civlog-list — List about general CivLog topics

[Civlog-list] Freeciv log patch idea

[Wong TM (H. Deming) <loo...@ma...>]

I see that the log file can get very big.

Idea:

- stop outputing dead player log and maybe add a 'dead' tag to signify
end of dead player log.

D  P YEAR DATA
-  - ---- ----
13 1 1100 100
13 2 1100 120
13 3 1100 dead <- player dead
13 4 1100 144
[...]
13 1 1120 110
13 2 1120 111 <-\
13 4 1120 111 <- dead player skiped

- Race info read along with player name.

I will do it in 0.6.2

 -Deming

--
L I N U X       .~.
The  Choice     /V\
of a  GNU      /( )\
Generation     ^^-^^

[Civlog-list] RE: Screenshots

[Wong TM (H. Deming) <loo...@ma...>]

> -----Original Message-----
> From: Raimar Falke [mailto:hawk@A315-2b.WH8.TU-Dresden.De]
>
> The announcement of CivLog was today in my freshmeat box. *nice*
> However I think the screenshots look ugly. You should IMHO increase
> the jpeg quality or use gif or png.

Done. I hope you mean the bigger screenshoot is ugly.

btw. Any free webhosting site you will recommend as the current asia1
doesn't support some advance ftp command(it tedious to update the site
manually).

  Deming

--
Hi! I'm a .signature virus! Copy me into your ~/.signature, please!

Re: [Civlog-list] Score.java v3

[Ulrich K. <ku...@un...>]

On Fri, 22 Sep 2000, Raimar Falke wrote:

> On Thu, Sep 21, 2000 at 11:16:03PM +0200, Ulrich Kuhn wrote:
> > Hi folks,
> > 
> > today I have taken a closer look at the CivLog code and I stumbled across
> > Score.java. It met my requirements: I understood it, I knew where to
> > improve it, it was somewhat separated from the rest.
> > So I wrote a in my opinion improved version of it.
> > 
> > I 've attached the three necessary files in a zip file. PlayerInfo.java
> > and Searcher.java are new files.
> > 
> > Please send comments.
> 
> It would be nice if you had sent a short diescription how the new score
> works and what kind of improvement you made. I took a look at the code 
> and have to following comments:
>  - can you go over the method declaritions and set the permission (private,
>    protected or public)? Some methods in PlayerInfo for example should be
>    private or public but not package

I think most of the method declarations are useful:
PlayerInfo is ment to be a helper and data storage for Score (methods
can't be private) but PlayerInfo is not intended to be given away by
Score (so no public methods).
But I need an value access/exchange object for Score that can other
objects (ie Graph) request. So maybe I'll change this...

>  - nice to have test methods in the main, but could you please move them to
>    the end? If I just want the pblic methods of the class these mains are 
>    in the way.

Yes.

>  - why do you use arrays in YearValueMap instead of a hash? Have you made
>    some measurements?

Not needed:
http://www.javaworld.com/javaworld/jw-11-1999/jw-11-performance-2.html

In short: (java) objects needs (besides the raw values) some memory to do
things like synchronization. This extra memory varies it's between 10-40
bytes per object. So the object (class IntW { int i; }) will not use 4
bytes but perhaps 20.
The idea to store the values in a hash was not bad but for pure value
storage objects are using way to much memory. BTW: arrays are objects too
but an array containing many values is only one object...

>  - the whole score thing should be using turns instead of years as primary
>    reference. At the beginning everything used years than the GUI was
>    changed to use turns and turn2year and year2turn was added. It
>    would now be time to change score to also use turns.

Yes.

>  - it would be nice to have an mapping (data type)->(boolean) which would
>    tell you which data types occur in the score file. Also a lastDataType
>    would be nice (can be used instead of 
>    FreecivConstants.NUMBER_OF_DATA_TYPES to declare arrays)

Question: Is the number and type of data types in a score file
arbitrary? Especially can there be some types missing (0,1,4,8,11,...)?

>  - in 
>     PlayerInfo(String name)
>     {
>         playerName = name;
>         typeList = new IndexedList(28);
>     }
>    Can you replace the 28 with a more meaningful number?

IndexedList is able to grow if needed. So 28 is just an initial value (I
think there are 27 or 28 data types but maybe there are (later) more.

>  - you can also replace the public variables with access methods in Score

With pleasure!

>  - the changes you made to loadScoreFile are nice

The reason here is similar to the memory usage thing: StringTokenizer is a
nice thing (if it wouldn't exist I had to programm it myself!) but it
returns new String objects instead of using String.substring which is much
faster.
Ooops false! I just looked in the source code of StringTokenizer and it
DOES use substring....


>  - the semantics of findEntry are wrong. findEntry in the current version
>    returns null if for this year no entry is found. We can get rid of this
>    kludge if we add an isValidYear() or maybe this also goes away if we
>    change everything to use turns.

