colossus-developers Mailing List for Colossus (Page 4)

Since Clemens is the only really doing anything, then whatever you
want to do is what will happen :)

Looking through the archives I think Bruno is the one re-packaging for
rpm distribution, so some insights on how to best support that would
be useful.

Some things you might want to consider Clemens:

Github is awesome.
You may find more people able to contribute by having the code available there.
Forking is easy, so it lowers the barrier for contributors.
It is even possible to make minor changes via a web browser on github.
You should check it out.

Maintenance should trump users.
If you are stuggling to support an old version of Java then move to
one you can support.
The existing users can still use an old version of Colossus until they
can upgrade.
But if you can't develop then no one gets new versions.
This may also be a barrier for people to contribute - i.e. they dont
have an old version of Java available to develop with.

Project structure.
I'm biased towards Maven (as a PMC member) but being able to
modularise the build bits might help.
If its layered right then it should be easier to understand things in
isolation (java packages are also doing this too).
Perhaps someone might be tempted to create a web based client (which
would mean they dont need all the java ui classes) which is hard to do
at the moment because everything is together.

regards
Barrie

On Sun, Apr 12, 2015 at 15:46:54 +0300,
  Clemens Katzer <cle...@cl...> wrote:
>
>What I see ahead then still is the question, how will we "nicely"
>identify builds and refer to them in terms of version on control. "this
>is a tes build from revision 5432" was somehow handy... at least for me.
>Git checksums are just ... unhandly.

You can use tags to label builds.

Correction:

Replace

> something at which I am the best imaginable candidate...

with

... I am NOT the best imaginable candidate

or even

... I am the worst imaginable candidate

:-)


