Page 1 of 1

V.3.8.1 adds extra blank lines when clipping Web pages

Posted: Mon Jul 10, 2017 10:55 pm
by Clairvaux
On some Web sites, v.3.8.1 adds 2 extra blank lines between paragraphs when clipping from Web pages. This happened to me on several sites.

Previously, CN respected Web pages text layout : 1 blank line on the Web = 1 blank line in a Cinta note.

Does not happen on all sites, though.

Some pages generating those extra blank lines :

http://www.lemonde.fr/planete/article/2 ... _3244.html

http://www.eurozine.com/germans-must-re ... r-own-sake

https://en.wikipedia.org/wiki/Cat

http://tass.com/politics/955619

Some pages producing normal results :

http://www.lefigaro.fr/flash-actu/2017/ ... iviles.php

https://www.nytimes.com/2017/07/10/opin ... manov.html

Re: V.3.8.1 adds extra blank lines when clipping Web pages

Posted: Tue Jul 11, 2017 10:27 am
by CintaNotes Developer
Thanks for the report!

However, I can't reliably reproduce. On the pages that you list as "producing extra blank lines", for me only EUROZINE produces an unnecessary blank line, that is explainable by a non-breaking space that has been put into the middle line.

Could you please add some screenshots? Thanks!

Re: V.3.8.1 adds extra blank lines when clipping Web pages

Posted: Tue Jul 11, 2017 2:58 pm
by Clairvaux
Hi Alex,

Here are screenshots of the clipped notes with v.3.8.1 and v.3.8.

In fact, only one of those Web pages clips correctly : Le Figaro. The New York Times does get extra blank lines too, only I had not noticed them because you have to scroll down to reach them.

I qualify my original post by saying that I did not inspect the Web pages html code, so I can't say for sure how many blank lines are actually there. What I mean is, the previous versions correctly interpreted the Web designers' intentions, and the reader's perception (those are two separate paragraphs here, the Web site put some white space between them => signal this by putting one blank line in the note).

Now it seems to be unpredictable. Temporarily reverting to v.3.8.

Just in case it is relevant, I use Firefox with No Script.

New version

Eurozine v.3.8.1.PNG
http://www85.zippyshare.com/v/ao9776WZ/file.html

Figaro Nice - v3.8.1.PNG
http://www85.zippyshare.com/v/aCTgmlfh/file.html

Le Monde ONU - v.3.8.1.PNG
http://www85.zippyshare.com/v/pApqw4Tx/file.html

NYT Romanovs - v.3.8.1.PNG
http://www85.zippyshare.com/v/fpzWddWW/file.html

Tass Syria - v.3.8.1.PNG
http://www85.zippyshare.com/v/SNY4rlnx/file.html

Wikipedia cat - v.3.8.1.PNG
http://www85.zippyshare.com/v/VNWoFB6L/file.html

Previous version

Eurozine v.3.8.PNG
http://www85.zippyshare.com/v/nsIKb0Vb/file.html

Figaro Nice - v3.8.PNG
http://www85.zippyshare.com/v/GygnaX7n/file.html

Le Monde ONU - v.3.8.PNG
http://www85.zippyshare.com/v/Jtrx8vrd/file.html

NYT Romanovs - v.3.8.PNG
http://www85.zippyshare.com/v/fucnXocX/file.html

Tass Syria - v.3.8.PNG
http://www85.zippyshare.com/v/EbDie7Wb/file.html

Wikipedia cat - v.3.8.PNG
http://www85.zippyshare.com/v/YWL5FeS7/file.html

Edit : is the attachment function broken ? It's the second time I try to upload some files with a post and this does not work. What happens is the green upload bars show up, something gets uploaded because the router light is flashing, the list of files shows up below the post, but there's a yellow triangle in the status column next to the files names. After submitting the post, the files disappear into thin air. Tried Firefox and Opera, to no avail.