Is fixed but I'll remove this method anyway because it uses ScoreEntry.

>  - also in your version any changes which are made to an ScoreEntry returned
>    by getEntry are discarded. AFAIK in the current code only Score changes
>    an ScoreEntry but is a difference.
>  - since the GUI got much smarter nobody uses maxValues anymore. This stuff
>    can be removed. Same with getMaxMinValue.

Ok.

>  - AFAIK this line from Score is a bit hard to understand correct: 
>      boolean notprecise = false;
>    IMHO we should avoid double negation.

ROFL! You're right! I just thought the variable name should have the same
meaning as it's value...

>  - lets try to describe the data structure:
>     * Score has PlayerInfoS which are indexed by player number
>     * PlayerInfo (holds all data about one player) has IndexedList
>     * IndexedList (holds all data about one player) has YearValueMapS which
>       are indexed by data type
>     * YearValueMap has pairs of (year, value) which are indexed by year

Seems to be correct but maybe this is easier to understand:
* Score contains PlayerInfoS indexed by player number.
* PlayerInfo contains a IndexedList indexed by data type.
* Each entry in this IndexedList is a YearValueMap (one for every data
  type) that contains all values for one data type (and player).
Well, not quite easier but at least another formulation. :)

>    Current structures:
>     * Score has pairs (EntryKey,ScoreEntry) which are indexed by EntryKey
>     * EntryKey describes all data of one player at one year
>     * ScoreEntry (holds all data of one player at one year) has pairs of
>       (type, value) which are indexed by type
>    So you don't use EntryKey. You added some layers to the prior flat
>    structure.

Yes one layer more.


I'm now diving into the code looking for some answers. :)

See you (<blubb>)

Ulrich

[Civlog-list] CivLog on fresh meat

[Wong TM (H. Deming) <loo...@ma...>]

I just submit CivLog to fresh meat. Some red tape in process.....
       __
.----.|__|.---.-..-----.
|  __||  ||  _  ||  _  |
|____||__||___._||_____|
 

RE: [Civlog-list] Score.java v3

[Wong TM (H. Deming) <loo...@ma...>]

> -----Original Message-----
> From: civ...@li...
> [mailto:civ...@li...]On Behalf Of Raimar
> Falke
>
>
> On Fri, Sep 22, 2000 at 01:34:27PM +0800, Wong TM (Huang Deming) wrote:
> > > -----Original Message-----
> > > From: civ...@li...
> > > [mailto:civ...@li...]On Behalf Of
> > > Ulrich Kuhn
> > >
> > > Hi folks,
> >
> > Howdy!
>
> I thought this is a greeting only used in Texas?!

Is there any ethical question in using 'Howdy'? I see 'Hi folk' as a
cowboy style of greeting and therefore 'Howdy'. Answer needed :-|.

> > I didn't go through all the code yet.
> > Here are some of my findings:
> >
> > - CivLog save previous settings like the players selected. If the old
> > setting saved that the 9th player is selected and I load a
> score log with
> > less than 8 players later. The MainFrame try to load the 9th
> player but
> > fail. Attached the debug log in 'cvslog'.
>
> This code in View should take care of it:
>             if(tmp<numOfPlayers)
>                 setPlayerSelected(tmp,true);
> And there is an error some lines later:
>             if(tmp<numOfPlayers)
>                 setDataTypeSelected(tmp,true);
> This should be something like:
>             if(score.containsDataType(tmp))
>                 setDataTypeSelected(tmp,true);

OK

> > The project had slow down a lot because the flame is smaller, the
> > direction to go is blur, and no time :).
>
> AFAIK this depends on the motivation. If there is no personal
> itch which you
> can scratch there is only low motivation. IMHO the project is in a nice
> condition. There are almost no personal itches left. For

Agreed. CivLog is now in a state I want when I first started thinking of
doing it. Now it is a stable platform for others to extend their
ambitious itch. I still see a lot of potentials in CivLog. I hope that it
isn't damper by Java(Java language is great but...).

> example I would
> like to try a second time to code some client side AI. This

Great! I look forward to it. AI is never easy. But it can satisfied a big
itch :)

> would get some
> motivation. Maybe we should go really public (announce at freshmeat and
> others) and wait for the acknowledgements and requests of some people.

Interesting. I will visit their website and see how.

> > Ulrich had brighten the flame. So which direction/goals should we
> > go/achieve?
>
> There are also some items on the TODO list left.

True :)

Ulrich, I welcome you to our prestigious CivLog club now that you pay the
membership fee(source code). :-))

So Ulrich, where/what is your itch(plan)?

 -Deming

Re: [Civlog-list] Score.java v3

[Raimar F. <hawk@A315-2b.WH8.TU-Dresden.De>]

On Fri, Sep 22, 2000 at 01:34:27PM +0800, Wong TM (Huang Deming) wrote:
> > -----Original Message-----
> > From: civ...@li...
> > [mailto:civ...@li...]On Behalf Of
> > Ulrich Kuhn
> >
> > Hi folks,
> 
> Howdy!

