Known problems with beta release V5

[wrensoft]

Known problems with beta release V5 beta build 15
Last update: 27/Nov/06

- We have seen an crash error. Access Violation at 0x0042563F (tried to write to 0x10932C16), program terminated. This occurs sometimes when attempting to use the PDF plugin. [Status: Fixed in V5.0 Beta 11b (18/Sept/06)]

- Crash problem and possible index corruption when using synonyms with incremental indexing. [Status: Fixed in V5.0 Beta 11c (21/Sept/06)]

- Poor CGI search times for some exact phrase searches. Some searches were 6 time slower than they should be. (e.g. 6sec instead of 1sec. [Status: Fixed in V5.0 Beta 11c (21/Sept/06)]

- Incremental indexing is not working for some combination of settings. In particular some character set settings. [Status: Fixed in V5.0 Beta 11c (21/Sept/06)]

- Crash problem when using some icon settings for file extensions [Status: Fixed in V5.0 Beta 11c (21/Sept/06)]

- When indexing in Offline Mode, all files indexed with plugins had a filesize of 1k. [Status: Fixed in V5.0 Beta 11c (21/Sept/06)]

- Double-quote and single-quote characters appear as ";" in the context description [Status: Fixed in V5.0 Beta 12 (5/Oct/06)]

- Incremental update fails for configurations using UTF-8, incorrectly reporting that the configuration file has changed when it has not [Status: Fixed in V5.0 Beta 12 (5/Oct/06)]

- Incremental update may cause subsequent incremental actions (such as "add new pages" or "add start points") to fail. [Status: Fixed in V5.0 Beta 12 (5/Oct/06)]

- Incremental update fails for configurations using no word join characters, incorrectly reporting that the configuration file has changed even when it has not [Status: Fixed in V5.0 Beta 13 (16/Oct/06)]

- FTP upload of large index files can fail on some severs as a result of a time out on the FTP control channel. Upload will lock up a couple of minutes into the upload [Status: Fixed in V5.0 Beta 13 (16/Oct/06)]

- Quote characters in page titles may appear with a backslash before them (\"), or may appear doubled up ("") in the search results. [Status: Fixed in V5.0 Beta 14 (1/Nov/06)]

- The help file and users guide have not been updated. However the information in this forum provides some details about the new features. [Status: Help file is now completed and included in V5.0 Beta 14 (1/Nov/06)]

- "Unauthorized..." usage error message [Status: Fixed in V5.0 Beta 14b (6/Nov/06)]

- Exact phrase matching with the ASP version sometimes returned incorrect results [Status: Fixed in V5.0 Beta 14b (6/Nov/06)]

- FTP issues on some machines. Occasionally freeze with the "Waiting for queued files" message or fail to upload the files with a "File already exists" error. [Status: Fixed in V5.0 Beta 15 (14/Nov/06)]

- Bugs with exact phrase matching for the ASP and PHP scripts [Status: Fixed in V5.0 Beta 15 (14/Nov/06)]

- Bug with searching for words containing apostrophes and ampersand characters even when they are enabled as word join characters [Status: Fixed in V5.0 Beta 15 (14/Nov/06)]

- Bug with picking up the ZOOMPAGEBOOST setting for the priority field on the XML sitemap output [Status: Fixed in V5.0 Beta 15 (14/Nov/06)]

- BSD search.cgi binary does not run on some newer versions of BSD [Status: Fixed in V5.0 Beta 15 (14/Nov/06)]

- Javascript search result links may not respond when trying to open them in Internet Explorer on a local computer due to MOTW used on the search.html page. See this thread (http://www.wrensoft.com/forum/showthread.php?p=4985#post4985) for more information and solution. [Status: Confirmed in V5.0 Beta 15 (20/Nov/06)]

- Ampersand characters in page titles and meta descriptions are not displayed as HTML entities. This causes the XML output to be considered invalid by some browsers. [Status: Confirmed in V5.0 Beta 15 (24/Nov/06)]

