TVxb 1.0108 - Official Release
Release date: September 26, 2007
New Features
Changes
Bug fixes
- Characters not
compliant with the W3C XML language specifications 1.0 are passed
through unchanged from the listing web-site to the xmltv xml output
file. Versions 3.0 and later of the MSXML parser strictly enforce
the valid XML character ranges thereby invalidating the xmltv
output.
Symptoms include:
- Error message
when opening the xmltv.xml file in Internet Explorer : "An
Invalid character was found in text content."
- BladeRunner will
not process the file for loading into Media Center.
The following
characters are now removed from all UTF-8 formatted xml output:
#x0 - #x8 (ASCII 0 - 8)
#xB - #xC (ASCII 11 - 12)
#xE - #x1F (ASCII 14 - 31)
This problem is known
to affect at least the Singapore listings.
TVxb.ini file changes
TVxb 1.0107 - Beta Release
Release date: July 29, 2007
New Features
Changes
Bug fixes
TVxb.ini file changes
TVxb 1.0106 - Official Release
Release date: May 21, 2007
New Features
-
The creation of dummy programmes to fill in gaps in the schedule can
be suppressed by adding the following item with no arguments to the
top section of the TVxb.ini file:
schedulefiller=
Changes
-
A wider range of
URLs can now be followed to fetch full descriptions from another
web-page
Bug fixes
- Part of the
program title text might be discarded if the program time and title
text are separated by only spaces on the listing page.
- An error message
is displayed in console when xmlencoding is set to something other
than iso-8859-1 or utf-8 (primarily eastern european character sets)
and no htmlencoding specified.
-
Hebrew text is not inverted by the
special=righttoleft item if the schedule is encoded in UTF-8.
-
Cyrillic text encoded in
iso-8859-xx or windows-xxxx characters sets might be discarded if
language is set to English (lang=en).
TVxb.ini file changes
TVxb 1.0099 - Official Release
Release date: August 13, 2006
New Features
-
Support for Vista Beta 2
-
Windows Installer kit
-
Multi-byte time separators are now recognized in time
strings. For example "10：00" is recognized as a valid time as
it "10:00".
(However other representations of digits 1,2,3,... in other
languages or encodings other than US-ASCII are not accepted at this
time. If there is a need for this, then let me know through the
feedback page.)
-
gb936 encoding is
now supported
-
Multiple
externally-generated xmltv files can now be merged
-
Channels names in
externally-generated xmltv files are now included in the MCE
ChannelInfo.xml file
Changes
-
The post-data used
to fetch the main listing pages is no longer sent when following
URLs to fetch detailed descriptions. (Most detailed listings can be
fetched with a single URL and without any postdata. On the otherhand
sending postdata when it's not expected is likely to cause the fetch
to fail. If is that's meaningless, then the bottom line is, this
change supports a wider range of on-line listings.)
-
The text found using
the textstart pattern is discarded instead of being used as program
data
-
The only valid pre-
and post- commands are now .bat files located in the TVxb\bin\
folder. The prepost flag file is no longer needed to run these
commands. This change was made for compliance with the Vista and
improve overall security.
-
The cached html
files are now purged every 10 days rather than monthly
-
<desc>, </desc> tags
and the description text is now always output to the xmltv.xml file
as a single line with no line-breaks.
Bug fixes
The following bugs are fixed in 1.0099:
TVxb.ini file changes
-
xmltvimport:
now accepts a comma-separated list of externally generated xmltv
files to merge
TVxb - V1.0085
Release date: November 20, 2005
Note:
There are a number of new features in
releases 1.0080 and 1.0085 that allow program attributes such as
categories, actors, directors etc. to be extracted. However the country
.ini files have not been updated to take advantage of these features as
yet.
New Features
·
Timezones can now be converted to any other timezone
(previously only UTC was possible) by using the xmltimezone=
item
·
Start/Stop times in
xmltv files imported from an external source can now be
converted to another timezone
·
Support for iso-8859-10, iso-8859-13 and iso-8859-16
character-sets
Changes
-
Eastern European characters are now output as single-byte characters
rather than character entities when an output encoding of iso-8859-*
or windows-* is used. This greatly improves the readability of the
xml produced. (However, there is a risk it might be slightly less
compatible with EPGs that are not fully internationalized. Let me
know if you have any problems.)