I thought this is a greeting only used in Texas?!
> I didn't go through all the code yet.
> Here are some of my findings:
> 
> - CivLog save previous settings like the players selected. If the old
> setting saved that the 9th player is selected and I load a score log with
> less than 8 players later. The MainFrame try to load the 9th player but
> fail. Attached the debug log in 'cvslog'.

This code in View should take care of it:
            if(tmp<numOfPlayers)
                setPlayerSelected(tmp,true);
And there is an error some lines later:
            if(tmp<numOfPlayers)
                setDataTypeSelected(tmp,true);
This should be something like:
            if(score.containsDataType(tmp))
                setDataTypeSelected(tmp,true);

> The project had slow down a lot because the flame is smaller, the
> direction to go is blur, and no time :).

AFAIK this depends on the motivation. If there is no personal itch which you
can scratch there is only low motivation. IMHO the project is in a nice
condition. There are almost no personal itches left. For example I would
like to try a second time to code some client side AI. This would get some
motivation. Maybe we should go really public (announce at freshmeat and
others) and wait for the acknowledgements and requests of some people.

> Ulrich had brighten the flame. So which direction/goals should we
> go/achieve?

There are also some items on the TODO list left.

	Raimar

-- 
 email: rf...@in...       phone: +49 351 4415773     icq: 54420251
 pgp 2: id: 0F9D7955 len: 1024 fingerprint: 7760F933D5478009 4FA0C56F1DC2FB8E
 "When C++ is your hammer, everything looks like a thumb."
    -- Steven M. Haflich

Re: [Civlog-list] Score.java v3

[Raimar F. <hawk@A315-2b.WH8.TU-Dresden.De>]

On Thu, Sep 21, 2000 at 11:16:03PM +0200, Ulrich Kuhn wrote:
> Hi folks,
> 
> today I have taken a closer look at the CivLog code and I stumbled across
> Score.java. It met my requirements: I understood it, I knew where to
> improve it, it was somewhat separated from the rest.
> So I wrote a in my opinion improved version of it.
> 
> I 've attached the three necessary files in a zip file. PlayerInfo.java
> and Searcher.java are new files.
> 
> Please send comments.

It would be nice if you had sent a short diescription how the new score
works and what kind of improvement you made. I took a look at the code 
and have to following comments:
 - can you go over the method declaritions and set the permission (private,
   protected or public)? Some methods in PlayerInfo for example should be
   private or public but not package
 - nice to have test methods in the main, but could you please move them to
   the end? If I just want the pblic methods of the class these mains are 
   in the way.
 - why do you use arrays in YearValueMap instead of a hash? Have you made
   some measurements?
 - the whole score thing should be using turns instead of years as primary
   reference. At the beginning everything used years than the GUI was
   changed to use turns and turn2year and year2turn was added. It
   would now be time to change score to also use turns.
 - Search is really an BinarySearchForIntArrays (don't change the name just
   add it in the header)
 - it would be nice to have an mapping (data type)->(boolean) which would
   tell you which data types occur in the score file. Also a lastDataType
   would be nice (can be used instead of 
   FreecivConstants.NUMBER_OF_DATA_TYPES to declare arrays)
 - in 
    PlayerInfo(String name)
    {
        playerName = name;
        typeList = new IndexedList(28);
    }
   Can you replace the 28 with a more meaningful number
 - you can also replace the public variables with access methods in Score
 - the changes you made to loadScoreFile are nice
 - the semantics of findEntry are wrong. findEntry in the current version
   returns null if for this year no entry is found. We can get rid of this
   kludge if we add an isValidYear() or maybe this also goes away if we
   change everything to use turns.
 - also in your version any changes which are made to an ScoreEntry returned
   by getEntry are discarded. AFAIK in the current code only Score changes
   an ScoreEntry but is a difference.
 - since the GUI got much smarter nobody uses maxValues anymore. This stuff
   can be removed. Same with getMaxMinValue.
 - AFAIK this line from Score is a bit hard to understand correct: 
     boolean notprecise = false;
   IMHO we should avoid double negation.
 - lets try to describe the data structure:
    * Score has PlayerInfoS which are indexed by player number
    * PlayerInfo (holds all data about one player) has IndexedList
    * IndexedList (holds all data about one player) has YearValueMapS which
      are indexed by data type
    * YearValueMap has pairs of (year, value) which are indexed by year
   Current structures:
    * Score has pairs (EntryKey,ScoreEntry) which are indexed by EntryKey
    * EntryKey describes all data of one player at one year
    * ScoreEntry (holds all data of one player at one year) has pairs of
      (type, value) which are indexed by type
   So you don't use EntryKey. You added some layers to the prior flat
   structure.

> <put on asbestos suite>

It is nice. No need to put on an asbestos suite. Some changes here and there 
and also some improvements which would change the interface and we are
ready.

