"Fossies" - the Fresh Open Source Software archive 
"Fossies" - the Fresh Open Source Software archive
extended the effet of --ascii: now also converts non-ASCII characters to ASCII equivalents
updated Html2cHtml with the code from Html2Wml 0.4.11
added support for <base href="..."> in CGI mode
corrected mishandling of URLs in xlate_url() and split_card()
added option -i as an alias for --ignore-images; also available in CGI mode
corrected a bug in the man page (Werner Heuser)
added first hooks support for HTML form conversion; supported inputs are: text, password, hidden and select
added support for internal hardcoding of proxy server
added missing proxy support (Frank Schwarz)
in Html2cHtml: the split_card() function has been partially fixed,
but as I'm not satisfied with it (and it's may be impossible in
cHTML the same way as in WML), the document splitting has been
completly disabled
added a default "encoding" parameter (valued to "iso-8859-1")
added a test to prevent incorrect auto-nesting (i.e.: <p><p>...); was also introduced a "unique" attribute in the tags tree
corrected a bug in the syntax repair engine (reverse tags were incorrectly added) (signaled by Werner heuser)
added the "skip"/"end_skip" actions
added the "for" action parameter
in comment_tag() handler: the "include" and "fsize"
actions now take into account the size of the text
being added to the output
xlate_url(): fixed two big bugs the URL translation in CGI mode
just saw that Getopt::Long::config() is deprecated; replaced by Getopt::Long::Configure()
get_url(): fixed a stupid bug in the output (Ouyang)
in start_tag() handler: rewrote the part of the
syntax repair engine that tries to insert the good
tags in order to restore a valid nesting;
thanks for Ouyang for pointing me to the base of
the problem
cgi_error(): changed the output so that the version number is printed
corrected a bug pointed by Werner Heuser: other schemes were messed up
removed the XML::LibXML validation test as it is currently too unstable (it always segfaults when the doc is not valid..)
added a small optimization when correcting incorrect entities
updated the security fix in get_url(); now "file:"
schemes are also forbidden in the LWP UserAgent
completed URL encoding
tracked down and corrected a bug that occured with
certain files: split_card() was incorrectly called
in some cases
unknown entities are now converted in order to keep the document well-formed
paths in include actions can now be relative
several special characters are escaped in URLs
align attribute are now always converted to lower case
corrected several braindead bugs: upper case <title> weren't recognized; links conversion was broken in CGI mode; plus a few other stupid bugs
added warnings in read_file() and get_url()
corrected a security issue pointed by Jukka Salomaa; URLs with the "file:" scheme are now forbidden
corrected a bug in the cgi mode: options weren't kept through the links
changed some parts of the code to try to gain speed
accidentaly corrected a bug in CGI mode :-)
corrected mistakes in the package
small bug corrected
documentation updates
Html2Wml comment deleted
convert_entities(): corrected a small bug
added small optimizations in the URLs constructions
documentation has been updated
added a workaround: wmlc doesn't like < in attributes
WML compilation now correctly works
corrected several minor things
end() handler: the card tag now cleans up the stack; the prev link
is added before the last card tag
output handling didn't work, it's corrected now
usage message has been updated
small optimization in the generation of the static part of self URLs in CGI mode
corrected a small bug in the entities conversion
corrected a warning
apply_options(): added security checks to forbid dangerous functions in href and src templates
convert_entities(): corrected another bug in the regexp that close incorrect named entities; now it should work in most cases
get_url(): corrected a bug reported by Igor Khristophorov
convert_entities(): corrected a bug in the regexp that close incorrect named entities
html2wml(): added a fix for badly formed DOCTYPE (pointed by Igor Khristophorov)
some clean-ups
added code to handle unclosed named entities
corrected a small bug in the code that handle unclosed numeric entities
corrected a mistake in the handling of frames and image maps links (pointed by Igor Khristophorov)
in CGI mode, now always converts options to their short names
split_card(): corrected a small bug: respect server port
changed a few things to clearly indicate the static configurable part of the code
corrected a stupid bug (pointed by Werner Heuser)
corrected some bugs that were occuring when processing image maps
corrected a few other bugs
changed a little the presentation of the source code
html2wml(): corrected a bug: https urls weren't handled correctly
the LWP agent is now created only once per session
added a few collapsing rules
get_url(): corrected two bugs; HTTP authentication is beginning to work
xlate_url(): corrected a bug: respect server port when creating relative URLs in CGI mode
work in progress: HTTP authentication support
added function cgi_error()
added option -o, --output to write result to the given file instead of stdout
opening tags forbidden inside <a> tags
split_card(): corrected a bug (deck-split in non-cgi mode)
some warnings were corrected
corrected several warnings; this also add another correction for very bad HTML documents
added support for rendering <frameset> and <area> tags
added full support for Latin Extended-A entities plus some other non standard entities
split_card(): doc_uri is a server-relative URL whenever possible in order to save up a few bytes
id bug corrected (Igor Khristophorov)
corrected a bug when closing anchor tags
collapse_spaces(): some improvments
added support for some non-standard entities
post_conversion_cleanup(): special characters present in a title are now converted (bug pointed by Igor Khristophorov)
some bugs corrections
xlate_url(): a part of the link is now generated only once, which should give a nice code speedup
htmlize(): links are now clickable
xlate_url(): now uses ';' as parameter separator instead of '&'
early support for deck splitting (in CGI mode only)
