Web interface & Firefox

[CCRDude]

Update: server restarted, no change.

The stylesheet is in it’s place, I can open it if I enter it’s address directly into the browser.

Idea: 9 HTML warnings… if you use XHTML Strict, you should follow the standard, otherwise use HTML Transitional 😉 The complete HTML header is invalid due to the (null)s, so maybe that’s why Firefox, strictly following the rules, doesn’t load the stylesheet?

I used the tarball with ./configure –enable-oggvorbis –enable-flac –enable-musepack –enable-sqlite. I had a 0.2.4 on the machine before that.

BUT: If it was a version mismatch, why would it work when I access it through Internet Explorer? I tried Firefox on another machine as well – same problems (all Firefox installations so far had Adblock and NoScript, but those are really two of the most-used Firefox installations).

I went and deleted the contents of /usr/local/share/mt-daapd/admin-root (where my mt-daapd.conf points to for the webroot), and copied the contents of mt-daapd-svn-1376/admin-root into that folder. I cleared the browser cache and accessed it again – same error.

I then used chmod to root:root and chown to 555 on the whole admin-root folder (the permissions and owners from the tarball where quite mixed) – no change.

I used locate smartpopup.html to see if any other directories exist (in case the webroot parameter wouldn’t be working) – that’s not the case.

The only executable I found is /usr/local/sbin/mt-daapd. It’s really to bad it doesn’t show any version :-/ But the file date is September 7, so it’s the result of my SVN installation (config.status in the svn folder etc. have the same date).

My final conclusion would be the (null)s in the header again:

(null)

(null)










(null)

(null)



(null)

(null)



(null)

(null)

Looks very much like under some conditions, ispage is returning (null) instead of an empty line. And comparing that to the hdr.html, I found a possible explanation for that:
* On the current page, the line is written, but cut too early, after character 62!
* On other pages ispage, src-fields with a file in the same folder return an empty line, and
* src-fields with a file in a subfolder return (null).

[rpedde]

@CCRDude wrote:

Idea: 9 HTML warnings… if you use XHTML Strict, you should follow the standard, otherwise use HTML Transitional 😉

I didn’t see that patch in my email? 😉

The complete HTML header is invalid due to the (null)s, so maybe that’s why Firefox, strictly following the rules, doesn’t load the stylesheet?

Very likely, but I didn’t know where the (null) was coming from, as I’ve never seen that behavior before.

Looks very much like under some conditions, ispage is returning (null) instead of an empty line. And comparing that to the hdr.html, I found a possible explanation for that:
* On the current page, the line is written, but cut too early, after character 62!
* On other pages ispage, src-fields with a file in the same folder return an empty line, and
* src-fields with a file in a subfolder return (null).

That gives me something to go on. Thanks.

— Ron

[CCRDude]

@rpedde wrote:

I didn’t see that patch in my email? 😉

Maybe because most errors are related to those (nulls) for now 😉

Patch for hdr.html adding new style file, making script lines shorting by removing closing tag and adding terminating to end of opening tag instead, and removing invalid li:

--- hdr.html	2006-07-06 06:42:30.000000000 +0200

+++ hdr.html	2006-09-27 16:15:39.000000000 +0200

@@ -4,22 +4,23 @@







-  @ispage index.html::@

-  @ispage index.html::@

-  @ispage index.html::@

-  @ispage index.html::@

-  @ispage config.html::@

-  @ispage config.html::@

-  @ispage config.html::@

-  @ispage config.html::@

-  @ispage smart.html::@

-  @ispage smart.html::@

-  @ispage smart.html::@

-  @ispage playlist.html::@

-  @ispage playlist.html::@

-  @ispage playlist.html::@

-  @ispage playlist.html::@

-  @ispage playlist.html::@

+  @ispage index.html::@

+  @ispage index.html::@

+  @ispage index.html::@

+  @ispage index.html::@

+  @ispage index.html::@

+  @ispage config.html::@

+  @ispage config.html::@

+  @ispage config.html::@

+  @ispage config.html::@

+  @ispage smart.html::@

+  @ispage smart.html::@

+  @ispage smart.html::@

+  @ispage playlist.html::@

+  @ispage playlist.html::@

+  @ispage playlist.html::@

+  @ispage playlist.html::@

+  @ispage playlist.html::@







@@ -49,7 +50,6 @@



about firefly


thanks


- 

Patch creating new css file for index.html since style inside the body is not allowed:

--- index.css	1970-01-01 01:00:00.000000000 +0100

+++ index.css	2006-09-27 16:07:23.000000000 +0200

@@ -0,0 +1,31 @@

+#service, #stat {

+  width: 70ex;

+}

+#thread {

+  width: 80ex;

+}

+#server_stopped_message {

+  position: absolute;

+  background-color: yellow;

+  padding: 1em;

+/*  border: 1px solid #8CACBB;*/

+  font-size: 120%;

+  margin-bottom: 1em;

+  z-index: 100;

+

+}

+#grey_screen {

+  position: absolute;

+  top: 0;

+  left: 0;

+  width: 100%;

+  height: 1000px;

+  background-color: #000;

+  filter:alpha(opacity=60);

+  opacity: 0.6;

+  z-index: 10;

+}

+#button_stop_server, #button_start_scan, #button_start_full_scan {

+  width: 10em;

+  margin-bottom: .2em;

+}

Patch removing invalid style definition from index.html:

--- index.html	2006-07-06 06:42:31.000000000 +0200

+++ index.html	2006-09-27 16:07:37.000000000 +0200

@@ -1,37 +1,4 @@

@include hdr.html@

-



Server Status

update: By the way, by shortening the length of the script lines (closing the tags inside the opening tag instead of the additional closing tag), the number of (null)s was reduced. So the problem about the (null) seems to have to do with the length cutting as well (maybe when I made that list above, the files in subdirectories where just unlucky to have a longer total line length).

[rpedde]

@CCRDude wrote:

Patch for hdr.html adding new style file, making script lines shorting by removing closing tag and adding terminating to end of opening tag instead, and removing invalid li:

The self-closing style of script tags make IE ill, but the rest I’ll happily take. Still haven’t looked closer at the null stuff… I’m still struggling with getting an automated build system. Might be this weekend before I get it all put together in a new nightly.

— Ron

[CCRDude]

@rpedde wrote:

The self-closing style of script tags make IE ill,

Are you sure about that, or is it because the forum added linebreaks (will send it as files by email then next time 😉 )? I have no problems at all with either IE 64 bit nor IE 32 bit (both 6.0). But then, that may have to do with the cutting – because of that, I guess the terminating isn’t considered by any browser in my environment, so you could be right 😉

[CCRDude]

Ok, as part of my attempts to make the system more Unicode even on the console, I did actually create a locale (de_DE.UTF8). Guess what this fixed as a side-effect? 🙂 I’ll play around some more tomorrow, since I also made changes to make paths shorter, but it seems it’s working now finally.

[rpedde]

@CCRDude wrote:

Ok, as part of my attempts to make the system more Unicode even on the console, I did actually create a locale (de_DE.UTF8). Guess what this fixed as a side-effect? 🙂 I’ll play around some more tomorrow, since I also made changes to make paths shorter, but it seems it’s working now finally.

Ooooooh… interesting. This was German, right? There isn’t any multi-byte characters in German, are there? Just a different codepage?

Or are there?

[Jxn]

In UTF-8 everything outside of pure ASCII (7 bit) is multi byte. So as in swedish (äöå) you have multi character codings of ü,ß and ã(?) in german UTF-8.

[rpedde]

@Jxn wrote:

In UTF-8 everything outside of pure ASCII (7 bit) is multi byte. So as in swedish (äöå) you have multi character codings of ü,ß and ã(?) in german UTF-8.

Yeah, for utf-8. I understand that. But some non-utf encoding systems are multibyte. Big5, shift-jis, etc. I was asking if the native (non-utf) encoding for german was multibyte. If strlen != bytelen, that would explain those problems.

— Ron

[fizze]

no, german is single-byte. just a different codepage. 850, iirc.

[Author]

Posts