Re: Binary File Data and Associated Tools

Hi Akarsh,

I grabbed the tarball and tested things out with binfiletester.  I got
an error message in trixel 1:
-------------------------------------
./binfiletester namedstars.bin 1 names.bin KStars Star Data v1.0. To
be read using the 32-bit starData structure only.
Data Field:
  Name: RA
  Size: 4
  Type: 5
  Scale: 1000000

Data Field:
  Name: Dec
  Size: 4
  Type: 5
  Scale: 100000

Data Field:
  Name: dRA
  Size: 4
  Type: 5
  Scale: 10

Data Field:
  Name: dDec
  Size: 4
  Type: 5
  Scale: 10

QA for 4.1

Hello,

I have posted our QA list for KStars in preparation for the 4.1 release:
http://techbase.kde.org/KStars_QA_List_4.1

Basically, this is an exhaustive list of every behavior that KStars is
supposed to do.  The idea is to go through the list and test each one,
and then color-code each entry according to the key at the top of the
page.  There are codes for "it works", "it's broken", "it sort-of
works, but has problems", "it causes a crash", "there's a usability
issue", and "not yet tested".

If you want to help polish KStars-1.4 for the KDE-4.1 release, please
visit that page and start putting kstars through its paces!

Note: this list was adopted from our preparations for 4.0, and hasn't
been substantively updated yet (e.g., there's no mention of the
conjunctions tool yet).  Feel free to edit as necessary.  Also, I went
ahead and reverted all previously green ("works") items back to black
("untested").  I left all non-green colors in place; so for those
please test the behavior and if the problem has gone away, change it
to green.

thanks!

Jason

Re: Binary File Data and Associated Tools

Hi Jason

It depends on which tarball you picked up. The new files don't support
names in the way they did earlier, so you can't query names with the
binfiletester. The reason for this is that we no longer store
'named_uid' fields, and the bright / named star file is designed to be
read sequentially.
Or, maybe the tarball was corrupt?

The trouble is that I haven't spent any time polishing the
binfiletester app. The newer version doesn't read names, but it
doesn't reflect that change in the usage statement as yet. I'll fix
these important things and commit to the branch today.

Regards
Akarsh

_______________________________________________
Kstars-devel mailing list
Kstars-devel <at> kde.org
https://mail.kde.org/mailman/listinfo/kstars-devel

KDE/kdeedu/kstars/kstars

SVN commit 815443 by harris:

We have 6 "object action" keystrokes that operate on the clickedObject: 
C = center the object
D = Show the object's Details window
P = Show the object's popup menu
O = Add the object to the observing list
L = Add/remove the object's name label
T = Add/remove a trail from the object (if isSolarSystem()==true)

If the Shift key is held, then these actions operate on the *centered 
object (i.e., focusObject) instead.  This is accomplished by setting the 
clicked object to be the focus object (setClickedObject( focusObject() 
)).  With this commit, the re-assignment of clickedObject when Shift is 
used with an action key is now temporary; it now reverts back to 
whatever it was before the action was invoked.

CCMAIL: kstars-devel <at> kde.org

 M  +3 -0      imageviewer.cpp  
 M  +2 -2      kstars.cpp  
 M  +60 -9     skymapevents.cpp  

--- trunk/KDE/kdeedu/kstars/kstars/imageviewer.cpp #815442:815443
@@ -113,6 +113,9 @@
     // check if download job is running
     checkJob();

+    //DEBUG
+    kDebug() << View << "::" << Caption << endl;

KDE/kdeedu/kstars/kstars

SVN commit 815445 by harris:

Reverting debug statements that should not have been part of the 
previous commit...

CCMAIL: kstars-devel <at> kde.org

 M  +0 -3      imageviewer.cpp  
 M  +0 -3      kstars.cpp  

--- trunk/KDE/kdeedu/kstars/kstars/imageviewer.cpp #815444:815445
@@ -113,9 +113,6 @@
     // check if download job is running
     checkJob();

-    //DEBUG
-    kDebug() << View << "::" << Caption << endl;
-
     delete View;
     delete Caption;
     if ( downloadJob ) delete downloadJob;
--- trunk/KDE/kdeedu/kstars/kstars/kstars.cpp #815444:815445
@@ -249,9 +249,6 @@

 void KStars::removeImageViewer( ImageViewer *iv ) {
     int i = m_ImageViewerList.indexOf( iv );
-    //DEBUG
-    kDebug() << i << "::" << m_ImageViewerList.size() << endl;
-
     if ( i != -1 )

Re: KDE/kdeedu/kstars/kstars

> If the Shift key is held, then these actions operate on the *centered 
> object (i.e., focusObject) instead.  This is accomplished by setting the 
> clicked object to be the focus object (setClickedObject( focusObject() 
> )).  With this commit, the re-assignment of clickedObject when Shift is 
> used with an action key is now temporary; it now reverts back to 
> whatever it was before the action was invoked.

