- New compiler security options, which are apparently in use by some packagers, threw a couple of warnings. Indirectly they indicated another potential bug. All of these are now corrected. Thanks to Graham Inggs for sending me the compile logs. - NOTE FOR PACKAGERS: Don't understand me wrong, I _really_ appreciate the work you're doing! I would appreciate it if you could warn the program author of any wrongdoings. I have no way (except bug reports from third parties) to know if my program is your catalog or not!
- Debian reported a warning about a missing parameter in misc.c call to recalculate_all. For some reason gcc didn't report any warnings for me. But it _is_ an error. Corrected. - Newer versions of gtk+ cause a warning about a deprecated _unref call. Though this doesn't cause any problem, I've updated the call to the new g_object format.
- A bug in confmgr.c caused errors when reading strings from
the configuration (thanks Graham Inggs)
- A rather important bug infiltrated when changing the configuration system. Though I carefully prepared the default values for the vector map, I entered them in hex which them caused a white chart. If you had/have this problem, the easiest way out is to use this new version, but first deleting the previous configuration file (in your home directory .gnome2/linsmith) - To encourage the use of the vector chart anyway, it is now the default for new installs.
- Print bug correct (Segfault on second call to print).
(Thanks for reporting, Clyde!)
- Cleanups in the configuration load/save functions
- Made some changes in automake.am and configure.ac to
avoid confusion if importing the project into Anjuta.
- Split cable table routines from misc.c into
a separate module.
- Added configuration of line width (Z, Y, Line, Z/Y) for
This release fixes quite a few bugs:
- Change papersize_combo to combobox, and load the list from
the available papersizes.
- Adapted confmgr to accept char * from comboboxes.
- Moving down elements caused a segfault, because of wrongly
g_free'ing an iter. (Thanks clydes!)
- Moving elements up implemented.
- A crash in the configuration dialog fixed, which made it
impossible to save the SWR circle color. (Thanks clydes!)
- Converted the strange negative default values for the color
definitions to their (easier to understand?) hex values.
- The 'standard cable' code mostly re-written, and a few
more cable types added.
Removed the (deprecated) gnomeprint depency. Now the complete print route relies on Pango/Cairo to generate the plot, and the GtkPrint interface to do the actual printing.
Also, some slight improvements in configuration settings.
This is a bugfix. Saving logs of the calculations generated error messages and, in some cases, segmentation faults.
Thanks to Peter for the Swedish language update.
Anatoly Fadeev reported a problem when changing
z0, which could cause some unexpected changes.
A few minor bugs after adding s2p file import have been corrected, and I have tried to compile .TGZ (Slackware) and .RPM (Red Hat) packages for linSmith. These are very experimental, but I'd really appreciate reports on their usefulness!
By popular demand, I've added s2p (Touchstone (r)) two-port s-parameter file import. For now, only Version 1.0 files are expected to import without problems. I'll add Version 2.0 later.
I don't have a lot of s2p files to play with, but data from several manufacturers imported without problems.
I've also updated the Spanish translation, and added the s2p import instructions to the manual.
Please report problems!... read more
Dustin Vaselaar reported that the input validator for loads didn't work correctly (the New button wasn't enabled, even with correct values). Corrected.
The Spanish translation file (es.po) was updated with the new texts.
Also, the manual PDF was synced with the latest changes in the program, and some minor syntax corrections were made.
I added CSV data input for the loads, as a first step. s2p format is on the way.
Solved a few erratic problems with colors disappearing on the screen (which in fact, converted into transparent).
Please consult the Changelog for more details.
Hans Nieuwenhuis reported an error in log.c which could cause crashes when generating a data log. Thanks!
A couple of other potential char array traps removed.
These are of less importance as the reserved space
was more than enough to accomodate the strings.
A bugfix release that avoids crashes with longer fontnames.
The main feature in this release is the replacement of the bitmap
background of the on-screen chart by a vector map. This vastly improves
the quality of the chart when zooming.
It is also possible to define the font and font size of the monospaced
characters below the chart and on the Results page. The original, default
values, were a little 'light'.
Thanks to Anatoliy Fadeev, a new version of the Russian translation is available.
To update your copy, download the ru-0.9.8.po file to the po/ directory in the source tree, and recompile. (please rename the file to ru.po!)
Peter Landgren sent me the latest update for the Swedish translation (thanks!).
To update your copy, copy the sv-0.9.8.po file to the po/ directory in the source tree, and recompile. (please rename the file to sv.po!)
Mainly minor improvements:
- Version in component files is now the complete
version in a single string. This should not
raise any incompatibilities.
- Updated es.po
- Modified the Results page: Circuit impedance
is now shown as either Z or Y depending on the
Z/Y blocks inserted in the element list. This
can be disabled (forced to impedance only) in
Preferences. (This also applies to Results
- Separate resolutions settable for impedances
Corrected the 'live' display of the cursor
values under the graph - they remained at
50 Ohm even after changing z0 (Thanks Anatoly!
To help students at the university, I added a mode
to input data directly from noise bridge measurements.
Next version will focus on the remote input (really!)
- Corrected the connecting arcs for 'to load'
- Modified the load impedance input to select
different input data sources and types.
- Don't enable the 'New' or 'Update' buttons if
the values are not valid (f != 0).
- Noise bridge input works! Input in R and C,
values are shown as R and X in the load table.
While preparing for the remote control part, I detected a bug which I considered important enough to merit a release. All changes:
- Added a 'Clear' button to the load and element lists.
- Replaced large 'Recalculate' button by a smaller version,
and moved it to the left, to make space for the rotation list.
- Added the Rotate list, to select rotation to generator or
to load. Added the option to the preferences record. Note
that though transmission lines already rotate correctly,
the connecting line is still wrong.
- Setting the Zo box on startup caused errors in the (not
yet existing) log_list. Solved. - Minor bugfix in configuration
window. Screen font name was not correctly saved and could
cause execution problems.
- After configuration, chart must be in non-logo mode, else an
error message is produced when trying to update the SWR circle.
While starting to look for the problems with the remote interface, I found a few other issues with the 'loads' part, which merited attention:
- Chart also updates when loads are added ot
- Infrequent bug detected in load management. Could
be the problem while importing from remote?
- Redefined (improved) quite a few routines in
the load.c file. This caused a few bugs (which I
hope, are resolved). Please report on any problems.