Changes included in version 2.0.4
---------------------------------
- Fixed printing under windows.
- You can now plot the real and imaginary parts of the
circuit input and output impedances
Changes included in version 2.0.3
---------------------------------
- Moved to Qt version 2.1.0
- The help browser is now integrated into the ViPEC application.
- Fixed a bug in calculating the circuit output impedance.
- Input and output impedances ouput to a graph now works.
Changes included in version 2.0.2
---------------------------------
- Improved data storage of graphs and Smith charts
- Double buffering implemented for drawing of graphs
and Smith charts to reduce flicker.
- Improved font management
- Fixed display bug in Table view
- Fixed bug with output to 2 port parameter files under Windows
- Added support for Group Delay, see help files for more details
as well as sample circuit: group_delay.ckt
o Added the missing support functions c_chstrlen, c_Screen2sv.
Oops. :)
o Updated DIAGNOSTICS section of pod.
o Added typedef detection support.
o Rearranged PANELS support a little.
o Removed support for wattr_get, wattr_off, wattr_on, and
wattr_set (for now). I got the prototypes from ncurses-1.9.9g,
but they apparently differ from the "standard" ones.
o Renamed "lines" everywhere, 'cause some curses library writer
was stupid enough to #define it. (Go figure.)
o Reworked "optional" arguments, which only affects newterm(),
panel_above() and panel_below(). Now, if you want to specify
"NULL" as an argument, use "undef" instead of omitting the
argument. (Don't worry if this doesn't make sense. You
probably aren't affected anyway.)
o Added the ncurses "curs_mouse" functions, but I have no idea
if they work. I couldn't even get the C versions to work, so
I had no way of testing them. Feedback on these functions
would be appreciated.
xsane-0.51 -> 0.52:
-------------------
- changed gimp_install_procedure(...) for specified device: "RGB, GRAY"
-> 0
- in xsane-scan.c and xsane-preview.c added " | GDK_INPUT_EXCEPTION" to
gdk_input_add(fd, GDK_INPUT_READ | GDK_INPUT_EXCEPTION,
preview_read_image_data, p);
- added static int xsane_set_modus_defaults(void)
zoomfactor is set to 1.0 if copy mode is selected
full preview area is selected if copy or fax mode is selected
- removed a second xsane_device_preferences_restore(); in static void
xsane_device_dialog(void)
- zoom factor keeps unchanged when printer is changed
- added psfile-geometry to setup/saving options, it is used when saving
postscript in scanmode
- moved exisiting routines to open functions to xsane-front-gtk.c:
void xsane_define_output_filename(void);
void xsane_identify_output_format(char **ext);
- define_maximum_output_size is called in each routine that changes
filename or fileextension because the output frame is also drawn in
scanmode with filetype postscript now
- included Stanislav Brabec's <utx@penguin.cz> sed-scripts and changes
for Makefile.in in xsane-*/po
- added creation of newlanguage.po in po/Makefile.in
- updated german translation file de.po
- updated czech translation file cs.po - thanks to Stanislav Brabec
<utx@k332.feld.cvut.cz>
- updated french translation file fr.po
xsane-0.52 -> 0.53:
-------------------
- added test if vendor, model and type strings do exists (pointer != 0)
otherwise "no vendor",... is printed
- changed
ext = &extension;
to
if (extension)
{
*ext = strdup(extension);
}
else
{
*ext = 0;
}
in xsane_identify_output_format - bad thing, I get old ;-)
- added new routine xsane_find_best_resolution that makes sure that the
selected resolution fits into range/quantization given by backend.
This is used by xsane_set_resolution and when resolution_list is
created.
- added xsane_set_all_resolutions to xsane_set_modus_defaults
- xsane resolution can be changed by backend now
xsane-0.53 -> 0.54:
-------------------
- updated french translation table
- updated german translation table
- changed type of umask in preferences from mode_t to int because
sizeof(mode_t) is on some machines only 2 bytes what cause problems.
all calls of umask(foo) are changed to umask((mode_t) foo)
- changed preferences update_policy from GtkUpdateType to int.
- selection box can not be moved out of the preview window, the
size of the selection keeps unchanged in that case now.
- removed some gtk_main_iteration and reordered some selection drawing
routines.