Re: V.3.8.1 adds extra blank lines when clipping Web pages

Posted: Thu Jul 13, 2017 7:23 am
by CintaNotes Developer
Thanks for the screenshots!

I have different results when I clip from same pages, so probably you're using a different browser than me.
Please tell which one! This is very browser dependent because each browser puts stuff into clipboard in its own unique way.

About attachments on the forum: Indeed, something is broken. Thanks for drawing my attention to it. I'll look into it asap.

Re: V.3.8.1 adds extra blank lines when clipping Web pages

Posted: Thu Jul 13, 2017 7:28 am
by CintaNotes Developer
Please try now - attachments should be working!

Re: V.3.8.1 adds extra blank lines when clipping Web pages

Posted: Thu Jul 13, 2017 8:49 am
by Clairvaux
Firefox 54.0.1, 64-bit. With No Script and Adblock Plus.

Just checking :

Re: V.3.8.1 adds extra blank lines when clipping Web pages

Posted: Fri Jul 14, 2017 7:45 am
by CintaNotes Developer
Nice bear!)

Re: V.3.8.1 adds extra blank lines when clipping Web pages

Posted: Wed Jul 19, 2017 10:39 am
by CintaNotes Developer
Hi,

could you please try this version?
Should fix the Firefox issues.

CintaNotes_3_8_2_Beta1.zip
(4.4 MiB) Downloaded 129 times

Re: V.3.8.1 adds extra blank lines when clipping Web pages

Posted: Wed Jul 19, 2017 12:35 pm
by Clairvaux
Hi Alex,

I've tried this new version, and all the pages above now clip fine, plus a few extra ones I also used it on.

I also tried it with Opera v.46.0.2597.57 (which I did not experiment with CN v.3.8.1), and in one case (http://www.eurozine.com/germans-must-re ... r-own-sake) it did a slightly better job than Firefox, not adding extra white lines at the top of the note (see attachments).

That's not really a problem however, as those extra white lines seem to be generated by the Twitter and Facebook buttons, so getting rid of them is part of the expected bit of cleaning work that occurs after clipping (if you want your note to be tidy and nice).

The problem with CN v.3.8.1 was it added white lines in totally unexpected places and wasn't as good as previous versions in that respect, which now seems to be corrected.

Kind regards,

Clairvaux

Re: V.3.8.1 adds extra blank lines when clipping Web pages

Posted: Thu Jul 20, 2017 3:01 am
by CintaNotes Developer
Thanks for taking time to try the new version out!
Well the main point was to get rid of the unexpected empty lines inside text blocks, which seems that now is achieved.
I'll take a look what can be done to further clear up the article beginning in FF, but it might be that the structure of HTML won't allow CN to make bold assumptions to clean up. For example, it just sees a DIV with an IMG inside, it can't possibly know that this is a Twitter button that can be stripped out.

I take it so that this version is definitely better than 3.8.1, the last question is - is it better than 3.8?

Re: V.3.8.1 adds extra blank lines when clipping Web pages

Posted: Thu Jul 20, 2017 11:33 am
by Clairvaux
Thank you for correcting this so fast.

Indeed, the nice thing about CN is it makes correct assumptions most of the time. I suppose assumptions is the best that can be hoped for when converting something as complex as a webpage to rich text, as opposed to a graphic object or something derived from html.

Also, different users might have different expectations regarding the design of their notes, so a core of common assumptions is what matters.

Regarding the progress between v.3.8.1 and v.3.8, others will be in a better position to chime in. I just jumped on this problem which was staring me in the face.

Re: V.3.8.1 adds extra blank lines when clipping Web pages

Posted: Fri Jul 21, 2017 8:23 am
by CintaNotes Developer
Well I guess we can consider the issue resolved for now.
However I'm definitely open to suggestions of any concrete HTML-to-rich text transformation algorithm improvements.
And huge thanks to you for taking time to report and test!