	Raimar

-- 
  Tank: So what do you need? Besides a miracle.
  Neo: Guns. Lots of guns.
    -- From The Matrix

RE: [Civlog-list] Score.java v3

[Wong TM (H. Deming) <loo...@ma...>]

> -----Original Message-----
> From: civ...@li...
> [mailto:civ...@li...]On Behalf Of
> Ulrich Kuhn
>
> Hi folks,

Howdy!

> today I have taken a closer look at the CivLog code and I
> stumbled across
> Score.java. It met my requirements: I understood it, I knew where to
> improve it, it was somewhat separated from the rest.
> So I wrote a in my opinion improved version of it.

Great work!

> I 've attached the three necessary files in a zip file. PlayerInfo.java
> and Searcher.java are new files.
>
> Please send comments.

I didn't go through all the code yet.
Here are some of my findings:

- CivLog save previous settings like the players selected. If the old
setting saved that the 9th player is selected and I load a score log with
less than 8 players later. The MainFrame try to load the 9th player but
fail. Attached the debug log in 'cvslog'.

- I stress test it by loading a 2MB log file(20 players). The resulting
debug log 'superlog' is attached. I don't know where the problem yet. The
2MB score log file can be found on CivLog Website(Download|Sample Log).

Sorry about complaining rather than fixing it ;-)

> <put on asbestos suite>
>
> Yours
> Ulrich Kuhn
>
> P.S.: Good work guys.
Yeah!

The project had slow down a lot because the flame is smaller, the
direction to go is blur, and no time :).

Ulrich had brighten the flame. So which direction/goals should we
go/achieve?

  -Deming

[Civlog-list] Score.java v3

[Ulrich K. <ku...@un...>]

Hi folks,

today I have taken a closer look at the CivLog code and I stumbled across
Score.java. It met my requirements: I understood it, I knew where to
improve it, it was somewhat separated from the rest.
So I wrote a in my opinion improved version of it.

I 've attached the three necessary files in a zip file. PlayerInfo.java
and Searcher.java are new files.

Please send comments.

<put on asbestos suite>

Yours
Ulrich Kuhn

P.S.: Good work guys.

[Civlog-list] RE: freeciv, CivLog

[Wong TM (H. Deming) <loo...@ma...>]

> -----Original Message-----
> From: mw...@mx... [mailto:mw...@mx...]On Behalf
        ^^^^^^^^^^^^^^^^^^^^^^^ Strange?
> Of Martin Willemoes Hansen
> Sent: Tuesday, September 19, 2000 11:33 PM
> To: loo...@ma...
> Subject: freeciv, CivLog
>
>
> Hi!

Hi!
>
> Moved your contribution CivLog-0.4 it is available from the download
> section, also I moved your readme file to something like this:
> CivLog-0.4.README, this is the prefered name syntax, please use it in
> the future :)

Great! Thank you, Martin.

I just finish CivLog-0.6(stable) with the help of Raimar Falke
<rf...@in...> and Michael Grundel <mic...@gr...>.

I uploaded two files into freeciv.org/incoming:

* civlog-0.6.tar.gz - contain just the source and samples file. unix file
attribute preserved. Good for linux users.
* civlog-0.6.zip - contain the above files plus precompiled class, for
people without JDK.

I hope that win9x didn't change CivLog-0.6.README to CivLog-0.6.Readme.

> Hope to see many new and great contributions from you!
:-)) Thanks to Raimar and Michael.


 -Deming

[Civlog-list] FW: freeciv, CivLog

[Wong TM (H. Deming) <loo...@ma...>]

Great work guys!

-----Original Message-----
From: mw...@mx... [mailto:mw...@mx...]On Behalf
Of Martin Willemoes Hansen
Sent: Tuesday, September 19, 2000 11:33 PM
To: loo...@ma...
Subject: freeciv, CivLog


Hi!

Moved your contribution CivLog-0.4 it is available from the download
section, also I moved your readme file to something like this:
CivLog-0.4.README, this is the prefered name syntax, please use it in
the future :)

Hope to see many new and great contributions from you!
 
Martin Willemoes Hansen

--------------------------------------------------------
E-Mail	mw...@st...	Website	mwh.stampede.org
IRC     MWH, openprojects	
--------------------------------------------------------

RE: [Civlog-list] CivLog Tarball

[Wong TM (H. Deming) <loo...@ma...>]

> -----Original Message-----
> From: Raimar Falke
>
> On Tue, Sep 19, 2000 at 05:08:27PM +0200, Ulrich Kuhn wrote:
> > On Tue, 19 Sep 2000, Wong TM (Huang Deming) wrote:
> >
> > >
> > > Seeing that the zip file didn't do justice to linux, I
> make a tarball of
> > > CivLog for Linux users.
>
> Since we are speaking about java AFAIK you can assume everybody has jar
> installed.

