Things noticed on first test of V 9

[MikeGale]

Hi,

I got around to a quick test of 9 today. (That's BETA 1.)

Here's what I saw:

1) A CSS sheet that previously validated fine now comes up saying 'The "right" property cannot have a negative value like "-1.5em".' for code 'right: -1.5em;'. I haven't examined the source in any detail, but my understanding is that negative values are permitted. (They work in the browsers.) See, for example, http://www.w3schools.com/htmldom/prop_style_right.asp

2) I noticed a series of false validation messages. In all cases characters, which aren't there were "seen" after the closing html tag. (These included, "he", "<", "</", '"/' and "me", with a couple of them having line breaks up front.) Without detailed checking I suspect that these particular pages are unchanged pages created on a Unix system. Other pages, which I have edited, don't show this. That might give a clue as to what is happening.

3) I changed the editor font a few times. When I tried Consolas the display for a tabs, at the beginning of lines was wrong. (I show whitespace in the editor.) In Courier New it is an arrow pointing right. In Consolas it's a little box with something inside it, not sure what. I guess that Consolas is becoming fairly popular. At last check I think it had more than 37% installed base on Windows machine. Probably much higher on recent machines. I'm guessing that the new character displayed for Tab is a glyph that is unreliable. (Version 8 shows, I think, the double chevron french quote character (Guillemet), for a tab.)

Thanks for the work on version 9, especially the utf-8, it's very welcome.

I look forward to the next Beta.

[MikeGale]

Another thing I found.

4) I couldn't paste text from the clipboard to file path text boxes in, batch wizard, target list, file options.

[MikeGale]

5) A single CSS of several similar files generated an error. The first line is an @import statement. CSE (batch) reported that the line did not have a closing semi-colon. The semi-colon is there. This is a utf-8 file. A diff tool showed this line and a similar line in another file to be identical, so there seem to be no "invisible characters" involved.

6) In the batch wizard the progress % indicator (bottom right) stopped at 59% (I think) even though the report was done.

[Albert Wiersch]

1) A CSS sheet that previously validated fine now comes up saying 'The "right" property cannot have a negative value like "-1.5em".' for code 'right: -1.5em;'. I haven't examined the source in any detail, but my understanding is that negative values are permitted. (They work in the browsers.) See, for example, http://www.w3schools.com/htmldom/prop_style_right.asp

Thanks. I have already fixed this bug for the next BETA.

2) I noticed a series of false validation messages. In all cases characters, which aren't there were "seen" after the closing html tag. (These included, "he", "<", "</", '"/' and "me", with a couple of them having line breaks up front.) Without detailed checking I suspect that these particular pages are unchanged pages created on a Unix system. Other pages, which I have edited, don't show this. That might give a clue as to what is happening.

Could you email me or send me a link to a document that generates these false validation messages?

3) I changed the editor font a few times. When I tried Consolas the display for a tabs, at the beginning of lines was wrong. (I show whitespace in the editor.) In Courier New it is an arrow pointing right. In Consolas it's a little box with something inside it, not sure what. I guess that Consolas is becoming fairly popular. At last check I think it had more than 37% installed base on Windows machine. Probably much higher on recent machines. I'm guessing that the new character displayed for Tab is a glyph that is unreliable. (Version 8 shows, I think, the double chevron french quote character (Guillemet), for a tab.)

Thanks. I've reproduced this problem and reported it to the editor developer.

Thanks for the work on version 9, especially the utf-8, it's very welcome.

Yes, I'm glad this support will be added! The tools and components I use are a bit "dated" when it comes to Unicode support, but that is changing.

[Albert Wiersch]

Another thing I found.

4) I couldn't paste text from the clipboard to file path text boxes in, batch wizard, target list, file options.

Thanks. That is strange. I will look into this.

[Albert Wiersch]

5) A single CSS of several similar files generated an error. The first line is an @import statement. CSE (batch) reported that the line did not have a closing semi-colon. The semi-colon is there. This is a utf-8 file. A diff tool showed this line and a similar line in another file to be identical, so there seem to be no "invisible characters" involved.

Was the semicolon the last character in the file? If so, then this may have been fixed when I fixed a problem that sometimes dropped the last character of UTF-8 files. If not, then can you send me files to reproduce this issue?

6) In the batch wizard the progress % indicator (bottom right) stopped at 59% (I think) even though the report was done.

That's actually a resource monitor measuring memory, not a Batch Wizard progress bar.

[MikeGale]

1) I noticed a series of false validation messages. In all cases characters, which aren't there were "seen" after the closing html tag. (These included, "he", "<", "</", '"/' and "me", with a couple of them having line breaks up front.) Without detailed checking I suspect that these particular pages are unchanged pages created on a Unix system. Other pages, which I have edited, don't show this. That might give a clue as to what is happening.

Could you email me or send me a link to a document that generates these false validation messages?

I was testing a staging version so I edited the batch file to point to the live version. Fired up CSE 9. First time no exceptions at all. Then retested staging version on IIS, got errors. Then retested live (Apache) version. This time I got some errors. I saved the test reports (4 sets), the source and the batch file that points to the live site so there is some material to investigate. (I'll sent that separately in a zip.)

Without any detailed analysis, my guess is:

1) This impacts both Apache and IIS hosted sites.

2) It could be that a newly fired up CSE 9 batch instance does not experience the issue.

3) This may only happen on repeat analysis with the batch tool.

4) I think the "ghost markup" found, is likely different every time.

5) These pages are, I think, dead / legacy pages that I have not worked on (other than via multipage S&R edits), so they may contain some markup I wouldn't use myself.

[MikeGale]

5) A single CSS of several similar files generated an error. The first line is an @import statement. CSE (batch) reported that the line did not have a closing semi-colon. The semi-colon is there. This is a utf-8 file. A diff tool showed this line and a similar line in another file to be identical, so there seem to be no "invisible characters" involved.

Was the semicolon the last character in the file? If so, then this may have been fixed when I fixed a problem that sometimes dropped the last character of UTF-8 files. If not, then can you send me files to reproduce this issue?

No the semicolon was at the end of the first line in a multiline file, like:



Code: Select all

@import url(denz.css);

body {
   font-size: 85%;
}
...

I have just rerun this test. (After seeing changes to another test.)

This time it did not show that same error. In fact it did not find any CSS sheets at fault. (There's may be a dozen of them in this test set.)

The test was a freshly instantiated version 9.

[Albert Wiersch]

Thanks! I received the files and plan to investigate tomorrow.

[Albert Wiersch]

Hi Mike,

BETA 1A is now out. Please try it and let me know how it works. I believe it should address many of the issues you posted about.

It does not address the Consolas font issue though.