-
Actor and
director name identification extended to eastern European languages
-
The encoding of
xmltv files imported from an external source is now checked and
corrected
more rigorously
-
Applied latest revisions to the Unicode mapping tables for more
accurate utf-8 conversion
Bug fixes
The following bugs are fixed in 1.0080:
-
Full-stops
“.” are no longer removed from the end of description lines
TVxb.ini file changes
·
htmltimezone:
replaces the timezone item
·
xmltimezone:
convert all start/stop times to the given timezone
·
special=utctime:
has been removed and replaced by the xmltimezone item
·
textinclude, textexclude:
only retains shows that include or do not include a given text string
·
sethtmlcontinue, sethtmlend: can be used to control the processing of sets of pages that
make up the schedule for a single day.
TVxb - V1.0080
Release date: October 13, 2005
New Features
Changes
- A wider range of title links can be
followed to load full descriptions
- Maximum full description downloads
increased to 300/day. The download quota is better spread across all
channels.
- Many online listings start 5am / 6am
in the morning. Whenever TVxb is run, there will be no schedule
information available between midnight and the first program time.
Bug fixes
The following bugs are fixed in 1.0080:
- Fix for looping problem affecting the
Finland listings.
- Fix for multi-day schedules starting on
a day other than Monday
TVxb.ini file changes
-
extractelement :
Text can be extracted from titles,
subtitles, descriptions, category and rating fields based on patterns.
The extracted text can then be used as a second title, subtitle,
description, category, rating, or for credits such as director or actor.
This replaces the splitoff... items
-
descriptionlclip,
descriptionrclip, descriptionlcliplast, descriptionrcliplast
: same as the title*clip items - clip text from the left or right
of the description
-
continueshow :
specifies the text that indicates a show is a continuation of the
previous show
TVxb - V1.0069
Release date: June 1, 2005
New Features
- Title, sub-title and
description text strings can be limited to a maximum length. This is
primarily for MCE EPG support where the length of these elements must be
constrained. (Alternatively upgrade to QuickGuide RC2.)
Bug fixes
The following bugs are fixed in 1.0069:
- It is possible for
TVxb to write out a programme element with a null title. This will cause
the present release of QuickGuide to fail to load into MCE.
Shows with blank titles, or titles containing only punctuation marks,
will now be discarded.
TVxb.ini file changes
-
maxtitlelength=
and
maxdescriptionlength= : limits the maximum length of programme
text strings – primary for MCE compatibility. (maxdescriptionlength
should be set to 500 for any TVxb.ini file that configures one or more
channels with sub-titles. An even better fix is to upgrade to QuickGuide
RC2 or higher.)
TVxb - V1.0068
Release date: May 8, 2005
New Features
- The xmltv output
file is no longer held open while downloading schedules. This was
causing errors in some EPGs as they attempted to periodically access the
file.
- Recognition of grid
schedules containing multiple days of program information on one page is
now improved.
- Minor text
formatting changes. Leading commas are removed from titles, descriptions
and other program elements.
Bug fixes
The following bugs are
fixed in 1.0068:
-
Summer time processing is fixed. A bug introduced in
1.0057 causes invalid xmltv to be generated when summer processing is
enabled using the summerstart= or summerend= items.
- A bug was introduced
in 1.0065 that broke manual shows. The following message is displayed in
the log file; no shows are defined for that channel:
Warning: No URL provided
for channel xxxxxxxxxx. Discarding...
- Html
links to a program description page were not always followed for
listings that show an entire week’s schedule on one page.
TVxb.ini file changes
-
special=
ignoreallformatting
: new item that instructs TVxb to ignore all text formatting tags. This
can be used where a listing source has been hand-edited or uses some
other none-regular html formatting and TVxb is unable to recognize
the format. It is the same as using an item of:
special=ignoreallformatting,ignorebtag,ignorebigtag,ignorebrtag,ignoredivtag,ignoreemtag,ignoreitag,ignoreptag,
ignoresmalltag,ignorestrongtag,ignoreutag
TVxb - V1.0065
Release date: May 1, 2005
New Features
The main features in 1.0065 are:
1.
support for East Asia encoding
2.
enhanced title processing to better support series recognition.