The main for the tar ball is that it retain the chmod attributes? And I
think tar ball is more popular among linux user.( they can use MC or
double click it) No need for command line for the lazy like me :-)

> > > * One new make target: 'archive' use to make a tarball
> > >
> > > archive: clean
> > > 	tar -C ./../ -cvf civlog.tar ./CivLog/.
> > > 	gzip -9 civlog.tar
> > > Is the tar command right?
> > > There seen to a extra '.' in the archive(CivLog/./org/ etc)
> > >                                                 ^
> > > created.
> >
> > Maybe little bit shorter (and without additional dot):
> > 	tar -cvz -f civlog.tgz -C .. CivLog
>
> You can remove the first two '-' to make it even shorter:
> (don't know about
> -C. -C can be replaced by "(cd ..;tar cfvf CivLog/civlog.tgz CivLog)")
>  tar cfvz civlog.tgz -C .. CivLog
> will do the job. IMHO this should be split since AFAIK tar
> uses just gzip
> and the default for gzip is -6. So maybe we can change this to
>   tar cfv civlog.tar -C .. CivLog
>   gzip -9 civlog.tar
> Since we speak about linux people which also have enough space for
> extentions we now get also ".tar.gz" instead of ".tgz".

Nod.

> > Alternativly you could have given the hint to do a
> > jar xvf civlog.zip
> > with the zip file. That's what I regularly do to uncompress zip files
> > under linux.
>
> Also for a lot of big images (about 30MB) jar is faster then
> winzip for just
> putting them together (no compression). I noticed this when I
> had to move a
> bunch (~30) of these over ssh.

After seeing that javac is slower than jikes and java application is
slow(endless tread) and a mem hog(84MB for CivLog on linux), didn't
expect people to say good thing about jar ;-).

btw. I read that 'javac -O' produce more optimized byte code than jikes.

Anyone try GNU gcj to compile java before?

  Deming

Re: [Civlog-list] CivLog Tarball

[Raimar F. <hawk@A315-2b.WH8.TU-Dresden.De>]

On Tue, Sep 19, 2000 at 05:08:27PM +0200, Ulrich Kuhn wrote:
> On Tue, 19 Sep 2000, Wong TM (Huang Deming) wrote:
> 
> > 
> > Seeing that the zip file didn't do justice to linux, I make a tarball of
> > CivLog for Linux users.

Since we are speaking about java AFAIK you can assume everybody has jar
installed.

> > Minor Changes:
> > 
> > * Add Raimar comment on log files added into the Readme.
> > 
> > * Michael patch applied.
> > 
> > * One new make target: 'archive' use to make a tarball
> > 
> > archive: clean
> > 	tar -C ./../ -cvf civlog.tar ./CivLog/.
> > 	gzip -9 civlog.tar
> > Is the tar command right?
> > There seen to a extra '.' in the archive(CivLog/./org/ etc)
> >                                                 ^
> > created.
> 
> Maybe little bit shorter (and without additional dot):
> 	tar -cvz -f civlog.tgz -C .. CivLog

You can remove the first two '-' to make it even shorter: (don't know about
-C. -C can be replaced by "(cd ..;tar cfvf CivLog/civlog.tgz CivLog)")
 tar cfvz civlog.tgz -C .. CivLog
will do the job. IMHO this should be split since AFAIK tar uses just gzip
and the default for gzip is -6. So maybe we can change this to
  tar cfv civlog.tar -C .. CivLog
  gzip -9 civlog.tar
Since we speak about linux people which also have enough space for
extentions we now get also ".tar.gz" instead of ".tgz".

> Alternativly you could have given the hint to do a
> jar xvf civlog.zip
> with the zip file. That's what I regularly do to uncompress zip files
> under linux.

Also for a lot of big images (about 30MB) jar is faster then winzip for just
putting them together (no compression). I noticed this when I had to move a
bunch (~30) of these over ssh.

	Raimar
-- 
 "Python 2.0 beta 1 is now available [...]. There is a long list of new 
  features since Python 1.6, released earlier today. We don't plan on 
  any new releases in the next 24 hours."
    -- Jeremy Hylton at Slashdot

Re: [Civlog-list] CivLog Tarball

[Ulrich K. <ku...@un...>]

On Tue, 19 Sep 2000, Wong TM (Huang Deming) wrote:

> 
> Seeing that the zip file didn't do justice to linux, I make a tarball of
> CivLog for Linux users.
> 
> Minor Changes:
> 
> * Add Raimar comment on log files added into the Readme.
> 
> * Michael patch applied.
> 
> * One new make target: 'archive' use to make a tarball
> 
> archive: clean
> 	tar -C ./../ -cvf civlog.tar ./CivLog/.
> 	gzip -9 civlog.tar
> Is the tar command right?
> There seen to a extra '.' in the archive(CivLog/./org/ etc)
>                                                 ^
> created.