small changes in the settings of the HTML parser
bug correction: HTML pure anchors (tag A with only a NAME attribute) are now correctly converted to WML anchors
split_card(): added a [prev] link; label is controled by the
--prev-card-label option
documentation updated
short options are available; they sould be used in the CGI mode in order to save up some bytes
xlate_url(): in CGI mode, generated URLs use server-relative URLs in order to save up a few bytes per link
htmlize(): small improvments in the syntax highlighting
HTTP headers: corrected the charset attribute; added the Content-Length header
added timing in the debug mode
unclosed numeric entites are now closed (after a question from Daniele Frijia)
xlate_url(): in CGI mode, generated URLs no longer use the absolute URL of the CGI in order to save up a few bytes per link; suggestion of Igor Khristophorov
the --img-alt-text option should now work correctly
doc updated
first attempts to fix some new bugs found by Igor Khristophorov
added the --img-alt-text option to control whether the images are replaced
by their alternative text; on by default; thanks to Igor Khristophorov for
the suggestion
cleaned up my debugging code; added the debug() function
corrected a weird bug pointed by Igor Khristophorov
several internal cleans-up
added support for sending compiled WML
added support for sending encoding information (from the patch submitted by Igor Khristophorov)
added the --numeric-non-ascii option to control whether the non-ASCII
characters are converted into numeric entities; off by default
non-ASCII characters are now converted to numeric entities
alone ampersand characters are now converted to the corresponding entity
corrected a small bug that prevent proper handling of https URLs
corrected a small mistake in the HTML syntax highlight
changed a few things in the doc
changed the default maximum size of a card to 1,000 bytes
corrected a mistake in the doc
debug mode doesn't use Text::Wrap anymore, it's really too messy.
I've written a small function, simple_wrap(), that does what I had in mind
added support for very bad-formed HTML documents (with no HTML or BODY tags)
added the --ignore-images option
corrected a small bug in doc/Makefile
added a section to the documentation
complete update of the documentation
improved the debug mode: now use Text::Wrap to avoid very long lines
corrected the parts of the code that generated the bunch of warnings; some of the bugfixes should speed up the code a little
lowered the default max-card-size value to 1,500 bytes
added attributes conversion
the software is now correctly packaged, and has a Makefile that checks the system configuration and installs the software and the documentation
changed the link reconstruction engine: it now uses Text::Template, so you can do whatever you want with the links
added support to compile WML using the compiler from WML Tools
corected a small internal flaw: now, in CGI mode, if the url parameter has
no value, use / as a default value
corrected a small bug in the link reconstruction engine
the card-split-threshold was not available because of a mistake; this is
correctd
the way spaces are collapsed has been changed a little
the no-pre option is now nopre
corrected a bug when getting the document title
corrected a bug in the text handler (skip mode was ignored)
ampersand characters are now escaped in URLs
added the no-pre option to disable the use of the pre tag
cleaned up things in the card splitting process; this correct a bug that could occur in certain circumstances
documentation has been updated
dollar caracters are now escaped
added a small HTML syntax highlighting when debugging in CGI mode
several bug corrections and improvments in the syntax correction engine
named HTML entities are now converted to numeric entities, this is safer because direct conversion to Unicode didn't seem to work
corrected the links reconstruction engine
named entities are now also converted when in attributes
new feature: the size of the cards can be limited using the max-card-size
option
corrected a bug in the de-nesting process
added features to the links reconstruction engine
corrected the local file disclosure vulnerability
until now, captions tags were always skipped; now they appears when tables are linearized
named entities are converted to real Unicode characters
non-breaking spaces ( ) are converted to spaces
links reconstruction is finalized
added the ascii option to convert entities to US-ASCII
named entities conversion is finalized
added preliminary support for named entities conversion and images support
oops! file download was broken in 0.3.1; this is corrected now
ampersand are now coded as & in links
a long standing bug that appeared in some cirumstances has been fixed
added a small SSI-like engine (not heavily tested)
corrected a stupid bug: I forgot to add the line that translate the internal links so that the urls point to the CGI
modification of the user agent signature so that it includes the signature of the original client
rendering of header tags H1, H2 and H3 has been improved
added a links reconstruction engine (currently unfinished)
improved the debug mode when launched from shell
the syntax repair engine has been improved and should now be able to correct most usual HTML pages (I hope)
corrected a stupid bug in the CGI mode
added the debug option
corrected some mistakes in the documentation
added the LINKS section in the documentation
added several improvements and corrected several bugs
added the collapse option to control whether redundant white space
chars are kept or not
added the documentation for the options
added the linearize option (on by default) so that it linearizes
tables
normal texts (i.e. not inside <pre> tags) are cleaned by stripping multiples spaces and lines
the syntax repair engine has been rewritten, but it does not work in all cases
first try to add syntax check/repair
added URI translation
added XML checker; the output is given to Expat to check that it is well-formed XML. Note that XML::Parser is loaded on demand.
added options support (using Getopt::Long)
cleaned up a few things;
added some attributes to the allow list
added url resolution
first functional version; now uses HTML::Parser v3
prototype version