- removed unused files/directories from package:
sanei, include/sane/sanei_(some).h, po_old
xsane-0.54 -> 0.55:
-------------------
- added preview support for 16 bpp (transformed to 8bpp)
- added copyright line in devcie selection dialog
- preview selection frame is only active when backend supports coordinate
selection
- added preview oversampling factor (multiplier for calculated preview
resolution)
- added lineart enhancement: preview can be done in grayscale now,
threshold value change is displayed without a new preview scan.
- added "device options" to setup dialog:
- Scanner dependant options for grayscale->lineart conversion have
to be entered there.
- a color preview is now displayed gray when grayscale mode is selected
- updated de.po, cs.po, fr.po
- updated po/xsane-pot.sed
xsane-0.55 -> 0.56:
-------------------
- added hold timer to preview (update-policy=delayed)
- added hold timer to xsane_slider (update-policy=delayed)
- moved help menu from file to help
- added swedish (sv) translation - thanks to Johan Bengtsson
<johanb@lls.se>
- removed bug in xsane_rc_io_w_string, it returns EINVAL now each time
when the string does not begin with "
xsane-0.56 -> 0.57:
-------------------
- made xsane_choose_device return type to void (was gint32)
- arranged device-preferences in struct xsane in the order they are used
as device preferences, made sure that all variable types are correct:
- changed type of xsane.mode from gint32 to int
- changed type of xsane.lineart_mode from gint32 to int
- changed type of xsane.show_preview from SANE_Bool to int
- changed type of xsane.enhancement_rgb_default from SANE_Bool to int
- changed type of xsane.fax_finemode from SANE_Bool to int
- added mask for *_lock in preview motion_notify_event handling
Preview selection can be done with locks activated now
- histogram is cleared before/redrawn after saving/loading device settings
- added xsane_control_option() that tests if option >= 0 and returns
SANE_STATUS_INVAL if not
- added xsane_get_option_descriptor() that tests if option >= 0 and
returns NULL if not
- changed parameters of preview_area_resize and added
preview_area_resize_handler
- removed a lot of gtk_main_iterations from xsane_preview.c
they make problems in the redrawing order and can cause display
problems
xsane-0.57 -> 0.58:
-------------------
- split the preview_event_handler into
- preview_press_button_event_handler
- preview_motion_event_handler
- preview_release_button_event_handler
- the preview_expose_event_handler now is called twice,
the second call has been placed in the preview_event_handler_before
- reactivated gtk_main_iteration after at the end of motion_event
- added warning window when xsane is started as root
- added selection of temporary directory in preferences/setup/saving
options
- added gdk_input_remove(xsane.input_tag); before xsane_start_scan is
called in 3-pass scanning mode (suggestion from
Heiko_Schroeder@t-online.de)
- changed output of "xsane -v"
- in xsane_scan_done moved xsane.header_size = 0; before test for ADF
Thanks to Russell Coggrave for the info about it.
- added gtk_set_locale() before gtk_init()
- added copyright sign "(c)" to translated texts because it is not
available on all charsets
- temporary filenames are created random and unique (mktemp + random
number) due to security problems with fixed/calculateable tempfiles.
- the preview-files are created (empty) at program startup now
- when program exits the previews are erased
there is no way to store the preview images any more when the program
is closed
- preview_restore_image() creates the xsane-startimage if no preview is
stored
- renamed directory frontend to src
- create xsane-help script that opens online helpfile
xsane-0.58 -> 0.59:
-------------------
- preview_do_gamma_correction does test for 3 pass scanner color mode
- changed fax x-resolution and standard resolution from 98 ppi to 204 ppi
and with of faxpage from 210.0 to 215.7 (so we get 1728 pixels in
x direction, hylafax-faq says that is the correct size)
- split preferences.printer[..].resolution -> lineart, grayscale and
color_resolution
- added setlocale(LC_NUMERIC, "C"); in xsane_init() and #include
<locale.h> in xsane.h this should solve the problems with empty
images (100%transparent) with gimp and other problems that were caused
by local settings (1.3 <=> 1,3).