Maybe little bit shorter (and without additional dot):
	tar -cvz -f civlog.tgz -C .. CivLog
This will create the gzip'd tar ball civlog.tgz in the actual directory.

Alternativly you could have given the hint to do a
jar xvf civlog.zip
with the zip file. That's what I regularly do to uncompress zip files
under linux.

Ulrich

[Civlog-list] CivLog Tarball

[Wong TM (H. Deming) <loo...@ma...>]

Seeing that the zip file didn't do justice to linux, I make a tarball of
CivLog for Linux users.

Minor Changes:

* Add Raimar comment on log files added into the Readme.

* Michael patch applied.

* One new make target: 'archive' use to make a tarball

archive: clean
	tar -C ./../ -cvf civlog.tar ./CivLog/.
	gzip -9 civlog.tar
Is the tar command right?
There seen to a extra '.' in the archive(CivLog/./org/ etc)
                                                ^
created.

It should be in CivLog website now.

       __
.----.|__|.---.-..-----.
|  __||  ||  _  ||  _  |
|____||__||___._||_____|

RE: [Civlog-list] Font problem in Linux

[Wong TM (H. Deming) <loo...@ma...>]

> -----Original Message-----
> From: civ...@li...
> [mailto:civ...@li...]On Behalf Of Raimar
> Falke
>
> On Tue, Sep 19, 2000 at 01:40:47PM +0800, Wong TM (Huang Deming) wrote:
> >
> > When starting CivLog in linux, there some minor error
> message about the
> > font not found. I notice that CivLog use a non-monotype font
> when display
> > the help|Content. Did any of you guys had similar thing happening?
>
> A lot of messages like
> "Font specified in font.properties not found
> [-*-symbol-medium-r-normal--*-%d-*-*-p-*-adobe-fontspecific]"
> when the VM starts? As long as I can remember using Java under
> Linux these
> messages were always there. Never searched and solved this.

'Font specified in font.properties not found [--zapf
dingbats-medium-r-normal--*-%d-*-*-p-*-adobe-fontspecific]' for mine
case.

I specified Courier New to be used for the text in Help|Content textbox.
Java display the text using a non-monotyped font.

Below are one email I just found,

-----------------------------------------------
Scot Bellamy [sbe...@ac...] wrote:

I found a way to at least eliminate the annoying font errors when
starting
Jext.  I use JBuilder3.5 Foundation and it had the same problem.  Borland
had
some information on their web site about it.

Here is what I did.

I found a copy of the symbol font, symbol.ttf and copied it to the
jre/lib/fonts subdirectory of my JDK installation. symbol.ttf can
typically be
found on any Windows machine.  I then edited the fonts.dir file, located
in the
same directory, adding the following line:
    symbol.ttf
-urw-symbol-medium-r-normal--0-0-0-0-p-0-adobe-fontspecific

Lastly, I edited the number at the top of the file, changing it from 12
to 13.
Once this is done, the font errors go away.

JBuilder also has the option of switching the L&F.  I changed it from
Metal to
CDE/Motif and it was MUCH better.  I know that this is only opinion but I
think
that it would be good for Jext to allow the same option.

Scot.

----------------------------

  Deming

Re: [Civlog-list] Font problem in Linux

[Raimar F. <hawk@A315-2b.WH8.TU-Dresden.De>]

On Tue, Sep 19, 2000 at 01:40:47PM +0800, Wong TM (Huang Deming) wrote:
> 
> When starting CivLog in linux, there some minor error message about the
> font not found. I notice that CivLog use a non-monotype font when display
> the help|Content. Did any of you guys had similar thing happening?

A lot of messages like
"Font specified in font.properties not found [-*-symbol-medium-r-normal--*-%d-*-*-p-*-adobe-fontspecific]"
when the VM starts? As long as I can remember using Java under Linux these
messages were always there. Never searched and solved this.

	Raimar

-- 
 "Reality? That's where the pizza delivery guy comes from!"

[Civlog-list] Font problem in Linux

[Wong TM (H. Deming) <loo...@ma...>]

When starting CivLog in linux, there some minor error message about the
font not found. I notice that CivLog use a non-monotype font when display
the help|Content. Did any of you guys had similar thing happening?

       __
.----.|__|.---.-..-----.
|  __||  ||  _  ||  _  |
|____||__||___._||_____|

[Civlog-list] Re: FreeCiv log files

[Raimar F. <hawk@A315-2b.WH8.TU-Dresden.De>]

On Mon, Sep 18, 2000 at 08:02:15PM +0200, Ulrich Kuhn wrote:
> On Mon, 18 Sep 2000, Raimar Falke wrote:
> 
> > The format you mentioned above is 2). CivLog currently reads 3). We also
> > considered to put the ability to read some parts of 2) into CivLog.
> 
> This will be problematic - I think. Because the gamelog is from what I
> have seen in no way standardized: Some texts are localized, some are not.
> If all is localized you will need a log parser for every language. In
> addition some information is missing ie when a country is split (there are
> suddenly cities removed from non existent countries...). And so on.