On 2015-04-12 15:46, Clemens Katzer wrote:
> ok, thanks.
>
> If Java 7 is at it's EOL. Well, then let's follow good Colossus
> tradition and go for it now ;-)
>
> I am not to keen on github; haven't used it and getting familiar with
> that as well will only add to the effort.
>
> Primary reasons reasons for change is, as said: getting the old Java 
> 5
> Ant stuff working on Fedora 21 is so messy, that it's incentive 
> enough
> to replace that part. And git, well, since SVN is too slow.
>
> If Peter could look then into the maven stuff, would very much be
> appreciated.
>
> The only "real" dependency we have is, AFAIK, jdom.
>
> That "shaded whatever" I have first to check what it is. All in all, 
> I
> feel web start *is* handy. Any other mechanism -- could people then 
> play
> on a computer where they can't install anything (library, company
> laptop)?
>
>
> What I see ahead then still is the question, how will we "nicely"
> identify builds and refer to them in terms of version on control. 
> "this
> is a tes build from revision 5432" was somehow handy... at least for 
> me.
> Git checksums are just ... unhandly.
>
>
> The projects I had to do with at work did not use "proper" maven
> versioning (x.y-SNAPSHOT and for "each" release change to a new 
> snapshot
> version). Instead it was 1.0-SNAPSHOT forever. Not sure how much work
> that causes. And if one did 1.2, is working on new stuff: "hm, we're 
> not
> sure yet will it deserve a x++ or y++" - what x.y-SNAPSHOT then to 
> do,
> and how to create minor bugfixes on top of the just done release". 
> Then
> building those sources several times always as 1.2. does not feel 
> right
> either... 1.2.1-SNAPSHOT etc.?
>
>
> So, technically changing to git and/and or maven might not be so much
> work, but coming to a set of agreements how will we develop and 
> release
> in the new world... something at which I am the best imaginable
> candidate...
>
> But perhaps I just worry too much.
>
> Let's see. I try to look into the "go for git, though on sourceforge"
> in near future.
>
> Thx,
> Clemens
>
>
>
>
> On 2015-04-12 13:10, Peter Becker wrote:
>> Reposting since I realized I had sent it to Clemens only last 
>> time...
>>
>>  I haven't been active much lately but I do have opinions - since 
>> you
>> asked you'll get them
>>
>>  On 11/04/15 00:17, Clemens Katzer wrote:
>>
>>> [...]
>>
>>> Aspects to consider:
>>>
>>> Java 1.5 vs. 1.7
>>> SVN vs. Git
>>> Ant vs. Maven
>>> Java web start ?
>>>
>>> And in which order....
>>>
>>> ~~~~~~~~~~
>>>
>>> Java - I'd say it's more a question of consensus to do it. It's
>>> simply
>>> difficult to have a 1.5 build environment around forever.
>>  I would even consider Java 8. It has plenty of goodies in there
>> (Lambdas, Date&Time API, lots of IO utilities). And if I am not
>> mistaken Oracle has pushed it via the Java update mechanisms by now.
>> Java 7 on the other hand is past the end of public updates.
>>
>>  I don't remember the Ant build enough, but unless it is horribly
>> complicated I would consider this as a change right now, before
>> everything else.
>>
>>> Git:
>>>
>>> I assume (hope) that there's no big objections to switch to git.
>>> Last
>>> days when I worked with Colossus with SVN again, ... it was
>>> unbearable
>>> to wait and wait and ....
>>>
>>> I suppose it makes more sense to first switch to git, then the
>>> working
>>> is more efficient. (Yes, I've been working with git-svn, still it's
>>>
>>> clumsy). I would like to stay on sourceforge (instead of e.g.
>>> github),
>>> so we can continue to use at least the existing project
>>> infrastructure
>>> (home page, downloads, developers, tickets, ... ) - at least for
>>> now.
>>> I personally would not bother setting up git
>>  is so much nicer on its own, whereas SF has always been a bit of a
>> slow and painful experience. And then you get the whole ecosystem
>> around github, like the various free CI engines such as
>> https://travis-ci.org/ [2] (just to name one, there are others).
>>
>>  Leaving the web site and the mailing list on SF is probably better
>> at
>> first, but I would move the code over, maybe issue tracking as well.
>> Downloads could go either way, or you could point straight to a CI
>> build.
>>
>>  ~~~~~~~~~~
>>
>>  Maven: Is it possible/useful to split Colossus into modules?
>>  Yes.
>>  Or do we
>>
>>> lithic build; but how to produce then two
>>> (three, if battleland builder is inclcuded) jar files?
>>  That is asking for pain since you will be fighting Maven. Maven
>> makes
>> it easy to manage lots of modules with their own artifacts, it is 
>> not
>> intended for complex builds per module. You can build multiple
>> artifacts if they belong toge
>>
>>> Doc JAR and a ZIP holding it together. But multiple binary JARs
>>> should have multiple modules.
>>>          What type(s)
>> eds to be produced: 1 jar, several
>>  jars, war, [ear, ?], ...
>>  I suspect several JARs is your answer. IIRC Colossus is
>> dependency-free, but if you ever want to use some, I'd probably have
>> at least a shaded fat client and a JWS client, plus the server. The
>> battlelands builder could just be a command line variation to the
>> normal client, it could be a separate tool. Using shaded JARs, you
>> could also have a build that create
>>
>>> just assembling that as a dependency with different main classes in
>>> executable JARs.
>>
>>  1) What are candidates for modules?
>>
>>  Ideally I'd like to see those following as parts, not sure if it's
>>  possible:
>>
>>  a) Common stuff
>>  b) The client part of playing (including AI?)
>>  c) The server part of playing
>>  d) The webserver client
>>  e) The webserver actual server
>>  f) Variant resources
>>
>>    From that one could pick together to build the following
>>  "applications":
>>
>>> jar), including game server side
>>> to host games
>>> - The "webserver" software/application (ATM ColossusWeb.jar)
>>>     Note that this ATM does not include game server side logic -
>>> it runs
>>> each game in an own JVM, though headless, using the "standard"
>>> Colossus.jar.
>>> - ? A thin client just for playing on the public server?
>>>
>>> I take it as granted that we would need to be able to get rid of
>>> cyclic
>>> dependencies between those which we want to split. I think existing
>>> the
>>> packages are "quite" much in tree-form already (yes, there's still
>>> issues).
>>>
>>> Or first change to a monolithic approach, and split into modules
>>> later
>>> (i.e., never)?
>>> I think all of that sounds about right. Inside the "Common stuff"
>>> you could probably consider more detail, such as model vs. utility
>>> code. I'd set up a parent/module configuration straight away, then
>>> start refactoring bottom up: put the utils in a separate module 
>>> that
>>> everything else can depend on, then the model, then higher level
>>> stuff.
>>>
>>> 2) How does that effect the deployment => web start?
>>>
>>> It seems that maven projects normally create many jar files, and
>>> build
>>> ear or war files out of tha
>> uld people play with that?
>>
>>  - can they locally run such an *ar file with a "normal" java? Or do
>>  they need some tomcat whatever?
>>  - can war files be used with Java web start?
>>
>>  Is there any alternatives - how is such stuff handled nowadays in
>> other
>>  projects?
>>  JWS never really caught on that much and
>>
>>> der. It's still viable to some extent, but in many ways just
>>> producing a shaded, executable JAR is superior. See
>>>
>>
>> 
>> https://maven.apache.org/plugins/maven-shade-plugin/examples/executable-jar.html
>>> [1]
>>>
>>> If the desktop environment is set up the right way, the JAR can be
>>> started like a normal application. I usually put some extra script
>>> files into a ZIP, just containing a single line calling "java
>>> -jar"/"javaw -jar". On Windows the latter stops the command window
>>> from appearing.
>>>
>>> These files don't need signin
>> s the whole thing much easier. They do not allow automated or 
>> partial
>> updates, but that doesn't seem that important for Colossus.
>>
>>  ~~~~~~~~~~~~~
>>
>>  Last not least, I'd like to see this new Colossus then emerge as
>>  Version 1.0 or even 2.0 ....
>>
>>  Given that it is pretty functional and stable a final number seems
>> appropriate. I propose to use semantic versioning 
>> (http://semver.org/
>> [3]), what number the MAJOR starts with doesn't really matter. I'd 
>> go
>> with the 1.0 since that seems the natural place to start, but I 
>> don't
>> really care.
>>
>>    Peter
>>
>>  PS: in terms of order: I don't think there are strong dependencies.
>> Any order should work, I'd go with whatever is either easier for you
>> or bothers you most. I could help you with the Maven migration if 
>> you
>> want me to, in which case I would have a preference for git first
>> since I don't even have a checkout at the moment. But I can still
>> handle Subversion, even on SF
>>
>>>
>>
>>
>> Links:
>> ------
>> [1]
>>
>> 
>> https://maven.apache.org/plugins/maven-shade-plugin/examples/executable-jar.html
>> [2] https://travis-ci.org/
>> [3] http://semver.org/
>
>
> 
> ------------------------------------------------------------------------------
> BPM Camp - Free Virtual Workshop May 6th at 10am PDT/1PM EDT
> Develop your own process in accordance with the BPMN 2 standard
> Learn Process modeling best practices with Bonita BPM through live 
> exercises
> http://www.bonitasoft.com/be-part-of-it/events/bpm-camp-virtual- 
> event?utm_
> 
> source=Sourceforge_BPM_Camp_5_6_15&utm_medium=email&utm_campaign=VA_SF
> _______________________________________________
> Colossus-developers mailing list
> Col...@li...
> https://lists.sourceforge.net/lists/listinfo/colossus-developers

