Test of 9.02 Beta 2C

[MikeGale]

I redownloaded to get C, the download said B still and a diff said it was the same, but the about box said 2C.

I tested with a few batch download setups. Findings:

1) The edit for a new target, go away, come back, no dialog issue is fixed. Thanks.

2) For js validations, http not file, I get an error that there is no closing script tag. There is no need for such a tag.

3) If I validate, then change to another batch in the recent list, I get a message that essentially says "Please save the last batch (lst) file". This when the last file is unchanged. Seems to happen every time. If I don't validate this doesn't happen.

4) If I validate one batch list after another. The current list has no external links, but the previous one did, the validator shows itself validating the links (at least the difficult items on it). There are no such links, they don't appear on the report but they do take time. If I restart CSE with the no-links list it doesn't go through that phase.

[Albert Wiersch]

1) The edit for a new target, go away, come back, no dialog issue is fixed. Thanks.

Great, thanks for checking.

2) For js validations, http not file, I get an error that there is no closing script tag. There is no need for such a tag.

OK, I will check this and see if I can reproduce.

3) If I validate, then change to another batch in the recent list, I get a message that essentially says "Please save the last batch (lst) file". This when the last file is unchanged. Seems to happen every time. If I don't validate this doesn't happen.

This could be because the "Process" setting is being changed when you validate. In the General tab of the Batch Wizard Options, do you have the "After validating a target list, set process to 'No'..." option checked? If so, this could be causing this issue.

4) If I validate one batch list after another. The current list has no external links, but the previous one did, the validator shows itself validating the links (at least the difficult items on it). There are no such links, they don't appear on the report but they do take time. If I restart CSE with the no-links list it doesn't go through that phase.

This is by design due to the link cache remembering the links since the program was started. The simplest way to address this may be to provide an option to automatically clear the link cache before processing a new job (but this could cause some links to be checked again). Do you think such an option would be useful?

[Albert Wiersch]

2) For js validations, http not file, I get an error that there is no closing script tag. There is no need for such a tag.

Hi Mike,

Can you give me more details on how to reproduce this? I tried but couldn't. Thanks.

[MikeGale]

Yes I had that set to No option on.

The link cache in this case turned an "instant" validation into a lengthy wait. The new validation didn't have any of those links in it.

A way to switch that off (or not invoke it where those links are not involved) would be good.

[MikeGale]

I have retested the js issue.

The message I get is:

Code: Select all

Error [JavaScript] html error: unable to find end of script tag

There is no need for such a tag.

I tested files with identical content twice in that test. Once where it's referred to by URL (this failed) the other where it's out of the file system.

The file is the jquery-1.2.6.js file. (This is the source as opposed to minified version.) I have added /*jsl:ignoreall*/ as the first line.

There is a leading URL and a trailing file based item in the list but both are turned off (no validation).

Hope that helps clarify the issue.

[Albert Wiersch]

Yes I had that set to No option on.

OK, I bet that was it then. If it changes the 'process' setting, then it will think the target list is unsaved.

The link cache in this case turned an "instant" validation into a lengthy wait. The new validation didn't have any of those links in it.

A way to switch that off (or not invoke it where those links are not involved) would be good.

Actually, I forgot that there is an option that would probably solve this for you. Try unchecking the "Recheck bad links at next link check" option in the Link Checking tab of the Validator Engine Options. Then the link checker should not try to recheck those bad links again. By default it tries to recheck bad links in case they are fixed. Otherwise it could cause confusion because the link cache will remember a fixed link as bad.

[Albert Wiersch]

I have retested the js issue.

The message I get is:

Code: Select all

Error [JavaScript] html error: unable to find end of script tag

There is no need for such a tag.

I tested files with identical content twice in that test. Once where it's referred to by URL (this failed) the other where it's out of the file system.

The file is the jquery-1.2.6.js file. (This is the source as opposed to minified version.) I have added /*jsl:ignoreall*/ as the first line.

There is a leading URL and a trailing file based item in the list but both are turned off (no validation).

Hope that helps clarify the issue.

This sounds like a JavaScript Lint bug. Can you reproduce it in the editor (not in the Batch Wizard)? If so, then can you send me the file that it can be reproduced with? I will pass it along to the developer of JavaScript Lint after I verify it's a JavaScript Lint issue. Thanks!

[MikeGale]

I can confirm that it also happens in the editor on that file. I haven't tested others.

I suspect it happens with other js's but haven't checked today.

I've sent a copy of that file by email in case it's specific.

(Mail readers might block a js file.)

(Side Note: For production I'm using more recent jquery versions.)

[Albert Wiersch]

I can confirm that it also happens in the editor on that file. I haven't tested others.

I suspect it happens with other js's but haven't checked today.

I've sent a copy of that file by email in case it's specific.

(Mail readers might block a js file.)

(Side Note: For production I'm using more recent jquery versions.)

Thanks. Got it but Outlook won't let me open it and I don't know of a way around this security feature (thanks Microsoft). Can you send it again with a "jss" or "txt" or other extension?

[MikeGale]

File resent.

This file thing is really annoying. The vitriolic ignorance of the media drives it as does people without firewalls and anti-virus. What we really need is protection from these!!

[Albert Wiersch]

File resent.

This file thing is really annoying. The vitriolic ignorance of the media drives it as does people without firewalls and anti-virus. What we really need is protection from these!!

Thanks!

What bothers me most is that Outlook doesn't seem (as far as I know) to let the "experts" open these attachments. There should be some way around it for people who know what they're doing.

[MikeGale]

There are registry fixes for these things.

I think this
http://support.microsoft.com/kb/829982
describes the technique.

[Albert Wiersch]

There are registry fixes for these things.

I think this
http://support.microsoft.com/kb/829982
describes the technique.

Thanks!

[Albert Wiersch]

I can confirm that it also happens in the editor on that file. I haven't tested others.

I suspect it happens with other js's but haven't checked today.

I've sent a copy of that file by email in case it's specific.

(Mail readers might block a js file.)

(Side Note: For production I'm using more recent jquery versions.)

Thanks again for the report. The problem appears to have been caused by the temporary filename that is used to pass the document to the JavaScript linter when it's opened with a URL (and doesn't have a local file path). CSE HTML Validator should now use a "js" extension for JavaScaript temporary files so the linter will recognize it as a JavaScript file. I think before it must have been trying to interpret it as an HTML file with JavaScript in it (the temporary file previously ended in "tmp").

So this should be fixed in the next BETA or release candidate.

[MikeGale]

Thanks for the revised DLL.

I can confirm that the linter no longer gives that unexpected error.

Thanks.