* template: Preprocess parameters before htmlizing.
* img: Fix a bug that could taint @links with undef values.
* setup automator: Configure Term::Readline to use bold for prompt, rather
than default underline. Closes: #517656
* Allow jumping directly into account registration process by going to
ikiwiki.cgi?do=register
* Improve display of openid in preferences page.
* Add link to userpage (or creation link) to top of preferences page.
* opendiscussion: This plugin will also now allow posting comments
to otherwise locked-down sites.
* auto-blog.setup: Lock all pages, so only admin can post to the blog
by default, and enable opendiscussion so others can comment.
* Fix color and format plugins to appear in the websetup interface.
* amazon_s3: Fix to support the EU S3 datacenter, which is more picky
about attempts to create already existing buckets.
* httpauth: When cgiauthurl is configured, httpauth can now be used
alongside other authentication methods (like openid or anonok). Rather
than always redirect to the cgiauthurl for authentication, there is now
a button on the login form to use it.
* httpauth: Add httpauth_pagespec setting that can be used to limit
pages to only being edited via users authed with httpauth.
* Allow globs to be used in user() pagespecs.
* Group related plugins into sections in the setup file, and drop
unused rcs plugins from the setup file.
uWSGI is a fast (pure C), self-healing, developer-friendly WSGI server,
aimed for professional python webapps deployment and development.
Backend support exists for Apache, nginx, cherokee and lighttpd.
Some of the features include:
- low memory footprint
- preforking and multithreaded operation
- logging, diagnostic and profiling support
- support for sendfile()
- self-healing
- graceful restart and hot-plug upgrading
Collection.
The Perl 5 module Web::Scraper is a web scraper toolkit, inspired
by Ruby's equivalent Scrapi. It provides a DSL-ish interface for
traversing HTML documents and returning a neatly arranged Perl data
structure.
NetBSD Packages Collection.
The Perl 5 module HTML::TreeBuilder::LibXML is a libxml based
compatible interface to HTML::TreeBuilder, which could be slow for
a large document.
* Update Polish language files, supporting really 2.8.RC2.
* Add missing language files which should be added: Bosnian, Croatian,
Finnish, Hungarian, Slovenian and Tagalog.
* Drop executable permission from translation files.
- fix various bugs, including a trivial to trigger OOM/DoS
- Allow support for checking for exec bit on cgi scripts
- Add support for TLS servername indication
- Add support client certificate verification
- Split off spawn-fcgi into a separate package
spawn-fcgi is a small utility to start external FastCGI applications.
It takes care of various common tasks to simplify deployment:
- setuid/setgid
- current working directory
- chroot
- socket creation and permissions in the case of Unix Domain Sockets
- forking a number of worker processes
- fix issues with Keep-Alive in various clients
- fix issue with internal redirects between proxy/fastcgi locations
- fix a bug with discarding request bodies
Add optional uwsgi backend.
NetBSD Packages Collection.
The Perl 5 module HTML::TreeBuilder::XPath adds typical XPath methods
to HTML::TreeBuilder, to make it easy to query a document.
Version 3.49
[BUG FIXES]
1. Fix a regression since 3.44 involving a case when the header includes "Content-Length: 0".
Thanks to Alex Vandiver (RT#51109)
2. Suppress uninitialized warnings under -w. Thanks to burak. (RT#50301)
3. url() now uses virtual_port() instead of server_port(). Thanks to MKANAT and Yanick Champoux. (RT#51562)
[SECURITY]
1. embedded newlines are now filtered out of header values in header().
Thanks to Mark Stosberg and Yanick Champoux.
[DOCUMENTATION]
1. README was updated to reflect that CGI.pm was moved under ./lib.
Thanks to Alex Vandiver.
[INTERNALS]
1. More tests were added for autoescape, thanks to Bob Kuo. (RT#25485)
2. Attempt to avoid test failures with t/fast, thanks to Steve Hay. (RT#49599)
1.4.4
Development of HTTP_Request package is halted, only bug fixing will be done.
Please submit feature requests for HTTP_Request2 package.