ok, thanks.

If Java 7 is at it's EOL. Well, then let's follow good Colossus 
tradition and go for it now ;-)

I am not to keen on github; haven't used it and getting familiar with 
that as well will only add to the effort.

Primary reasons reasons for change is, as said: getting the old Java 5 
Ant stuff working on Fedora 21 is so messy, that it's incentive enough 
to replace that part. And git, well, since SVN is too slow.

If Peter could look then into the maven stuff, would very much be 
appreciated.

The only "real" dependency we have is, AFAIK, jdom.

That "shaded whatever" I have first to check what it is. All in all, I 
feel web start *is* handy. Any other mechanism -- could people then play 
on a computer where they can't install anything (library, company 
laptop)?


What I see ahead then still is the question, how will we "nicely" 
identify builds and refer to them in terms of version on control. "this 
is a tes build from revision 5432" was somehow handy... at least for me. 
Git checksums are just ... unhandly.


The projects I had to do with at work did not use "proper" maven 
versioning (x.y-SNAPSHOT and for "each" release change to a new snapshot 
version). Instead it was 1.0-SNAPSHOT forever. Not sure how much work 
that causes. And if one did 1.2, is working on new stuff: "hm, we're not 
sure yet will it deserve a x++ or y++" - what x.y-SNAPSHOT then to do, 
and how to create minor bugfixes on top of the just done release". Then 
building those sources several times always as 1.2. does not feel right 
either... 1.2.1-SNAPSHOT etc.?