Does this affect any feature or was this a code quality fix?

Regards
Akarsh

_______________________________________________
Kstars-devel mailing list
Kstars-devel <at> kde.org
https://mail.kde.org/mailman/listinfo/kstars-devel

crash condition, possibly due to KIO bug

Hello,

Just a heads up: I have uncovered a crash condition in KStars.  If you
invoke the Image Viewer using an object's right-click menu, and the
image cannot be downloaded (due to an incorrect URL or unresponsive
server), then KStars will crash.

After some investigation, I believe the bug is in KIO, not KStars.  I
have filed a bug report here:
http://bugs.kde.org/show_bug.cgi?id=163023

If anyone has ideas of how we can avoid exposing this crash, let me know!

thanks,
Jason

branches/kstars/summer/kdeedu/kstars/kstars/data

SVN commit 815460 by asimha:

+ Moving tools used to analyze data into kstars/data/tools from kstars/data

+ Adding tools used to handle binary data

+ Adding a README that describes some of the data tools

CCMAIL: kstars-devel <at> kde.org

 D             check_url.pl  
 D             cull-cbounds.pl  
 D             mag-linenum.pl  
 D             max-length.pl  
 D             read-hippo.pl  
 D             reformat-hip.pl  
 D             sort-hip-by-pm.pl  
 A             tools (directory)  
 A             tools/Makefile  
 A             tools/README.tools  
 A             tools/binfiletester.c   [License: UNKNOWN]
 AM            tools/check_url.pl  
 AM            tools/cull-cbounds.pl  
 AM            tools/datatomysql.pl  
 AM            tools/mag-linenum.pl  
 AM            tools/max-length.pl  
 A             tools/mysql2bin.c   [License: UNKNOWN]
 AM            tools/read-hippo.pl  
 AM            tools/reformat-hip.pl  
 AM            tools/sort-hip-by-pm.pl  

branches/kstars/summer/kdeedu/kstars/kstars/data/tools

SVN commit 815487 by asimha:

+ Adding new makefile target to generate the data files
+ Updating documentation
+ Making minor changes to mysql2bin.c

CCMAIL: kstars-devel <at> kde.org

 M  +10 -1     Makefile  
 M  +55 -0     README.tools  
 M  +7 -4      mysql2bin.c  

--- branches/kstars/summer/kdeedu/kstars/kstars/data/tools/Makefile #815486:815487
@@ -9,6 +9,9 @@
 #

 MYSQL_CONFIG=/usr/bin/mysql_config
+KSTARS_MYSQL_DB_TO_BIN=./mysql2bin
+KSTARS_MYSQL_DB_DB=$(KSTARS_MYSQL_DB_USER)
+KSTARS_MYSQL_DB_TBL=allstars

 all: mysql2bin binfiletester

@@ -20,6 +23,12 @@

 clean:
 	-rm binfiletester mysql2bin
-
+datafiles: mysql2bin
+	echo "If this step hangs, please reduce the value of MYSQL_STARS_PER_QUERY in mysql2bin.c and try again."

Re: KDE/kdeedu/kstars/kstars

the old code could lead to unexpected behavior in some cases.  It's
not a major fix by any means.  If you're wondering if you should port
it to the branch, I would say don't bother (actually, I wouldn't
bother trying to keep the branch synched to trunk at all; we can worry
about merging at the end).

regards,
Jason

On Sun, Jun 1, 2008 at 1:31 PM, Akarsh Simha <akarshsimha <at> gmail.com> wrote:
>> If the Shift key is held, then these actions operate on the *centered
>> object (i.e., focusObject) instead.  This is accomplished by setting the
>> clicked object to be the focus object (setClickedObject( focusObject()
>> )).  With this commit, the re-assignment of clickedObject when Shift is
>> used with an action key is now temporary; it now reverts back to
>> whatever it was before the action was invoked.
>
> Does this affect any feature or was this a code quality fix?
>
> Regards
> Akarsh
>
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1.4.6 (GNU/Linux)
>
> iD8DBQFIQwcpkAo+aFFLL8MRAqb4AKCYpbmXcjxGW7+1mzeHGTeLR1fP2gCaA8Bq
> ea+dzOQJ7NB/wMCLAwZXDSw=
> =rDds
> -----END PGP SIGNATURE-----
>