- added setup-option "disable gimp preview gamma"
- added selection (preferences/setup/enhancement options) if
autoenhancement changes gamma value
- renamed Preferences/Setup/"Device options" to "Enhancement options"
- removed xsane internal callbacks from dialog struct, the functions are
called directly now
- changed the reading of the preview image header. The last newline
character before the image data starts again made some problems, this
seems to depend on the data that follows. So I added a fgets() command
to skip the newline character - this seems to work.
- changed increase_filename_in_counter, does work now if the only
character of the base-filename is the counter
- reduced size of src/xsane-startimage.pnm (1MB => 0.6MB)
xsane-0.59 -> 0.60:
-------------------
- after filename counter is increased the cursor in the outputfilename
entry is moved to the right
- in preview_read_image_data when image is complete:
preview_update_surface(p, 0);
is called before
preview_display_image(p);
this way the geometry of the first preview scan is correct.
- moved xsane_set_all_resolutions in xsane_update_xsane_callback(),
it now also is called when the backend only uses one reesolution
- added calibration image (xsane-calibration.pnm) that is loaded when
setup window is created.
The xsane enhancement functions are disabled in calibration mode, the
image can not be resized (xscale = yscale = 1.0).
The previsous image is reloaded when the setup window is closed again.
- enviornment variable SANE_DEFAULT_DEVICE is used to activate the
radiobutton in the device selection menu
- removed unused item "custom" from preview surface menu
- changed preview surface menu items:
A3 -> A3P (portrait)
A4 -> A4P (portrait)
A4H -> A4L (landscape)
A5 -> A5P (portrait)
A5H -> A5L (landscape)
- removed script xsane-help, necessary command now is listed in
xsane.PROBLEMS
- after filename counter is increased the cursor in the outputfilename entry is moved to the right
xsane-0.60 -> 0.61:
-------------------
- added debug output function and messages
- moved all global variables and struct GSGdialog to global struct xsane
- cleaned up some variable names
- added free gamma curve dialog (not ready)
- prepared xsane for win32:
- when BUGGY_GDK_INPUT_EXCEPTION is defined only blocking mode is used
(needed because GDK input exception does not work on win32)
- on win32 you cannot get a homedirectory, for this
XSANE_FIXED_HOME_PATH has to be set to the path that shall be
used as homedirectory
- replaced all gtk_dialog_new by gtk_window_new and added hbox and
vbox to the window (gtk_dialog_new does not work on WIN32)
- added xsane.WIN32 and xsane.WIN32-compilation that describes how xsane
is build on WIN32 with CYGWIN
- replaced "/" in all paths, added label SLASH that is predefined to
'/' if it is not defined by compiler option. For WIN32 it is set to
'\' by compiler option
- all fopen functions use argument "rb" or "wb" now (b=binary). This
is needed for WIN32 and the b is ignored on unix systems.
- published binary version xsane-win32-0v01 and xsane-win32-0v02 based
on xsane-pre0.61 versions
- removed an ugly bug: in struct Preferences there was predefined a
string ("/tmp/") that caused a bad memory state when the string memory
was freed, instead the string is defined in xsane_pref_restore now
- changes for tiff format:
- added support for 16 bit tiff format.
- predefined tiff-compression is changed from 5 to PACKBITS(16bit),
JPEG(8bit), FAXG3(1bit).
- add resoultion-tags and datetime-tag to tiff files.
- added "change directory" entry to preferences menu
- added "scanning for devices" dialog
* New backends: mustek_pp, plustek, st400, v4l.
* Updated backends: Artec (v0.5.14), Coolscan (v0.4.3), Epson (v0.1.33),
HP (v0.88), Mustek (v1.0-96), Sharp (v0.31), UMAX (v1.0-21).
* Removed backend info sagitta.desc due to lack of interest.
* Installing no longer depends on 'rev'.
* Added and structured documentation.
* Every backend has a man page now.
* Debug messages are now sent to syslog if stderr is a socket.
* New 16 bit support to scanimage.
* Aliased and hidden backend support.
* Port to UnixWare 2.x and UnixWare 7.
* New script 'sane-config' to help frontend developers.
* Added Ricoh IS50 support.
* Updated backends: Artec (v0.5.13), CANON (v991108), DC-25 (1.2), HP
(v0.86), Microtek (0.12.0), Microtek2 (v0.8), UMAX (v1.0-build-18),
Epson (v0.1.15), Sharp (v0.30), sp15c (1.12) and m3096g (1.11).