So, technically changing to git and/and or maven might not be so much 
work, but coming to a set of agreements how will we develop and release 
in the new world... something at which I am the best imaginable 
candidate...

But perhaps I just worry too much.

Let's see. I try to look into the "go for git, though on sourceforge" 
in near future.

Thx,
Clemens




On 2015-04-12 13:10, Peter Becker wrote:
> Reposting since I realized I had sent it to Clemens only last time...
>
>  I haven't been active much lately but I do have opinions - since you
> asked you'll get them
>
>  On 11/04/15 00:17, Clemens Katzer wrote:
>
>> [...]
>
>> Aspects to consider:
>>
>> Java 1.5 vs. 1.7
>> SVN vs. Git
>> Ant vs. Maven
>> Java web start ?
>>
>> And in which order....
>>
>> ~~~~~~~~~~
>>
>> Java - I'd say it's more a question of consensus to do it. It's
>> simply
>> difficult to have a 1.5 build environment around forever.
>  I would even consider Java 8. It has plenty of goodies in there
> (Lambdas, Date&Time API, lots of IO utilities). And if I am not
> mistaken Oracle has pushed it via the Java update mechanisms by now.
> Java 7 on the other hand is past the end of public updates.
>
>  I don't remember the Ant build enough, but unless it is horribly
> complicated I would consider this as a change right now, before
> everything else.
>
>> Git:
>>
>> I assume (hope) that there's no big objections to switch to git.
>> Last
>> days when I worked with Colossus with SVN again, ... it was
>> unbearable
>> to wait and wait and ....
>>
>> I suppose it makes more sense to first switch to git, then the
>> working
>> is more efficient. (Yes, I've been working with git-svn, still it's
>>
>> clumsy). I would like to stay on sourceforge (instead of e.g.
>> github),
>> so we can continue to use at least the existing project
>> infrastructure
>> (home page, downloads, developers, tickets, ... ) - at least for
>> now.
>> I personally would not bother setting up git
>  is so much nicer on its own, whereas SF has always been a bit of a
> slow and painful experience. And then you get the whole ecosystem
> around github, like the various free CI engines such as
> https://travis-ci.org/ [2] (just to name one, there are others).
>
>  Leaving the web site and the mailing list on SF is probably better 
> at
> first, but I would move the code over, maybe issue tracking as well.
> Downloads could go either way, or you could point straight to a CI
> build.
>
>  ~~~~~~~~~~
>
>  Maven: Is it possible/useful to split Colossus into modules?
>  Yes.
>  Or do we
>
>> lithic build; but how to produce then two
>> (three, if battleland builder is inclcuded) jar files?
>  That is asking for pain since you will be fighting Maven. Maven 
> makes
> it easy to manage lots of modules with their own artifacts, it is not
> intended for complex builds per module. You can build multiple
> artifacts if they belong toge
>
>> Doc JAR and a ZIP holding it together. But multiple binary JARs
>> should have multiple modules.
>>          What type(s)
> eds to be produced: 1 jar, several
>  jars, war, [ear, ?], ...
>  I suspect several JARs is your answer. IIRC Colossus is
> dependency-free, but if you ever want to use some, I'd probably have
> at least a shaded fat client and a JWS client, plus the server. The
> battlelands builder could just be a command line variation to the
> normal client, it could be a separate tool. Using shaded JARs, you
> could also have a build that create
>
>> just assembling that as a dependency with different main classes in
>> executable JARs.
>
>  1) What are candidates for modules?
>
>  Ideally I'd like to see those following as parts, not sure if it's
>  possible:
>
>  a) Common stuff
>  b) The client part of playing (including AI?)
>  c) The server part of playing
>  d) The webserver client
>  e) The webserver actual server
>  f) Variant resources
>
>    From that one could pick together to build the following
>  "applications":
>
>> jar), including game server side
>> to host games
>> - The "webserver" software/application (ATM ColossusWeb.jar)
>>     Note that this ATM does not include game server side logic -
>> it runs
>> each game in an own JVM, though headless, using the "standard"
>> Colossus.jar.
>> - ? A thin client just for playing on the public server?
>>
>> I take it as granted that we would need to be able to get rid of
>> cyclic
>> dependencies between those which we want to split. I think existing
>> the
>> packages are "quite" much in tree-form already (yes, there's still
>> issues).
>>
>> Or first change to a monolithic approach, and split into modules
>> later
>> (i.e., never)?
>> I think all of that sounds about right. Inside the "Common stuff"
>> you could probably consider more detail, such as model vs. utility
>> code. I'd set up a parent/module configuration straight away, then
>> start refactoring bottom up: put the utils in a separate module that
>> everything else can depend on, then the model, then higher level
>> stuff.
>>
>> 2) How does that effect the deployment => web start?
>>
>> It seems that maven projects normally create many jar files, and
>> build
>> ear or war files out of tha
> uld people play with that?
>
>  - can they locally run such an *ar file with a "normal" java? Or do
>  they need some tomcat whatever?
>  - can war files be used with Java web start?
>
>  Is there any alternatives - how is such stuff handled nowadays in
> other
>  projects?
>  JWS never really caught on that much and
>
>> der. It's still viable to some extent, but in many ways just
>> producing a shaded, executable JAR is superior. See
>>
> 
> https://maven.apache.org/plugins/maven-shade-plugin/examples/executable-jar.html
>> [1]
>>
>> If the desktop environment is set up the right way, the JAR can be
>> started like a normal application. I usually put some extra script
>> files into a ZIP, just containing a single line calling "java
>> -jar"/"javaw -jar". On Windows the latter stops the command window
>> from appearing.
>>
>> These files don't need signin
> s the whole thing much easier. They do not allow automated or partial
> updates, but that doesn't seem that important for Colossus.
>
>  ~~~~~~~~~~~~~
>
>  Last not least, I'd like to see this new Colossus then emerge as
>  Version 1.0 or even 2.0 ....
>
>  Given that it is pretty functional and stable a final number seems
> appropriate. I propose to use semantic versioning (http://semver.org/
> [3]), what number the MAJOR starts with doesn't really matter. I'd go
> with the 1.0 since that seems the natural place to start, but I don't
> really care.
>
>    Peter
>
>  PS: in terms of order: I don't think there are strong dependencies.
> Any order should work, I'd go with whatever is either easier for you
> or bothers you most. I could help you with the Maven migration if you
> want me to, in which case I would have a preference for git first
> since I don't even have a checkout at the moment. But I can still
> handle Subversion, even on SF
>
>>
>
>
> Links:
> ------
> [1]
> 
> https://maven.apache.org/plugins/maven-shade-plugin/examples/executable-jar.html
> [2] https://travis-ci.org/
> [3] http://semver.org/