- Recommended links for keywords/phrases in UTF-8 and some foreign languages fail and do not return the recommended link when a user searches for that word. [Status: Confirmed in V5.0 Beta 15 (24/Nov/06)]

- Multiple ZOOMCATEGORY meta tags are not being picked up, and assigning the file to multiple categories. [Status: Confirmed in V5.0 Beta 15 (24/Nov/06)]

- We need to do more testing on Windows 98, 2000, Server 2003 and Vista. To date most testing has been done on XP. We might even drop support for Windows 98 if there are incompatibility issues with V5.

- There appears to be a problem with FTP uploading on Vista with the built-in Windows Firewall enabled. Adding Zoom to the firewall exclusion list did not seem to help. This however, appears to be a Vista bug as other FTP clients have similar issues. We are continuing to investigate this, and it may be limited to a certain beta version of Vista.

As we find more issues we'll update this list.

[cardiogr]

Hi,
Version 5 is amazing. It introduces many important features! Thank you for all your efforts.

I have come to a small problem. I don't know if it is my fault or if it is a bug.The problem is the following:

I have tried to build my own Search form using checkboxes. This is essential for me in order to apply my own javascript code to the search form.
I have noticed however that multiple categories search does not work in manual generated search forms unless the input type checkboxes are named zoom_cat[] (or zoom_cat%5B%5D in Hex) instead of zoom_cat.

Although after renaming zoom_cat to zoom_cat[] multiple categories search worked with no problem the addition of the [] signs creates problems in manipulating the checkboxes in javascript code, for example in order to save a cookie and remember the previously entered values.

It would be great if you fixed this issue in a future release so that multiple categories search (in manual generated forms) works without the need to rename zoom_cat to zoom_cat[]

[Ray]

Yes, there are some changes to the "zoom_cat" parameter (used in search forms and in the URL for a search query as a GET parameter) for the new categories in V5.

First of all, the old "zoom_cat" (without square brackets) should still work. This will only allow you to specify a single category to search within at any one time (as in previous versions).

The new "zoom_cat[]" parameter is required for searching multiple categories at any one time (by enabling "Allow searching in multiple categories" in the Categories Configuration window). This was necessary because of the limitations on the scripting platforms supported: PHP, for example, can not handle multiple parameters with the same name, unless it is specified as an array (thus the square brackets). I think the ASP platform might be able to handle it without square brackets, but we wanted the usage to be consistent across the platforms so it became a requirement.

This should however, not restrict you in your scripting or creation of a custom search form. If you previous script only handled "zoom_cat", then you should only need to update it to use "zoom_cat[]" instead.

I'm not sure how it could hinder saving a cookie or remembering previous entered values. I think this depends on your code and you can find a way to do this if you look around online. It should definitely be possible since the "zoom_cat[]" method is the only way to accept checkboxes via PHP.

[cardiogr]

Hi Ray,
Thank you for your answer.
The problem is that [ and ] are special characters that are recognised as special command by javascript when entered.
So when I try to refer to zoom_cat[] (for example document.form_name.zoom_cat[]) this is not recognised.
I have changed zoom_cat[] to zoom_cat12 (and zoom_cat%5B%5D to zoom_cat%31%32) hexedititing the search.cgi file
I hope there is no problem with this, but this solved the problem.
This could work for anyone having the same problem.

[Ray]

Your abovementioned method is not necessary.

You can retrieve the zoom_cat[] element in Javascript by use of the getElementsByName function, eg.


var cats = document.getElementsByName("zoom_cat[]");


This would give you an array named cats, with each item comprising of a .checked property which you can change or verify.

We do not recommend hex editing the CGI file. Or at the least, please be aware that you could easily break functionality and we would be unable to provide any support for a hex-edited CGI file.

[cardiogr]

Ray, thank you for the code and for your time.
I replaced the original cgi and tested it by replacing my code with yours.
Unfortunately this didn't work either.
It still can't get the element by name because of the [].

