Re: [Htmlparser-developer] Line Separator

[Somik R. <so...@ya...>]

Hi Dhaval,
    Two of the bugs were in HTMLImageScanner.extractImageLocn() and its =
corresponding method in HTMLLinkScanner(). Fixed now, by removing \r as =
well.
    Also - there was another bug in HTMLLinkScanner.evaluate() - we need =
to check if next char after A is \r (we were only checking \n).
    All tests passing now.

BTW, currently all the test cases are geared to work on Windows. Do u
really think we need to open up all the test cases and set the line
separator character in case anyone happens to run the test cases on
UNIX. It does seem to be a terrific waste of time specially if we know
that all developers are going to be working on Windows.

There are a lot of people using the parser on linux. And I know at least =
one dev who uses only linux :)
Its not that bad, we just need to insert =
parser.setLineSeparator("\r\n"), and leave the rest of the testcase as =
is. If I find some time I will do it. Will be grateful if u can, as I =
will be on the road next week..

I might just find time to make the next release with your integration =
work. Thanks again for the great work.

Regards,
Somik

  ----- Original Message -----=20
  From:    =20
  To: so...@ya...=20
  Sent: Monday, August 26, 2002 3:42 PM
  Subject: RE: [Htmlparser-developer] Line Separator


  Hi Somik,

  I made the two changes that you pointed out. And yes there were lots =
of
  failures in the test cases (14 to be precise). I have fixed 11 of =
those
  and have no idea y I am getting the remaining 3 and hence am sending
  over the sources to you so that you can check it out at your time.=20

  BTW, currently all the test cases are geared to work on Windows. Do u
  really think we need to open up all the test cases and set the line
  separator character in case anyone happens to run the test cases on
  UNIX. It does seem to be a terrific waste of time specially if we know
  that all developers are going to be working on Windows.



  Regards,=20

  Dhaval Udani=20
  Senior Analyst=20
  M-Line, QPEG=20
  OrbiTech Solutions Ltd.=20
  +91-22-8290019 Extn. 1457=20



     -----Original Message-----
     From: somik [mailto:so...@ya...]
     Sent: Monday, August 26, 2002 10:25 AM
     To: Udani, Dhaval H.
     Cc: somik
     Subject: Re: [Htmlparser-developer] Line Separator
    =20
    =20

    =20
     Hi Dhaval,
     Sorry for that total goof-up regarding setLineSeparator() in
     HTMLParser.
     I really don't know how that happened. Anyway I've attached the
     required
     file. You can check it in.=20
    =20
     No problem - happens all the time with me. I've updated the release
     file.=20
    =20
     Regarding only 3 scanners having line separators, I had a look at =
all
     the scanners initially and once again today and do not find any of
     them
     requiring the line separator. Do let me know if I am mistaken and I
     will
     fix it up for u.
    =20
     Hmm.. I think there will be changes in a lot of testcases - if you
     see any of the tagTests, you will find lots of hard-coded tests =
with
     '', or "". That has to be changed to be consistent - surprising =
fact
     is all tests are passing. This is bcos the stuff that is sending =
the
     contents back is not using your latest integration - this is =
probably
     bcos of :
     HTMLTag.java (Method : incrementCounter())
     HTMLRemarkNode.java (Method : find())
    =20
     When I made the change, a lot of testcases failed - but this =
doesent
     mean there are bugs.
    =20
     Most of the tests should be re-written to use a known initialized
     end-of-line-char, or use the detected end-of-line char in creating
     the result test strings, or else, tests will fail in one platform =
and
     pass on another.
    =20
     The former option might be simpler, just setting the end-of-line =
char
     to a known value.
    =20
     Regards,
     Somik

    =20