Reposting since I realized I had sent it to Clemens only last time...

I haven't been active much lately but I do have opinions - since you 
asked you'll get them

On 11/04/15 00:17, Clemens Katzer wrote:
> [...]

> Aspects to consider:
>
> Java 1.5 vs. 1.7
> SVN vs. Git
> Ant vs. Maven
> Java web start ?
>
> And in which order....
>
>
> ~~~~~~~~~~
>
> Java - I'd say it's more a question of consensus to do it. It's simply
> difficult to have a 1.5 build environment around forever.
I would even consider Java 8. It has plenty of goodies in there 
(Lambdas, Date&Time API, lots of IO utilities). And if I am not mistaken 
Oracle has pushed it via the Java update mechanisms by now. Java 7 on 
the other hand is past the end of public updates.

I don't remember the Ant build enough, but unless it is horribly 
complicated I would consider this as a change right now, before 
everything else.
>
> ~~~~~~~~~
>
> Git:
>
> I assume (hope) that there's no big objections to switch to git. Last
> days when I worked with Colossus with SVN again, ... it was unbearable
> to wait and wait and ....
>
> I suppose it makes more sense to first switch to git, then the working
> is more efficient. (Yes, I've been working with git-svn, still it's
> clumsy). I would like to stay on sourceforge (instead of e.g. github),
> so we can continue to use at least the existing project infrastructure
> (home page, downloads, developers, tickets, ... ) - at least for now.
I personally would not bother setting up git on SF. Github is so much 
nicer on its own, whereas SF has always been a bit of a slow and painful 
experience. And then you get the whole ecosystem around github, like the 
various free CI engines such as https://travis-ci.org/ (just to name 
one, there are others).