Encoding
Support
-
Support for East-Asia
and other encoding: big5 (HK & Taiwan), gb2312 (china), shift_js
(japan), ksc5601 (korea). Also koi8-u (ukrainian), koi8-r (russian), and
iso-8859-15
-
Encoding is now determined from the
html meta data; in most cases htmlencoding= does not need to be
configured
-
xmltv output defaults to utf-8 encoding
although this can be overridden by the xmlencoding= item
MCE
Support
Enhanced
matching of subtitles, categories, descriptions etc.
-
Enhanced title
processing. Titles can be split off using a selection of search patterns
and distributed between title, sub-title, description, category and
rating elements.
-
TVxb attempts to
recognize category information occupying its own column (but not
embedded in the title)
Performance Improvements (for TVxb and the listing providers)
-
TVxb CPU usage should be
reduced 2 - 40 times (although download speed is still the main
bottleneck.)
-
On the otherhand, since release 1.0057,
memory usage has increased. TVxb consumes ~32MB
-
Downloads of separate description pages
by following links off the schedule is now capped to reduce loading on
listing provider servers. This is in the interests of us all. (Please -
download only as much as you need.)
Misc.
Features
-
Additional non-standard
time patterns recognized e.g. 7:2am
-
The lang=”xx” attributes
can now be optionally suppressed in the xmltv output. This is required
for some EPGs that that do not fully comply with the xmltv.DTD.
-
The terms and conditions
have been updated and the new file is required to run 1.0065
-
The xmltv output is
watermarked to deter commercial distribution or other non-personal use
of grabbed schedules.
Bug fixes
The
following bugs are fixed in 1.0065:
-
Errors in program schedule when using
multiple htmlstart or textstart patterns
-
TVxb might exit due to
an array out-of-bounds access
-
Out-of-order show
processing was broken in release 1.0057
-
There was a bug in the
html comment processing – 4 characters beyond the end of the comment
were being wasted (I’m surprised this day-one bug did not cause more
havoc)
-
href references
to description pages were being converted to lower case. This prevented
some description pages being downloaded where the URL is case-sensitive
-
downloadmethod=force
should have been limiting downloads to the next 3 days ahead and
reverting to cached files after that, however, this was not working.
-
Invalid pages were not always flushed
out of the cache and replaced with new downloaded copies
-
Invalid stop times might
sometimes be calculated resulting in missing shows. (This affected the
Swedish listings in particular.)
TVxb.ini file changes
-
mcechanneloutput= : new item. Specifies where to write the
ChannelInfo.XML file
-
splitofftitle, splitoffsubtitle, splitoffdescription, splitoffcategory,
splitoffrating replace the titleseparator and
titleposition items
-
localday= : new item can be used to specify Chinese days of the
month (or other non-European number systems)
-
special=nolang : new item to suppress the lang= attributes from the
xmltv output
-
downloadmethod=cache : “cache” replaces “reuse”
Known Errors
Summer-time processing is seriously messed
up. Although most countries will have made the transition by now, if you
have either summerstart= or summerend= items in TVxb.ini,
and you process schedules overlapping one of these dates, then the xmltv
will be invalid.
Workaround: remove summerstart and
summerend items from TVxb.ini. A permanent fix will be
released before the end of summer.
TVxb - V1.0057
Release date: 24th
February 2005
New Features
-
All shows are now buffered in memory and only written out to
the xmltv file after all downloads are complete.
A future release of TVxb will use this to provide features such as
automatic series and episode detection; enhanced schedule time error
detection and so on.
-
html tag pattern matching improved
-
< em>, <big>
and <small> html tags now recognized
-
tagtr= and tagtd= matching is now case-independent
-
Multiple htmlstart and textstart items can be specified for
each channel (up to 10 per channel.)
i.e.
textstart=[day=dd Mmm yyyy]
textstart=TV National
will first search for the current date, then search for the text “TV
National” and only then start processing listings.
