103539 - Find does not always find strings containing spaces.

Status: VERIFIED FIXED

(SeaMonkey :: UI Design, defect, P3)

Description

[Mehmet D. Akýn]

Steps:

-download the html from the URL (Beware, very big one)
 http://www.tuxedo.org/~esr/jargon/jargon.html

Try to find the sentence from second paragraph

"The Jargon File is a common heritage of the hacker culture. Over the years a
number of individuals have volunteered considerable time to maintaining the File
and been recognized by the net at large as editors of it." 
in the file 
-copy the text from HTML
-paste it to find text edit box in find dialog 
-and click Find button

Expected result: Mozilla Finds the text and highlights it
Actual Results: Mozilla pops up an "text you entered cannot be found" message

is there a limitation on string size in dialog?

Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:0.9.4+) Gecko/20011004

Comment 1

[Ashley Bischoff (blog at handcoding.com)]

-> XP Apps: GUI Features, for initial triage

Comment 2

[Christopher Aillon (sabbatical, not receiving bugmail)]

The given string contains more than one space.  Dupe of bug 98730 -
"Find/Replace not finding a string with more than one space"

*** This bug has been marked as a duplicate of 98730 ***

Comment 3

[Mehmet D. Akýn]

But I tried other shorter strings on page with more than 1 spaces in it, it
found them.. I dont think this is a dup.
it finds the string
"The Jargon File is a common heritage of the hacker culture."
but cannot find the fallowing
"The Jargon File is a common heritage of the hacker culture. Over"

Comment 4

[Blake Ross]

--> kin

Comment 5

[kinmoz]

The problem is that the whitespace ignoring code in find is broken.

In this particular case the user sees only one space between certain words, and
in the actual text that find is searching through, some of these words are
separated by a space and a newline (2 spaces in a row).

I'll have to rework the find code's whitespace ignoring code so that we match at
least n spaces (n being the number of spaces in a row in the pattern string)
within the run of spaces in the string being searched, and then ignore any extras.

Comment 6

[kinmoz]

Attachment: Simple Testcase for testing whitespace ignoring code.

Comment 7

[kinmoz]

Attachment: Patch Rev 1

Comment 8

[kinmoz]

Changing summary from:

  "Find" does not find long strings in big HTML pages

to:

  Find does not always find strings containing spaces.

Comment 9

[kinmoz]

Attachment: Patch Rev 1.1 (Minor changes based on review with sfraser)

Comment 10

[Simon Fraser [no longer active]]

Comment on attachment 55136 [details] [diff] [review]
Patch Rev 1.1 (Minor changes based on review with sfraser)

sr=sfraser

Comment 11

[Joe Francis]

I recommended a perf enhancement and found a correctness prob.  Kin has fixed
both of these (fixesnot included here), so r=jfrancis

Comment 12

[kinmoz]

Attachment: Patch Rev 1.2 (based on review with jfrancis)

Comment 13

[kinmoz]

Fix checked in on trunk:

    mozilla/editor/txtsvc/src/nsFindAndReplace.cpp  revision 1.8

Comment 14

[sairuh (rarely reading bugmail)]

vrfy fixed using 2001.11.08.0x-comm bits on linux rh6.2, winnt and mac os 10.1.

using kin's attached test case, Find was able to locate the following strings:

"words in" [1 space]
"words  in" [2 spaces]
"words   in" [3 spaces]

this is expected, since the source contains "words   in".