Fixes:
* Improved memory usage of _buildRequest() method (bug #14574)
* Clarified documentation for addFile() method to mention that it is useful
only for POST method file uploads (bug #14635)
* Do not send "Content-Length: 0" header for methods other than
POST and PUT, as some servers may return error 400 (bug #14740)
1.4.3
* Added possibility to get reason phrase from HTTP response (request #12352)
* PHP 4.2.0 should be the minimal PHP version (bug #12354), also updated other
dependencies' versions to saner values
* Send a Content-Length header in request even if body is empty (request #12900)
* Do not pass length parameter to gzinflate(), it could cause problems in some
corner cases (bugs #13135, #14370)
* Return an error if trying to do a HTTPS request without OpenSSL support
(bug #14127)
* Do not stop reading chunked response body prematurely (bug #14200)
1.4.2
* The final CRLF was not properly added to request headers on POST request
with no post data (Thanks to Brock Weaver)
* Added error codes (request #12335, thanks to Joe Stump for the patch)
* HTTP_Request sent broken requests on redirects with no trailing slash
(bug #12308, thanks to Joe Stump for the patch)
* Requests with a body consisting of only a symbol '0' were sent without body
(reported privately by Sam Ghods)
* download-progress.php example was broken since 1.4.0 due to addition of
'connect' and 'disconnect' events. Now works again.
1.4.1
* Removed bogus parameter for getURL() (Bug #9586, thanks to Martin Jansen)
* Improved API documentation (Bug #9984, thanks to Martin Jansen)
* Fixed wrong Content-Length if using mbstring function overloading (bug #10605)
* Fixed bogus "data CRC check failed" error on 64-bit systems (bug #10790,
thanks to Bill Moran)
* Redone the way package handles mbstring function overloading, this will allow
using gzip Content-Encoding when said overloading is switched on
* Added proper header comment blocks, improved phpdoc comments
1.4.0
* Added Keep-Alive support (request #4806), thanks to Justin Patrin for the
initial patch. Please note that "Connection: close" header is still added
by default, you need to explicitly add "Connection: Keep-Alive" header or
remove "Connection" header if using HTTP 1.1
* A new disconnect() method was added which forces disconnection from the
server if Keep-Alive is used. Also two new events are sent to the Listeners:
"connect" and "disconnect"
* Added getUrl() method (request #6589)
* Added method to properly parse header of gzip-encoded data (see RFC 1952).
This takes care of situations when the server adds some additional data
to the header (bug #8245) or sends data that is not gzip-encoded when
"Content-Encoding: gzip" header is present (bug #8213)
* "Proxy-Authorization" header is now properly set by constructor
(bug #5913)
* Fixed doc comments mentioning addBody() method instead of proper setBody()
(bug #5969)
* Fixed erroneous removal of "Content-Type" header from request
(bug #7922)
* Bogus HTTP headers are now ignored (bug #8214)
* Path is set to "/" if an URL without path (http://www.example.com)
is given (bug #8662)
* Moved to package.xml version 2.0
1.2.5
Whitespace is no longer trimmed from quoted attribute values in
_parseAttribute() method (bug #15787)
1.2.4
Migrate to package.xml version 2.0, use proper header comment block.
No code changes since previous release.
Pkgsrc changes:
- Add LICENSE
- Adjust dependencies
- Add recommended dependencies
Upstream changes:
1.13
- ignoring dirs starting with '.'
- no HTML rendering for perltoc (too slow)
- support for perlvar
- computing TOC layout, works for MSIE6, MSIE8, FF, Chrome and Safari
- many JS fixes (done in Alien-GvaScript 1.20)
- applied patch by Elliot Shank for specifiying the page title
- tested on MSIE8, FF3.5, safari4, Chrome3
1.12
- sync TOC, automatically open nodes in TOC
- support aliases in links L<alias|http://...>
- fixed bugs in URLs containing special chars like L<http://foo$/...>
- disabled automatic links within text nodes
- links to external resources will open in a blank page
- fixed bugs in displaying scripts (as opposed to modules)
- toc ignores directories starting with '.'
- toc_pragmas ignores "pod", "pods" and "inc" subdirs
Pkgsrc changes:
- Add LICENSE
Upstream changes:
1.21 18.01.2010
- bug in Builder, Gvascript.js was written to wrong location
1.20 07.01.2009
- Element.autoScroll : fix scroll amount when the scrolling block has
an offsetTop > 0
- removed PAGE_UP/DOWN handlers in TreeNavigator -- let the browser deal
with those
- treeNavigator select() has an additional param "prevent_autoscroll"
- treeNavigator autoscroll only active through kbd, not through click
- treeNavigator observes "mouseup" (instead of mousedown or click)
- autoCompleter bug fix on detach (stop observing "click")
- protoExtensions : removed RegExp.escape() (now in prototype.js)
- autoCompleter div, display above only if enough vertical space
- autoCompleter : code layout improvements
1.18 13.10.2009
- fix some minor doc glitches released in 1.17
1.17 13.10.2009
- major revision from Mona Remlawi
- new Grid component
- new CustomButtons component
- new Paginator component
- new test suite