-
htmlstart= and htmlend= strings : commas can now be escaped
with %2c
Bug fixes
- Fix for missing
stop= attributes in the xmltv output. This could happen when
out-of-sequence shows are detected and removed. (In particular this
could cause the EPG load into MCE to fail.)
- Minor fix to
fixedday= date patterns for fixedday=d+10 and above
- Fix to days=
item which could cause the same day to be reloaded several times
- Obsolete downloaded
html files in the \html folder (last month’s files) are now deleted.
-
TVxb will now exit
if the Terms and Conditions file is not present
TVxb.ini file changes
- Multiple
htmlstart= and
textstart= items can be specified for each channel (Refer
to the Developer's Guide for more information.)
TVxb - V1.0052
Release date: 30th
December 2004
New Features
-
Pop-up
wget windows are now hidden. To restore pop-ups, set
trace=2
in TVxb.ini .
To also hide the Console window, add "-NoConsole" to
the end of the TVxb command line. e.g.
"C:\Program
Files\TVxb\bin\TVxb.exe" -NoConsole
-
Externally generated xmltv files can now be imported and combined
together with TVxb's grabbed guide to make a single xmltv file.
Bug fixes
TVxb.ini file changes
-
xmltvimport=
item specifies an externally generated xmltv file to merge together with
the TVxb output to make a single xmltv file.
-
xmltvimportencoding=
item optionally specifies the character-set/encoding of the imported
xmltv file (although TVxb should be able to work this out itself)
-
downloadwait=
item optionally specifies the maximum number of seconds to wait for a
schedule download before giving up (default is 300 seconds)
TVxb - V1.0050
Release date: 21st
December 2004
Bug fixes
Error: Encoding conversion from ISO-8859-1
to iso-8859-1 not supported
Error: Encoding conversion from UTF-8 to
utf-8 not supported
Otherwise there are no changes from 1.0049
TVxb - V1.0049
Release date: 9th
December 2004
Bug fixes
TVxb - V1.0048
Release date: 8th
December 2004
New Features
-
XML
start and stop times now include seconds (always "00") for WebGuide and
therefore Media Center Edition EPG support (Use special=hhmmtime
to restore the original behavior)
-
UTF-8 text grabbed from web-pages is now validated and illegal encoding
removed
-
Missing schedule information (by default any gap 60 minutes or more) is
filled with dummy shows so you have something to click on in the EPG
-
Substitutions can be made for text in the title and description text.
e.g. Change all "&" strings to "and"
-
Multiple <display-name> xml elements can be defined for each channel
-
Multiple
<title> elements or <sub-title> xml elements and can be defined for each
show
-
Auto-recognition of multiple titles and sub-titles is better (although
needs further improvement)
-
Trailing period (“.”) is (usually) removed from titles
-
TVxb.ini special characters in a comma-separated list, can now be
escaped by a wider range of characters. For example comma (“,”) can be
escaped by %2C and ,
-
It is now possible to skip ahead more than one day after processing one
schedule. For example, process a weekly schedule from a single page,
then skip ahead 7 days and process the next weekly schedule
-
Extra URL date patterns added
Bug fixes
TVxb.ini file changes
-
schedulefiller=
item fills gaps in the schedule. Enabled by default; use
schedulefiller=,0 to disable
-
substitution=
items can be used to substitute text in titles and descriptions. For
example the following will substitute “ and “ for the ampersand entity
“&”
substitution=&, and
-
name=
item passes channel information to the EPG. Now you can pass multiple
values for each channel. This could be used to map the channel name to
the physical channel number. For example:
name=TVB Pearl
name=21
creates two <display-name> elements in the xml file:
<display-name
lang="en">TVB
Pearl</display-name>
<display-name
lang="en">21</display-name>
-
title=
item now supports multiple titles and sub-titles. For example:
title=2,3 creates two <title> elements in
the xmltv file using the text in positions 2 and 3
title=2;3 creates a <title> element using
the text in position 2 and a <sub-title> element using the text
in position 3.