* Automatic document feeder support for scanimage.
* New backend for Avision AV 630 CS.
* New Linux SCSI Generic driver.
* Upgraded to libtool 1.3.4
* GIMP compatibility patch.
* OS/2 patches.
* Bugfixes.
* Fix a bug in Latest Release Info under the Help Menu.
* Fix a bug in handling the Edit Attribute in Editor Context Menu when the
edit_attrs_in_context_menu attribute is used.
* Fix a bug that may cause seg faults when reading tgif files.
* Fix a bug in Print Selected Objects that imported EPS objects are not
exported.
* For a tgif file in landscape mode, generate the correct (rotated) PDF
file.
* Update ja.po.
* Add a README file in the spice subdirectory to include some notes on the
ports and components feature of tgif.
* Add a compiler option -D_PRTGIF_USE_ISO_LATIN_1_ALWAYS to always use ISO
8859-1 character set for single-byte characters when running tgif with
the -print commandline option. Please only use this if you are sure that
you will *never* use other encoding. You can selectively turn off
reencoding by using either the Tgif.DontReencode X default or the
-dontreencode commandline option.
* 'sub handler : method {}' is now treated as a method handler
* variable $Apache::Server::ConfigTestOnly is set to 1 if Apache
is running in configuration test mode (httpd -t)
* add PerlAddVar directive, like PerlSetVar, but uses ap_table_add
instead of ap_table_set
* Apache::Table->unset can now be called with an array reference
* Apache::PerlRun::flush_namespace fixes, so aliased (imported)
code/hash/array/scalar are undefined without undef-ing the pointed-to
data and without using B.pm, thanks to Richard Chen for the suggestion
* document Apache::print's special behavior wrt references
* fix core dump triggered by <Perl>do 'foo.pl'</Perl>,
* pass $Config{ccflags} to apache for 5.6.0+
* fix Apache->httpd_conf, thanks to Will Trillich for the spot
* fixed 'make clean' to remove FILES => passed to MakeMaker
* add Apache::FILENO method for 5.6.0+ tied filehandle support
thanks to Richard Titmuss for the spot
* --disable-rule=EXPAT is passed to Apache's configure to avoid
XML::Parser conflicts
* fix PerlSetVar inheritance bug
* set r->notes("error-notes") to $@ if $@ after a Perl*Handlers is run
* added INSTALL.raven
* $c->remote_ip($ip) now also sets conn->remote_addr to make
IP-based access control work correctly
* Apache::URI::port wasn't setting the port correctly
* $Apache::Registry::NameWithVirtualHost fix take 2
(move is_virtual logic from mod_perl.c to Registry.pm)
* added Apache::OPEN method for 5.6.0+ tied filehandle support
* Apache::Table is now loaded by default (if enabled), rather than
autoloading when first needed
* $r->document_root can now be modified
* upgrade license to apache software license version 1.1
* $r->the_request can now be modified
* warn that dso+Perl malloc needs a Perl built with -Ubincompat5005
* add Apache::BINMODE stub for 5.6.0 tied filehandle support (fixes cgi.t #4)
* add $r->server->loglevel() and relevant constants.
* $Apache::Server::AddPerlVersion uses $^V for Perl/v5.6.0+
* reset the stack pointer after calling perl_require_module() in
perl_call_handler(). this fix will most likely cure the reports of
"Can't upgrade that kind of scalar at ..."
* change $Apache::VERSION check to use a string instead of SvNV, which
was troublesome with Perl 5.6.0
* Apache::src fixup to quiet warnings under Perl 5.004
* Apache::src::apxs fixups to quiet warnings if apxs doesnt exist or
httpd is not dso enabled
* fix #ifdef SGI_BOOST typo
* proxy support fixed wrt RFC2068
* set Apache::Resource::DEFAULT_RLIMIT_AS (address space) to 64Mb
* Bundle::Apache will now install Bundle::LWP before mod_perl
* support Apache::FakeRequest::args in a list context
* compile fixes for 5.6 + -Duse5005threads
* log to parms->server instead of stderr if PerlRequire fails
* make sure lib/Apache/MyConfig.pm is written before used
* fixup the output of apxs -q CFLAGS to preserve quoting (again)
* removed the generated apaci/mod_perl.config from MANIFEST
* new variable $Apache::Server::StrictPerlSections, if true, will croak
<Perl> section generates invalid Apache configuration syntax
* new Apache::src->ccflags method to combine $Config{ccflags} and
apxs -q CFLAGS
* added Apache::src->apxs method
* make use of $Config{shrpenv} to find CORE/libperl.so
* fix Apache::Registry bug when filename starts with a 0
* Apache::Registry will restore $r->status to the original value
* fixup the output of apxs -q CFLAGS to preserve quoting
* dso fix fix: modules with END blocks calling xs code (e.g. DBI) would
core dump because the xs .dso had been unloaded, now dso's are not
unloaded until after END blocks are called.