[Ray]

I have tested my method before I posted it, so I have confirmed that it works. Perhaps there's something else wrong in your script, or your search form is declared differently than standard? The [] characters should not be the cause of the problem when using the getElementsByName function call, because it is specified as a string, so Javascript would not have any problem with it.

Here's the full example I used to test this method:


<script>
function test()
{
var el = document.getElementsByName("zoom_cat[]");
for(i=0;i<el.length;i++)
{
if(el[i].checked)
{
return true;
}
}
alert("Please check at least one box!");
return false;
}
</script>

...

<form method="get" action="search.cgi" class="zoom_searchform" onsubmit="return test();">
Search for: <input type="text" name="zoom_query" size="20" value="*e*" class="zoom_searchbox" />
<input type="submit" value="Submit" class="zoom_button" />
<span class="zoom_categories">
Category: <ul>
<li><input type="checkbox" name="zoom_cat[]" value="-1" checked="checked">All</input></li>
<li><input type="checkbox" name="zoom_cat[]" value="0">News</input></li>
<li><input type="checkbox" name="zoom_cat[]" value="1">Articles</input></li>
</ul><br /><br />
</span>
</form>


The above example should pop up a message window if you have no checkboxes selected.

[cardiogr]

You are right Ray.
I tested your code and it works. It is logical what you say, that it is now a string, it should work.
I changed again my code and now it works in my code too.
I can't understand what happened at my first attempt.
Maybe I had forgotten to change some functions, maybe I didn't clear the cached files in my browser.
Anyway , it is working now!
Thank you for your time and your support

[daniswired]

Hi, I can't wait for V5 to come out in final release. I tried the beta on a live site and came across one issue, the direct link to the page is incorrect. i.e. The files are placed in a folder called "files", but the final link (after the results have been returned) links to "http://www.mysite.com/files/ww.mysite.com/files/page.html" instead of "http://www.mysite.com/files/page.html.

Thanks in Advance!

Dan

[wrensoft]

This is probably not a bug. More likely you are using offline mode and haven't entered the correct base URL. What did you enter as the start point and base URL?

[daniswired]

I see where the issue is in the resulting HTML. The result is returning a link with the path <a href="www.website.com/folder/item.html"> instead of <a href="http://www..... ">. Both Netscape and Safari Browsers (mac) don't seem to like this.

Here is the actual test:

http://www.101hardware.com/search_test/search.php?zoom_query=air&zoom_per_page=10&zoom_and=0&zoom_sort=0

I really appreciate your input!

Dan

[Ray]

Your ZDAT files appear to have been corrupted by uploading the files in text mode. Are you using an external FTP program instead of Zoom to do your uploading? If so, you need to make sure the FTP program you are using uploads in BINARY mode. Do not use "Auto" modes in FTP programs, they rarely ever guesses correctly.

You can also use Zoom's built-in upload feature to ensure that the files are uploaded correctly.

If you continue to get this problem, can you ZIP up your search files (all ZDAT files including search.php and settings.php) along with your ZCFG file containing your indexer settings, and e-mail these to us (http://www.wrensoft.com/contactus.html). We will then try to reproduce this problem (we have tried indexing your test site already with our own settings and did not have any problems).

[daniswired]

Thank you very much for your quick responses.

When you mentioned ftp apps, red lights started flashing. I uploaded the files through Zoom and it worked like a charm. I was uploading the files via Fetch on a mac. The files were indeed getting corrupted.

After trying all methods of transfer, I think it's safe to tell all mac users that Fetch does not work, in ANY setting (raw data and ASCII failed).

Again, I can't wait for the final release!

Dan

[AG!]

Hello, its me again,

Yesterday, using the latest build Beta 12, my config file had 11 start points. I had a limit of 300 files per start point in the limits tab (for testing). Zoom only indexed the first 9 start points. I then added another start point bringing the total to 12 and zoom indexed 10 this time. If you want to see the cfg file, I can email it to you when I get home from uni.