-
subtitleformat=
formats
the secondary title or a sub-title by adding text before and optionally
after the title or sub-title text. This is useful for those EPGs that
concatenate the title and sub-title or secondary title together. (myHTPC
does this for example)
-
days=totaldays,+inc
item can now take a comma-separated increment value: the number of days
to skip forward after processing the present date. For example the
follow items both process 3 weekly schedules:
days=21,+7
days=15,+7
-
fixedday=
date pattern in the URL now supports tomorrow, t-8, t-9 and
t+1,… t+14 values
-
multiday=
item now supports yesterday, tomorrow, t+10,…
t+14 values
-
special=
items
specified in the “general” section at the start of the TVxb.ini file
apply to all channels
TVxb - V1.0042
Release date: 3rd November 2004
New Features
Bug fixes
-
Bug fix to rowspan and colspan
processing. (required for Channel U and Channel i support in Singapore)
TVxb.ini file changes
- xmlcr=
item used to suppress LF/CR at end of line in the xmltv file, or to
specify an alternative end-of-line character sequence
- precommand=
item used to specify a .bat file or executable that should run before
TVxb
- postcommand=
item
used to specify a .bat file or executable that should run after TVxb
- icon=
item to specify a channel logo picture
- multisection*
parameters -
used to support multiple day/date
schedules on a single page
-
multisectionchannel=
-
multisectionday=
-
multisectionrowstart=
-
multisectionrowend=
-
multisectioncolstart=
-
multisectioncolend=
TVxb - V1.0033
Release date: 4th September 2004
New Features
-
All shows in
xmltv.xml now have stop times. (For ShowShifter and other PVR
software that needs them.)
-
Title links can now be
followed to load full descriptions
from other web pages
-
Full description
downloads can be restricted to certain times of the day; for
example, all day during the weekends, and between 6pm and midnight
on other days
-
Improved handling of
listings with out-of-order times
-
Improved recognition of
times (and improved rejection of text strings that are not times)
-
Extra text matching
features to strip unwanted text from show titles such as episode numbers
etc.
-
Listings text can now be
delimited by date patterns as well as by html and text strings. Used to
check the schedule downloaded is for the correct date.
Bug fixes
-
Bug fix for handling of
12-hour times
-
Bug fix to handling of
special characters in channel names in the TVxb.ini file
-
Bug fix to time handling
that caused shows to be dropped from the Danish listings
TVxb.ini file changes
-
duplicatehandling=
item used to control which show is used when two shows share a start
time. (valid options are: keepfirst, keepsecond)
-
special=startprevday
used where the first show on each listing might start on the previous
day and span midnight
-
special=outofordershows
used where the listings source often displays shows not in time order
(however all shows on the page are assumed to be on the same day between
00:00 and 23:59 hours.)
-
special=fulldescriptions
used to download full descriptions from other web-pages using title
hyperlinks
-
descriptionhtmlstart=
used to specify an html or text string that will be used to delimit
the start of the description on the downloaded description web-page
-
descriptionhtmlend=
used to specify an html or text string that will be used to delimit
the end of the description on the downloaded description web-page
-
descriptionhours=
used to restrict the hours between which full descriptions will be
downloaded. This should be used to reduce download times.
-
descriptionsplit=
used to break description text into fixed line lengths
-
titlelclip=
replaces titleskip; finds the first instance of the specified
pattern, keeps text to right, and clips text to the left of the pattern
-
titlerclip=
keeps text to left, and clips text to the right of the pattern
-
titlelcliplast=
finds the last instance of the specified pattern and clips text to the
left
-
titlercliplast=
finds the last instance of the specified pattern and clips text to right
-
downloadmethod
can now be specified for
individual channels as well as generally
TVxb - V1.0025
Release date: 10th August 2004
New Features
-
Hebrew and other
bi-directional languages are now supported. Visually-encoded listings
can be reformatted using the special=righttoleft item.
-
Schedule times can be
automatically adjusted when scheduling over daylight saving transitions.
Use the
summerstart= and summerend= items to specify the date/time of
the time change. (However it is still necessary to edit the TVxb.ini
file with the date/times once every year)
-
Show start and stop
date/times can be converted to UTC-based time using the special=utctime
item. This might be useful if your PVR software have difficulty working
with timezone information.
