The underscore in css class selector

[nikita-volkov]

Queries like css ".some_class" produce no results when in fact there are elements with that class

[nikita-volkov]

Just noticed that it isn't stably reproducable

[egonSchiele]

Can you give me an example where this issue occurred? With my test file in tests/test.html, the selector works just fine.

[nikita-volkov]

Here's an example of multiple correct selectors which produce no results:

main = do 
  print $ runLA (hread >>> tags //> getText) $ html
  where
    tags = css "div .fl_r" <+> css ".info .audio_add_wrap" <+> css ".fl_l .fl_l" <+> css "div.title_wrap"
    html = "<div class=\"audio  fl_l\" id=\"audio126257070_150084772\" onmouseover=\"addClass(this, 'over');\" onmouseout=\"removeClass(this, 'over');\">  <a name=\"126257070_150084772\"></a>  <div class=\"area clear_fix\" onclick=\"if (cur.cancelClick){ cur.cancelClick = false; return false;} playAudioNew('126257070_150084772')\">    <div class=\"play_btn fl_l\">      <div class=\"play_btn_wrap\"><div class=\"play_new\" id=\"play126257070_150084772\"></div></div>      <input type=\"hidden\" id=\"audio_info126257070_150084772\" value=\"http://cs1-4.userapi.com/d33/859470469ef948.mp3,221\" />    </div>    <div class=\"info fl_l\">      <div class=\"title_wrap fl_l\" onmouseover=\"setTitle(this);\"><b><a href=\"/search?c[q]=Michel%20Tel%5C%F3&c[section]=audio\" onclick=\"if (checkEvent(event)) { event.cancelBubble = true; return}; Audio.selectPerformer(event, 'Michel Tel\&#243;'); return false\">Michel Tel&#243;</a></b> &ndash; <span class=\"title\"><a href=\"\" onclick=\"Audio.showLyrics('126257070_150084772',24208242,1); return cancelEvent(event);\">Bara Bar&#225; Bere Ber&#234;</a> </span><span class=\"user\" onclick=\"event.cancelBubble = true;\"></span></div>      <div class=\"actions\">        <div class=\"audio_add_wrap fl_r\" onmouseover=\"Audio.rowActive(this, 'Добавить в мои аудиозаписи', [9, 5, 0]);\" onmouseout=\"Audio.rowInactive(this);\" onclick=\"Audio.addShareAudio(this, 150084772, 126257070, 'a74352b70e39439b99', 0, 1); return cancelEvent(event);\">  <div class=\"audio_add\"></div></div>      </div>      <div class=\"duration fl_r\">3:41</div>    </div>  </div>  <div id=\"lyrics126257070_150084772\" class=\"lyrics\" nosorthandle=\"1\"></div></div>"

[egonSchiele]

Here's what I get:

["  ","3:41","  ","3:41","  ","3:41","  ","  ","      ","      ","    ","      ","Michel Tel"," "," ","Bara Bar"," Bere Ber"," ","      ","        ","  ","      ","      ","3:41","    ","Michel Tel"," "," ","Bara Bar"," Bere Ber"," ","Michel Tel"," "," ","Bara Bar"," Bere Ber"," ","Michel Tel"," "," ","Bara Bar"," Bere Ber"," "]

What versions of GHC and HXT do you have, and what platform are you on? Does upgrading ghc / hxt fix the issue?

[nikita-volkov]

Strange it is stably reproducable on my two machines: OSX 10.8.2 with GHC 7.4.2, HXT 9.3.0.1, upgraded to 9.3.1.1. Ubuntu 12.10 (64-bit, AMD processor) with GHC 7.4.2, HXT 9.3.1.1.

Just in case, the import statements are:

import Text.XML.HXT.Core
import Text.HandsomeSoup

Could it probably be a locale/UTF/special symbols related issue?

I must note that most other selectors work fine.

[egonSchiele]

Does it work for you with the special symbols removed? Or does the following work for you?

runX $ parseHtml html >>> multi (hasAttrValue "class" (elem "info" . words)) //> getText

That's the equivalent translation to pure HXT. And what happens when you add a multi in front of tags?

print $ runLA (hread >>> multi tags //> getText) $ html

I would expect this to give you duplicated text.

[nikita-volkov]

Adding multi in front of tags still produces an empty list.

Clearing the HTML from ampersands does not help.

Concerning your second question, I think you've presented a selector for css ".info", which works fine for me. The one not working is css ".info .audio_add_wrap", and I've tried the following HXT translation of it and it worked fine

runX $ parseHtml html >>> multi (hasAttrValue "class" (elem "info" . words)) >>> multi (hasAttrValue "class" (elem "audio_add_wrap" . words)) //> getText

So I guess the problem is still somewhere in HandsomeSoup

[pooya-raz]

I think I'm having similar problems, but with hyphens. This doesn't work links <- runX $ doc >>> css ".item-result" but this works: links <- runX $ doc >>> css "div" >>> hasAttrValue "class" (=="item-result")

And this is the html source I'm parsing: http://pastebin.com/HNmRvFC6

[egonSchiele]

I'll take a look.

[ibotty]

@pooster: are you running version 0.3.2 or the latest hackage 0.3.1. 0.3.2 fixed a few bugs that look similar to yours.

[egonSchiele]

@pooster your example works for me with version 0.3.2 (it doesn't work with version 0.3.1).

[egonSchiele]

@nikita-volkov it's been a while but could you verify that you're still having this issue. I don't see it.

[nikita-volkov]

In a couple of days. Yes. 17.08.2013 7:56 ÐÏÌØÚÏ×ÁÔÅÌØ "Aditya Bhargava" notifications@github.com ÎÁÐÉÓÁÌ:

@nikita-volkov https://github.com/nikita-volkov it's been a while but could you verify that you're still having this issue. I don't see it.

Reply to this email directly or view it on GitHubhttps://github.com/egonSchiele/HandsomeSoup/issues/8#issuecomment-22805069 .