You can ask freeciv-dev what they think the format should be. A problem is
that the gamelog should be readable by humans but it should also contain
enough format to be parsed by a program. Maybe another log file is needed.
Or the scorelog can be extended.

> Does the information in the score log file include the one in the game
> log? At least for the science this is the case...

No. The total amount of technologies discovered is contained. But for
example no info by which way (hut, research, spy).

> Where is the score log format defined? 

server/report.c:log_civ_score

> > > P.S.: The last three days I have written a FreeCiv log analyser for the
> > > 'old' log format that displays the map and the placement of the cities
> > > varying with time.
> > 
> > In what language? 
> 
> Java.
> 
> > Have you looked at the attached civpow.pl (I'm sorry I
> > don't know who is the author)?
> 
> I could not start it. (With the odd message 'file or dir not found'.) 

Maybe perl is not at the place mentioned in the first line. Try:
  perl civpow.pl *.sav

This will produce a final.gif which is animated. You can use Netscape or
xanim to view it. I'm sorry but civpow.pl doesn't use the game log like I
first mentioned but the savefiles.

There is another one: civscore2gnuplot is like CivLog. It produces graphs
based on the scorelog. But CivLog is faster and better ;-)

	Raimar

-- 
 "Just because you put a flag on the moon doesn't make it yours, it just
  puts a hole in the moon."

[Civlog-list] Re: FreeCiv log files

[Ulrich K. <ku...@un...>]

On Mon, 18 Sep 2000, Raimar Falke wrote:

> On Mon, Sep 18, 2000 at 12:18:09PM +0200, Ulrich Kuhn wrote:
> > Hi,
> > 
> AFAIK there are three log files
>  1) the "plain" logile: set with command line arguments -l and --log
>  2) the game log file: set with command line arguments -g and --gamelog
>  3) the score log file: set inside the server with "set scorelog 1"

Thanks, this is exactly what I wanted to read. :)
(Thanks too to Deming for his efforts!)

> 
> The format you mentioned above is 2). CivLog currently reads 3). We also
> considered to put the ability to read some parts of 2) into CivLog.

This will be problematic - I think. Because the gamelog is from what I
have seen in no way standardized: Some texts are localized, some are not.
If all is localized you will need a log parser for every language. In
addition some information is missing ie when a country is split (there are
suddenly cities removed from non existent countries...). And so on.

Does the information in the score log file include the one in the game
log? At least for the science this is the case...

Where is the score log format defined? 


> 
> > P.S.: The last three days I have written a FreeCiv log analyser for the
> > 'old' log format that displays the map and the placement of the cities
> > varying with time.
> 
> In what language? 

Java.

> Have you looked at the attached civpow.pl (I'm sorry I
> don't know who is the author)?

I could not start it. (With the odd message 'file or dir not found'.) 

May the performance be with you
und wünsche wohl zu ruhen

Ulrich Kuhn

[Civlog-list] Re: FreeCiv log files

[Raimar F. <hawk@A315-2b.WH8.TU-Dresden.De>]

On Mon, Sep 18, 2000 at 03:33:03PM +0200, Raimar Falke wrote:
> In what language? Have you looked at the attached civpow.pl (I'm sorry I
> don't know who is the author)?

Forgot to attach the script. It's a known problem. I also noticed you are on
the civlog list.

	Raimar

-- 
  One nuclear bomb can ruin your whole day.

[Civlog-list] RE: FreeCiv log files

[Wong TM (H. Deming) <loo...@ma...>]

> -----Original Message-----
> From: Ulrich Kuhn [mailto:ku...@un...]
> Sent: Monday, September 18, 2000 6:18 PM
> To: loo...@ma...
> Subject: FreeCiv log files
>
>
> Hi,
>
> I have some questions about the log files in FreeCiv:
> Are there different versions? I found the gamelog option in the FreeCiv
> server and it produces a human readable file (with things like -3950
> Hobeljup (45, 47) founded by the Scots). This is not the same
> as the log
> files CivLog expects.
> I suppose the above format is the old one and CivLog expects a
> new one. If
> this is correct: Starting with which version FreeCiv produces
> the desired
> log file?

I afraid you had mistook CivLog for what it is not(my Readme isn't good
enough). CivLog read in 'civscore.log' Freeciv produce when the server
option 'scorelog'=1(sample log files are available from
http://myhome.asia1.com/home/f/furyhawk/civlog/index.html). It just
product a power graph of players' Population over time/turns for example.
It doesn't read in the 'old log'. Anyway I thanks you for your interest
in CivLog. I hope that you would try it and see.

I am interested what is the system you using?(win9x or linux)

>
> Yours Sincerely
> Ulrich Kuhn
>
> P.S.: The last three days I have written a FreeCiv log analyser for the
> 'old' log format that displays the map and the placement of the cities
> varying with time.

