xml - Using XPATH to search text containing   -

i use xpather browser check xpath expressions on html page.

my end goal use these expressions in selenium testing of user interfaces.

i got html file content similar this:

 <tr>   <td>abc</td>   <td>&nbsp;</td> </tr> 

i want select node text containing string " ".

with normal string "abc" there no problem. use xpath similar //td[text()="abc"].

when try an xpath //td[text()=" "] returns nothing. there special rule concerning texts "&" ?

it seems openqa, guys behind selenium, have addressed problem. defined variables explicitely match whitespaces. in case, need use xpath similar //td[text()="${nbsp}"].

i reproduced here text openqa concerning issue (found here):

html automatically normalizes whitespace within elements, ignoring leading/trailing spaces , converting spaces, tabs , newlines single space. when selenium reads text out of page, attempts duplicate behavior, can ignore tabs , newlines in html , assertions based on how text looks in browser when rendered. replacing non-visible whitespace (including non-breaking space " ") single space. visible newlines (<br>, <p>, , <pre> formatted new lines) should preserved.

we use same normalization logic on text of html selenese test case tables. has number of advantages. first, don't need @ html source of page figure out assertions should be; " " symbols invisible end user, , shouldn't have worry them when writing selenese tests. (you don't need put " " markers in test case asserttext on field contains " ".) may put newlines , spaces in selenese <td> tags; since use same normalization logic on test case on text, can ensure assertions , extracted text match exactly.

this creates bit of problem on rare occasions when want/need insert whitespace in test case. example, may need type text in field this: "foo ". if write <td>foo </td> in selenese test case, we'll replace spaces 1 space.

this problem has simple workaround. we've defined variable in selenese, ${space}, value single space. can use ${space} insert space won't automatically trimmed, this: <td>foo${space}${space}${space}</td>. we've included variable ${nbsp}, can use insert non-breaking space.

note xpaths not normalize whitespace way do. if need write xpath //div[text()="hello world"] html of link "hello world", you'll need insert real " " selenese test case match, this: //div[text()="hello${nbsp}world"].