|
Release Notes
 Version 3.3.1- The problem of an href not being found in an anchor (<a) if it was not preceded by a space has been corrected. For example, if the entry appears as <a style="...."href=..., the href will now be found.
- The problem of the program crashing or having to be aborted when a comment tag (<!--) did not have a correct closing tag (-->) has now been corrected.
- Version 3.3.0
- New Feature
- The
program now includes an option to allow the rechecking of just the URLs
that were reported as broken in a previously run test (option appears
when Recheck is selected).
- Bug Fixes
- When
"Stop" was clicked the program will now stop cleanly.
- The "Abort" command (under the "File" menu) will now cause the program to abort immediately, as intended.
- An error in the parsing of styles that would sometimes result in the program looping has been corrected.
- Version 3.2.2
- In spite of other fixes, CyberSpyder Link Test would still occasionally
crash with an access violation error. This appears to be caused
by a bug in the Delphi (language used to program CyberSpyder) libraries
but, after much searching on the Internet, a solution has been found
and implemented.
- Version 3.2.1
- A long <style
entry (longer than 1275 characters) within an HTML page would cause an
error in CyberSpyder
Link Test and make it impossible to complete a test. This has
been corrected.
- An empty URL entry in a style sheet (url()) would cause
an error that would result in the program aborting. This has
been corrected.
- A style entry using @import would cause
the program to go into a loop if a url
entry did not appear somewhere after it in the style entry. A
valid @import
can be in the form @import
url('...the url...') or @import '...the url...' and
the program was expecting the first form or a url entry later in
the style entry. This has been corrected.
- A
program error in a procedure call would cause an error
requiring a
program abort. The error message would sometimes refer to the
AddNameIDRec
procedure in the DBProcessModule.
It was caused by a failure to prevent more than one thread
from trying
to access this routine at the same time. This has
been corrected.
- Version
3.2.0
- New
Features
- The program will now parse for URL references in
all style
entries associated with all tags.
- The program will now parse for URLs contained in style sheets.
- An option has been added to suppress the
reporting of empty items (such as within an <IMG item
where the alt
value may appear as alt=""
- i.e. empty). The empty items often result when a WYSIWYG
editor to create Web pages.
- Bug
Fixes
- Fixed a counter problem where, when the program
had completed a test, it would sometimes show 1 item left to check
instead of 0.
- Fixed
a problem associated with the use of Tag IDs as HREF targets.
The
original logic could result in a significant slowdown of parsing (and
testing) if a page contained many IDs. The logic has been
changed
to avoid this slowdown.
- If a URL was longer than 255
characters, it could cause the program to go into an endless loop
rechecking the particular URL. This has been
corrected.
- Fixed
a problem associated with the use of a GET request when a site did not
respond to a HEAD request. In some cases, the complete file
would
be downloaded whether it was needed or not. This has been
corrected so that the program only downloads the header portion of
these files. This should speed up the testing of some sites.
- Version
3.1.0
- New
Features
- The
program help system has been updated to use HTML help instead of the
older WinHelp. This change should make the program compatible
with Windows Vista.
- CyberSpyder
Link Test
is often used by a single user to test more than one site. To
add
to the functionality of the program, it now includes an option to
select a template directory other than the default so the user can
define
different report templates for different sites, if desired.
- Some major
changes are planned for URL retrieval and page parsing to speed up
processing. An initial change has already been implemented.
To save time when checking URLs, the program first requests
the
HTTP header (via a HEAD)
request. Some servers do not accept this request, either
ignoring
it or returning an error. In such cases the program retries
with
a GET
request, hence
receiving the entire document whether it is needed or not.
The
program has now been modified so that, when a GET is being used
because HEAD
did not work,
it will now watch for the end of the HTTP header, then terminate the
connection. For some sites, this will result in a significant
speedup in testing.
- Bug
Fixes
- If a style
is defined for the <BODY
tag, the program will now check to see if a background image is
specified in the style and, if found, will check that it exists.
- Changed
in Version 3.03
- Added:
When Javascript is used as the link within an anchor, the href is often
entered as href="#".
While this is technically incorrect and normally
reported as an error, an option has been added (see Test Options tab of
the Test Configuration
Screen) to suppress the reporting of this as an error/
- Added:
Similarily, are sometimes intentionally written
with the href
URL ending with a "#"
.
Again, this is technically wrong but is a common practice,
therefore an option has been added to suppress the reporting of this as
an error.
- Fix: The
Site Map and Site File Structure Reports would contain duplicate
entries if name
anchors or a query
string were part of a URL - there woukl be an entry for
every name
anchor or query
string used instead of one entry for each basic URL.
This has been corrected.
- Fix: When
a URL appeared in the form "http://some.host.name?query=something",
CyberSpyder would misinterpret this as being a URL with only the host
name and add a trailing slash. This gave an incorrect URL of
"http://some.host.name?query=something/" resulting in
CyberSpyder
parsing and getting a host name of "some.host.name?query=something",
resulting in broken links being reported with the error "11004:
[11004] Valid name, no data record (check DNS setup)".
This has been corrected.
- Fix: When
a URL that contained the entity "&",
it would not be interpreted correctly. The "&"
should be converted to "&",
but it was not. This has been corrected.
- Fix: An
encoded URL (like "mailto:s..."),
a form sometimes used to encode email addresses, was not properly
handled. This has now been corrected.
- Fix: The
HTML (and XHTML) head tag can be in the form <head>
or <head
profile=...URI.... CyberSpyder found the simple
form but not the <head
profile... form and reported a missing head tag.
This has now been corrected.
- Fixed
in Version 3.0.2
- A problem was discovered with extracting the
correct path from some
URLs (where one of the directory names contained a '.') This
has
been corrected.
- If the header data returned by a request
included a "Content-Location", CyberSpyder failed to recognize it,
potentially resulting in an error when expanding any relative URL(s)
that appear in the page. This has been resolved.
- Fixed in Version 3.0.1
- A problem was discovered with the
entry of registration
information. It was not being recorded. Also, the
program is supposed
to recognize the registration information already recorded for previous
versions so that the information does not need to be re-entered.
This was not working correctly. Both problems have
been resolved.
|