* force environ[] to be copied before any Perl code is loaded
* localize $_ in Apache::send_cgi_header
* fix precedence problem in StatINC.pm
* if hash value has magicalness in <Perl>, invoke FETCH
* sync with 1.3.12's virtual_host section logic for httpd -S + <Perl>
* remove t/httpd symlink during 'make realclean' and when rebuilding
* add Apache::src::otherldflags method
* __THE__ dso fix: dlclose() xs shared objects and call perl_shutdown()
before mod_perl's libperl.so is unloaded/reloaded by mod_so
* add /D "_MSWSOCK_" to ApacheModulePerl.dsp
* Apache::RegistryBB returns DECLINED instead of FORBIDDEN if -d $r->finfo
* add Apache::Util::validate_password()
* call (void)perl_request_rec(r) asap, instead of waiting till the
response phase, e.g. so CGI.pm can be used before PerlHandler, without
having to call Apache->request($r);
* Apache::perl_hook() fixes for Perl*Api
* save/restore $^W in Apache::PerlRun
* fix Apache::Registry when $r->uri ends with a /
* add virtualhost support to Apache::RegistryLoader
* $Apache::Registry::NameWithVirtualHost fix
* allow Apache::PerlRun::handler to be subclassed
* fix $r->connection->remote_ip for SGI performance patches (#ifdef SGI_BOOST)
* mod_perl.exp is now installed
* pod fixups to avoid WARNING from newer pod2*
* Apache::RedirectLogFix now uses $r->last
* tweaks for sfio
* update Apache::SIG docs
* fix get_set_PVp in mod_perl_xs.h
* tweaks to compile/run with 5.5.670 + -Dusethreads
* $r->hostname now documented
* Apache::File methods are now documented
* avoid quoting <Perl> config if args_how == RAW_ARGS
* PerlPassEnv maintains its value beyond the first request,
* prevent possible core dump in $r->pnotes if get_module_config returns
NULL
* fix Apache::Status::status_cv_dump for Perl/5.6-dev
* use perl_get_sv("]") instead of patchlevel for $AddPerlVersion
* Apache::Util functions $r->unescape_uri, $r->unescape_uri_info,
$r->size_string are now documented
* $r->request_time is now documented
* Look for the makepl_args.mod_perl file as .makepl_args.mod_perl in
./ and ../ too since the book says it works so.
* PerlRestartHandler is now enabled w/ ALL_HOOKS=1 (or EVERYTHING=1)
* $r->no_cache(1) will now set the r->headers_out "Pragma" and
"Cache-control" to "no-cache"
* `PerlSetEnv PERL5LIB ...' now works again
* PerlFreshRestart semantics changed to preserve order of PerlModule's
and not to reload modules while actually iterating over %INC
* $r->current_callback now works properly when PerlHandler invokes
subrequests
* if an Apache::Table is NULL return undef (e.g. $r->notes during ChildInit)
* ActivePerl compile patches
* document that Apache 1.2.x is no longer supported
* Apache::src now supported with USE_APXS=1
* disable experimental perlrunxs until future notice
* add sanity check for libgdbm
* change PL_siggv usage to gv_fetchpv("SIG", ...) for Perl 5.6-tobe
* document $r->get_handlers, $r->set_handlers, $r->push_handlers
* document PerlSetVar
* document the fact that $r->dir_config handles keys in a
case insensitive manner
* avoid logging "rwrite returned -1" multiple times in the
same request
* bugfix in write_client() when running under APACHE_SSL
* document the fact that $r->connection->user can be used to set
the authenticated user name
* document the fact that $r->uri sets/gets the URI minus the query string,
not the complete URI
* Apache::RegistryLoader::handler was rewritten to allow a better
diagnostics of scripts preloading problems.
