KevinH Sigil Developer Posts: 8,759 Karma: 5706256 Join Date: Nov 2009 Device: many	Did you mark the parenthesis or brackets too or just the string "but"? Or does this happen with any word being selected? If only with the bracketed word, have you tried inserting a single space between "but" and the surrounding parentheses or brackets? I will look into it when I get back to my dev machine.

02-10-2020, 01:07 PM	#3
Patrick V Junior Member Posts: 8 Karma: 10 Join Date: Sep 2019 Device: Kobo Aura H2O	I marked a word outside the parenthesis. But if there is a parenthesis everywhere within the <a> </a>, the word is not indexed. No error is reported when you mark a word in a sentence with parenthesis or when you create the index. I did not try to mark the word inside the parenthesis.

02-10-2020, 01:30 PM	#4
KevinH Sigil Developer Posts: 8,759 Karma: 5706256 Join Date: Nov 2009 Device: many	Okay, this is caused by the () or [] being interpreted as QRegularExpression special characters instead of normal characters. But since QRegularExpressions are also used to build indexes, we have to be careful how we fix this without breaking things. Thank you for your bug report, and testcase.

02-10-2020, 02:07 PM	#5
KevinH Sigil Developer Posts: 8,759 Karma: 5706256 Join Date: Nov 2009 Device: many	The fix for this will appear in the next release.

02-11-2020, 03:08 AM	#6
Patrick V Junior Member Posts: 8 Karma: 10 Join Date: Sep 2019 Device: Kobo Aura H2O	Another problem I discovered. If you want to mark a word in this block: <a class="myclass">This is not a pipe</a> The class "sigil_index_marker" will replace "myclass". It'd be better to add "sigil_index_marker" alongside the existing class. Patrick

02-11-2020, 09:12 AM	#7
KevinH Sigil Developer Posts: 8,759 Karma: 5706256 Join Date: Nov 2009 Device: many	Yes, this problem only happens if the text you mark for index inclusion's immediate parent is already an anchor and one that has an existing class. The more general case typically causes a new anchor to be inserted and so no class collision is possible. This should now be fixed in master. Again, thank you for your bug report.

Advert

Advert