Interesting, feel free to join us, the CivLog small community. We
currently had CivLog setup on SourceForge and a mailing-list
<civ...@li...> provided. I will try to help you if
possible base on mine experience on CivLog.

Current CivLog members:
Wong Teck Meng <loo...@ma...>
Raimar Falke <rf...@in...>
Michael Grundel <mic...@gr...>

With you, three Germans and a Singaporean :)
You will also proof whether the discourse 'All Germans sleep on 2am' is
true :) Sorry it's just a joke among the CivLog's members.

  Deming

[Civlog-list] FW: FreeCiv log files

[Wong TM (H. Deming) <loo...@ma...>]

Hey guys here is our first customer :-) .

-----Original Message-----
From: Ulrich Kuhn [mailto:ku...@un...]
Sent: Monday, September 18, 2000 6:18 PM
To: loo...@ma...
Subject: FreeCiv log files


Hi,

I have some questions about the log files in FreeCiv:
Are there different versions? I found the gamelog option in the FreeCiv
server and it produces a human readable file (with things like -3950
Hobeljup (45, 47) founded by the Scots). This is not the same as the log
files CivLog expects.
I suppose the above format is the old one and CivLog expects a new one.
If
this is correct: Starting with which version FreeCiv produces the desired
log file?

Yours Sincerely
Ulrich Kuhn

P.S.: The last three days I have written a FreeCiv log analyser for the
'old' log format that displays the map and the placement of the cities
varying with time.

Re: [Civlog-list] Why oh why?

[Raimar F. <hawk@A315-2b.WH8.TU-Dresden.De>]

On Sat, Sep 16, 2000 at 10:41:42PM +0800, Wong TM (Huang Deming) wrote:
> 
> > > I try to use jikes. I compile jikes src on my box and the
> > resulting ELF
> > > exec. is 11 megabyte!
> >
> > My jikes is stripped and about 2.1 mb big.
> > Type
> >   file jikes
> > it should say something "...not stripped..."
> >   strip jikes
> >   file jikes
> > /usr/bin/jikes: ELF 32-bit LSB executable, Intel 80386,
> > version 1, dynamically linked (uses shared libs), stripped
> 
> Thanks alots. (I a linux newbie) What do I pass to ./configure in the
> first place to disable debug info?

It looks like there is no one. But you can use the install_strip target of
the makefile.

> > > I try make all2 and results:
> > > jikes: class.h:1119: void
> > > LocalVariableTable_attribute::AddLocalVariable(short
> > unsigned int, short
> > > unsigned int, short unsigned int, short unsigned int, short unsigned
> > > int): Assertion `end >= start' failed.
> >
> > I don't know. As you know my version is much older. My advise
> > is to take an rpm.
> > Or find and fix the bug. I don't know your C++ skills.
> 
> Oh I thought end >= start is from CivLog. Didn't expect jikes to have an
> internal error.

No software is bug free.

	Raimar

-- 
 "Are you saying that you actually used the Classpath Java AWT classes in 
  addition to the GTK peers and got them to display something?
  Wow.  That's way better than I did and I wrote the code!"
    -- Aaron M. Renn in the classpath mailing list

RE: [Civlog-list] Why oh why?

[Wong TM (H. Deming) <loo...@ma...>]

> -----Original Message-----
> From: civ...@li...
> [mailto:civ...@li...]On Behalf Of Raimar
> Falke
>
> On Sat, Sep 16, 2000 at 02:14:18PM +0800, Wong TM (Huang Deming) wrote:
> > > -----Original Message-----
> > > From: civ...@li...
> > > [mailto:civ...@li...]On Behalf Of Raimar
> > > Falke
> > >
>
> You are working as root. This isn't recommended.

:-) I am very lazy. I use 'su' sometime and find that the path change to
a shorter version '/usr/bin:/sbin....'.

> > I try to use jikes. I compile jikes src on my box and the
> resulting ELF
> > exec. is 11 megabyte!
>
> My jikes is stripped and about 2.1 mb big.
> Type
>   file jikes
> it should say something "...not stripped..."
>   strip jikes
>   file jikes
> /usr/bin/jikes: ELF 32-bit LSB executable, Intel 80386,
> version 1, dynamically linked (uses shared libs), stripped

Thanks alots. (I a linux newbie) What do I pass to ./configure in the
first place to disable debug info?

> > I try make all2 and results:
> > jikes: class.h:1119: void
> > LocalVariableTable_attribute::AddLocalVariable(short
> unsigned int, short
> > unsigned int, short unsigned int, short unsigned int, short unsigned
> > int): Assertion `end >= start' failed.
>
> I don't know. As you know my version is much older. My advise
> is to take an rpm.
> Or find and fix the bug. I don't know your C++ skills.

Oh I thought end >= start is from CivLog. Didn't expect jikes to have an
internal error.

> 	Raimar

	Deming