* document the fact that $r->args can be used to set the query string
as well as get it
* add $r->connection->fileno, provides the client file descriptors. This
can be used to detect client disconnect without doing any I/O, e.g. using
IO::Select
* if $r->bytes_sent has not been calculated, then do so before returning
the value.
* ensure that all -I directories printed by Apache::src actually exist
* don't assume $uri =~ /$path_info$/ in Apache::Registry/PerlRun
* fix segfault when using Limit in <Perl> sections
* make $c->remote_host writeable
* add per-server PerlSetVar variables, accessed using
Apache->server->dir_config or $r->server->dir_config.
$r->dir_config merges them with per-directory variables.
* documentation typo
* add $r->server->uid and $r->server->gid
* add per-server configuration merge
* `_' is no longer escaped in Apache::Registry package names
* fix $ENV{PATH} corruption
* mod_perl_version.h removed from MANIFEST
* preload Apache::Constants::OPT_EXECCGI() for Apache::{Registry,PerlRun}
* Apache::SizeLimit now uses $r->child_terminate instead of Apache::exit(-2)
* get rid of various 5.005_60 warnings
* add Status{Terse,TerseSize,Deparse,Fathom,OptionsAll} options to
Apache::Status
* adjust mod_perl.h for 5.005_59 perl_eval_{pv,sv} rename
* fix flush_namespace undef logic in Apache::PerlRun
* Tweak Apache::content to comply with media-types embedded in content-types
(RFC 2616 section 3.7)
* Removed and added a few modules from the Bundle file.
* Renamed the Apache::StatINC parameters to StatINC_<something>. Changed
StatINC_Debug to take a number as parameter instead of "on" to support
"debug levels"
* fix refcount bug in %VirtualHost [] values
* $r->read will now properly check the return value of
ap_setup_client_block() and return undef if != OK,
e.g. if LimitRequestBody is exceeded
* fix bug in register_cleanup()
* added StatusLexInfo option to Apache::Status
* turn off warnings in Apache::Status::as_HTML symdump code
* tweak Apache::RegistryNG so it works with virtual includes
* change Apache::File->tmpfile flags from O_WRONLY to O_RDWR
* fix Apache::RegistryBB when $r->filename NOT_FOUND
* cygwin port
* Documented $r->pnotes
* Change "setting auth_name" trace message so we don't get it without
having enabled a MOD_PERL_TRACE level
* s/class/pclass/g so we can compile w/ c++
* mod_perl will now save/restore SIGALRM
* fix Apache::Status bug prototype check of 'use constant' subs
* added set_byterange() and each_byterange() methods
* added proper support for {Location,Directory,Files}Match in <Perl> sections
* rework PerlFreshRestart to avoid core dump triggered w/ 5.005_03
* fix $Apache::Server::Starting under dso
* use ap_bfileno() instead of r->connnection->client->fd directly
* fix possible overwrite in perlio.c:sfapacheread() #ifdef USE_SFIO
* dont use DEFAULT_PATH if r->subprocess_env->{PATH} is already set
* ensure perl_setup_env() is called just once per-request
* :common Apache::Constants are compiled at bootstrap time, so CVs are
shared and AUTOLOAD is avoided, which was known to cause a spin
while loading SERVER_ERROR for particular Perl errors
* <Perl> sections now use 1.3.7-dev's cmd_parms->context to cure many
known, old bugs, e.g. @DirectoryIndex outside of %Location, .htaccess
and many more
* fix bug where top-level PerlSetEnv's would be lost after 1st request
* fix bug in Apache::ModuleConfig->get
* fix rwrite error message
* Change Apache::SizeLimit so that it supports Solaris 2.6 and above
* Fixed apaci/load_modules.pl.PL to properly fixup log_agent and
log_referer when loaded as DSOs.
uri_unescape() did not work when given multiple strings
to decode. Patch by Nicholas Clark <nick@ccl4.org>.
ldap URIs now support _scope() and _filter() methods that
don't have default values. Suggested by Graham Barr.
Incorporated old rejected MSWin32 patch to t/old-base.t.
Hope it works.