Re: [development] Xpath query for class matching

[Domenic Santangelo]

You don't encode the query string, just the path.

-D

nitin gupta wrote:

Yeah, I understand . it is the thumbnail maker for google images page. But if I want to have a specific script for this conversion, one can never be sure. If you notice, there are two colons in the query, which if percent encoded will render the URL invalid.

Thanks for all your inputs.

--
Regards,
Nitin Kumar Gupta
http://publicmind.in/blog/

On Sat, Oct 3, 2009 at 3:06 AM, Domenic Santangelo <dome...@workhabit.com <mailto:dome...@workhabit.com>> wrote:

    nitin gupta wrote:

    @URL: Maybe in hurry, I did not explain it properly. This was
    exactly what I was doing. look at the project,
    https://sourceforge.net/projects/absoluteurl/

    until I came across this image url on the google images page:

    
http://t3.gstatic.com/images?q=tbn:7i1D2KAZcCd8yM:http://www.flash-slideshow-maker.com/images/help_clip_image004.jpg

    What would you consider here the query string or the path?
    Anything we do, the URL is going to break unless we do nothing.
    Any ideas?(may be just for encode the spaces).

    scheme: http
    host: t3.gstatic.com
    
<http://t3.gstatic.com/images?q=tbn:7i1D2KAZcCd8yM:http://www.flash-slideshow-maker.com/images/help_clip_image004.jpg>
    path: images
    
<http://t3.gstatic.com/images?q=tbn:7i1D2KAZcCd8yM:http://www.flash-slideshow-maker.com/images/help_clip_image004.jpg>
    query:
    
q=tbn:7i1D2KAZcCd8yM:http://www.flash-slideshow-maker.com/images/help_clip_image004.jpg
    
<http://t3.gstatic.com/images?q=tbn:7i1D2KAZcCd8yM:http://www.flash-slideshow-maker.com/images/help_clip_image004.jpg>

    That url is some sort of thumbnail maker or something... compare
    to http://www.flash-slideshow-maker.com/images/help_clip_image004.jpg

    -D

    --
    Regards,
    Nitin Kumar Gupta
    http://publicmind.in/blog/


    On Sat, Oct 3, 2009 at 2:22 AM, Domenic Santangelo
    <dome...@workhabit.com <mailto:dome...@workhabit.com>> wrote:

        nitin gupta wrote:

        Thanks for your help. but this query will probably select
        "abcfooxyz" as well when "foo" is supplied. (untested),
        although it will definitely select "foo" in "foo bar" (tested)

        $xpath->query("//div[contains(@class, '" . $class . "')]")

        How can we be more specific?

        Try this example:
        http://westhoffswelt.de/blog/0036_xpath_to_select_html_by_class.html

        @URL: Actually I am maintaining the module feedapi
        imagegrabber, which downloads images from external websites.
        Now sometimes the url I parse has spaces, so I am unable to
        decide whether or not  to percentage encode the URL, because
        percentage encoding will make this URL valid but will break
        the following URL:
        http://www.google.com/search?q=hello
        by converting it to
        http://www.google.com/search?q%3Dhello

        "q=hello" is a query string, not strictly part of the path. I
        would strip the url to its component parts (parse_url) and
        encode the path, then re-append the query string.

        http://us2.php.net/manual/en/function.parse-url.php
        http://www.faqs.org/rfcs/rfc1738.html

        HTH,

        -D

        Looking forward.

        --
        Regards,
        Nitin Kumar Gupta
        http://publicmind.in/blog/


        On Sat, Oct 3, 2009 at 1:30 AM, Domenic Santangelo
        <dome...@workhabit.com <mailto:dome...@workhabit.com>> wrote:
        >
        > nitin gupta wrote:
        >
        > for eg <div class="foo"></div> is selected by the
        following query.
        > $class = "foo";
        > $xpath->query("//*...@class = '".$class."']");
        > But, this fails in case of multiple classes, i.e. does not
        select this :
        > <div class="foo bar"> </div>
        >
        > The "contains" function would probably work, eg (untested):
        > $xpath->query("//div[contains(@class, '" . $class . "')]");
        >
        > One more question: is this url valid : http://xyz.com/path
        with space.html
        >
        > Nope. You can escape the url if you just HAVE to have spaces:
        >
        > http://n00b.com/path%20with%20space.html
        >
        > but why not use dashes?
        >
        > http://pro.com/path-with-properly-indexed-spaces.html
        >
        > -Dom