-
Overlapping schedules
are now handled correctly. Repeated entries are discarded.
(Although it discards the 2nd
overlapping entry. In most cases discarding the first entry would be a
better choice. In another release…)
Bug fixes
-
Stop attributes in the
xmltv.xml file now explicitly state the timezone as for start
attributes. This corrects a problem where myHTPC/GotTV would display
incorrect show information while watching Live TV.
-
Corrected bug in HTML
rowspan
and colspan attribute handling. xxxspan=’n’ is now recognized as
well as xxxspan=”n”. This could have caused shows to be dropped.
TVxb.ini file changes
-
timezone=
item can now be specified a default for all channels as well as for
individual channels. This can be used when most or all channels are
located in the same timezone.
-
summerstart=
and summerend= can be used to specify the date/time for daytime
saving transitions
-
special=righttoleft
used to correct visual encoding for languages such as Hebrew. It will
also switch to right-to-left text searching and truncate long titles on
the left.
-
special=fewertimechecks
can be used is partially disable some of the new date/time sanity
checking in this release
-
titleflip=
replaces fliptitle= for consistency with other “title” items.
(TVxb.ini will be autoconverted)
Known Errors
- Showshifter
compatibility: Not directly compatible with Showshifter. The
HMNXMLTV.exe importer
tool expects stop attributes for each show. (Will be provided in the
next TVxb release.) Workaround for the moment is to process xmltv.xml
with the xmltv tv=sort function.
(Not provided with TVxb.)
TVxb - V1.0021
Release date: 23rd July 2004
New Features
-
The TVxb.ini file can
now be encoded in UTF-8 format as well as the usual ANSI format.
-
Web-pages encoded in
UTF-8 and Windows proprietary encoding (such as WINDOWS-1255) are now
supported.
-
Conversion from ISO-8859
and WINDOWS encoded character-sets to UTF-8 is now supported
-
Performance is
significantly improved when downloading large numbers of channels or
days, especially when using dial-up or slow internet connections
-
Show start and stop
times embedded in the title can now be recognized by using the new
special=startintitle
and special=stopintitle items
-
Title and show
descriptions that are combined into a single text string can now be
separated based on punctuation marks. The minimum title length can be
set by the
titlesplit= number item; the punctuation characters can be specified
by the titlesep= item
-
Auto-recognition of show
details on web-pages is improved including:
-
Improved recognition
of show descriptions
-
Improved recognition
of multiple days on a single page
-
Improved recognition
of start and duration times
-
The maximum number of
manually defined shows allowed is now increased
Bug fixes
-
Program exception due to
array overflow when processing large numbers of manually defined shows
-
Parsing of "&" and
character-entities could under some situations cause an attempt to
extract a string of negative length resulting in program termination
-
Some <br> html tags were
not being recognized
-
Program times are now
correctly propagated down each row of an html rowspan column
-
Program titles are now
correctly propagated across each column of an html colspan row
TVxb.ini file changes
-
Comments are now
preserved when new TVxb.ini files are written
-
Error checking and
reporting for
special= items is improved
-
end=
is replaced by stop= in line with xmltv naming (TVxb.ini will be
autoconverted)
-
period=
is replaced by multiday= (TVxb.ini will be autoconverted)
-
alttimesep=
is replaced by timesep= (TVxb.ini will be autoconverted)
-
special=splitlongtitles
item is replaced by
titlesplit=
item (TVxb.ini will be autoconverted)
-
titlesep=
item added to param section and channel section (defaults to “ . : ,
* ”)
-
htmlstart, htmlend,
titleskip
and titleblock patterns are now case independent
Known Errors
-
UTF-8 downconvert to
other encoding schemes: Conversion from UTF-8
to other encoding schemes is not supported. There are no plans to
support this.
-
Unicode
(UCS-2 and UCS-4)
encoding:
Unicode
(UCS-2 and UCS-4)
encoding is not supported in web-pages, in the TVxb.ini file, or as
output to the XMLTV file.
|