Thanks

[wrensoft]

Yes, please E-mail us the config file.

[Ray]

AG,

We did some tests with your ZCFG file.

The file you sent us had 13 start points to begin with. We ran a full index on this, and it succeeded to index 12 out of the 13 start points. The last start point returned this error message:

Additional start URL invalid or already scanned: http://www.troid.org/store/home.php

This is correct behaviour. This last start point in your list was already indexed by your second start point (http://www.troid.org/).

We then removed 2 of the start points, returning it to the 11 start point configuration you mentioned. We ran a full re-index with this configuration, and it succeeded without problems (11 out of 11 start points were indexed).

We then used the "Incremental" -> "Add start points to existing index" option to add the 2 start points back, again following your description. This also ran without problems.

So to conclude, we did not see the same problem you appear to be having.

But here are some points and possible explanations for what you are seeing:

- As we saw above, you may have been adding start points that were already indexed by your existing start points, and thus, Zoom skips over them. This is expected and normal.

- Did you check/read the error messages that popped up in red in the Index Log? Particularly near the end of indexing. These message would tell you why Zoom stopped indexing. It is important to make note of these messages.

- Some of the domains you were indexing from may have been down during your indexing session. Or your own Internet connection may have been unavailable. There should be error messages in the Index Log which indicate download problems if there were any. Note that you can save the index log to file (see "Index Log" tab in the Configuration window) if you wish to search through your index log in more detail via Notepad etc.

- We noticed one minor problem with the Status reporting "Start points indexed: 1 / 2" when performing an incremental add start point with only 1 start point. It should say "1 / 1" (the number of incremental start points reported in the status is one more than the actual), we'll fix this. It does not appear to be related to your described problem though. Unless you were basing your counts on this.

If you continue to have problems, please take note of the Index Log messages, and check what reason Zoom gives you as to why a start point was not indexed. Save this index log and send it to us as well if you can pinpoint the problem.

[cardiogr]

Hi, I tried Zoom Search Engine V5 beta 14 .
I used an enterprise key sent to me by you!
After Indexing and searching I clicked on a result. It connected to your site and I received this error message
"Unauthorized copy of Zoom Search Engine detected "
This is the first beta release that gives me this message.
I have used most of the previous beta versions with Professional Edition (before you introduced Enterprise edition) and I didn't get this message.
Thank you!

[Ray]

This appears to be a bug in the beta 14 build. Please contact us by email for more information. We will provide a new build addressing this issue ASAP.

[cardiogr]

Thank you for your answer.
I rolled back to Beta 13 and reindexed with Beta 13.
I will wait for the next version, since Beta 14 has this bug.

[Ray]

After Indexing and searching I clicked on a result. It connected to your site and I received this error message
"Unauthorized copy of Zoom Search Engine detected "

This problem has now been fixed in a patch release we made today. You can download it from the same download link found on the "How to download Zoom V5 (http://www.wrensoft.com/forum/showthread.php?t=1125)" thread.

The patch release is named "Version 5.0 Beta 14b". You can verify if you are using this patched build by clicking on Help->About within the Indexer.

Sorry for the confusion and the inconvenience.

[modernbaby]

why is the recommanded links can't support foreign language? i input foreign language in recommanded links's words,but when i input this words in my search,can't show recommanded links.

[wrensoft]

What is the foreign word being entered for the recommended link?
What character set (e.g. UTF-8) are you using on your web pages?

[modernbaby]

i used chinese,i used utf-8 in search page,but the search is no problem,only recommanded link have error.

[Ray]

This seems like a bug (unable to match recommended link keywords in UTF-8 encoding). We will look into it on Monday, and hopefully have it fixed for the final release.

[wrensoft]

V5.0 has now been publicly released on the 30/Nov/06. So this thread for beta testing is now closed. Any problems found in the public V5.0 release should be posted in a new thread.