Leaving the web site and the mailing list on SF is probably better at 
first, but I would move the code over, maybe issue tracking as well. 
Downloads could go either way, or you could point straight to a CI build.
>
> ~~~~~~~~~~
>
> Maven: Is it possible/useful to split Colossus into modules?
Yes.
> Or do we
> simply stay in one big monolithic build; but how to produce then two
> (three, if battleland builder is inclcuded) jar files?
That is asking for pain since you will be fighting Maven. Maven makes it 
easy to manage lots of modules with their own artifacts, it is not 
intended for complex builds per module. You can build multiple artifacts 
if they belong together, e.g. a JAR, the source JAR, the JavaDoc JAR and 
a ZIP holding it together. But multiple binary JARs should have multiple 
modules.
>          What type(s) of binary needs to be produced: 1 jar, several
> jars, war, [ear, ?], ...
I suspect several JARs is your answer. IIRC Colossus is dependency-free, 
but if you ever want to use some, I'd probably have at least a shaded 
fat client and a JWS client, plus the server. The battlelands builder 
could just be a command line variation to the normal client, it could be 
a separate tool. Using shaded JARs, you could also have a build that 
creates a JAR with all the GUI, then two on top just assembling that as 
a dependency with different main classes in executable JARs.
>
>
> 1) What are candidates for modules?
>
> Ideally I'd like to see those following as parts, not sure if it's
> possible:
>
> a) Common stuff
> b) The client part of playing (including AI?)
> c) The server part of playing
> d) The webserver client
> e) The webserver actual server
> f) Variant resources
>
>   From that one could pick together to build the following
> "applications":
>
> - A full application as now (Colossus.jar), including game server side
> to host games
> - The "webserver" software/application (ATM ColossusWeb.jar)
>     Note that this ATM does not include game server side logic - it runs
> each game in an own JVM, though headless, using the "standard"
> Colossus.jar.
> - ? A thin client just for playing on the public server?
>
>
> I take it as granted that we would need to be able to get rid of cyclic
> dependencies between those which we want to split. I think existing the
> packages are "quite" much in tree-form already (yes, there's still
> issues).
>
> Or first change to a monolithic approach, and split into modules later
> (i.e., never)?
I think all of that sounds about right. Inside the "Common stuff" you 
could probably consider more detail, such as model vs. utility code. I'd 
set up a parent/module configuration straight away, then start 
refactoring bottom up: put the utils in a separate module that 
everything else can depend on, then the model, then higher level stuff.
>
>
> 2) How does that effect the deployment => web start?
>
>
> It seems that maven projects normally create many jar files, and build
> ear or war files out of that. But how would people play with that?
>
> - can they locally run such an *ar file with a "normal" java? Or do
> they need some tomcat whatever?
> - can war files be used with Java web start?
>
>
> Is there any alternatives - how is such stuff handled nowadays in other
> projects?
JWS never really caught on that much and all the security enhancements 
make it harder. It's still viable to some extent, but in many ways just 
producing a shaded, executable JAR is superior. See 
https://maven.apache.org/plugins/maven-shade-plugin/examples/executable-jar.html 


If the desktop environment is set up the right way, the JAR can be 
started like a normal application. I usually put some extra script files 
into a ZIP, just containing a single line calling "java -jar"/"javaw 
-jar". On Windows the latter stops the command window from appearing.

These files don't need signing, which makes the whole thing much easier. 
They do not allow automated or partial updates, but that doesn't seem 
that important for Colossus.
>
> ~~~~~~~~~~~~~
>
> Last not least, I'd like to see this new Colossus then emerge as
> Version 1.0 or even 2.0 ....
>
Given that it is pretty functional and stable a final number seems 
appropriate. I propose to use semantic versioning (http://semver.org/), 
what number the MAJOR starts with doesn't really matter. I'd go with the 
1.0 since that seems the natural place to start, but I don't really care.

   Peter


PS: in terms of order: I don't think there are strong dependencies. Any 
order should work, I'd go with whatever is either easier for you or 
bothers you most. I could help you with the Maven migration if you want 
me to, in which case I would have a preference for git first since I 
don't even have a checkout at the moment. But I can still handle 
Subversion, even on SF

On 04/10/2015 10:17 AM, Clemens Katzer wrote:

> making new builds with existing codebase becomes more and more 
> difficult. Had to go back to an ancient WinXP desktop to be able to use 
> svnkit and jarsigning...
> Perhaps it's really time to port something to today's world.
> 
> So I'd like to share my thoughts on this with you, and collect 
> suggestions especially from those who have experience with this type of 
> legacy porting. I would expect that some of you have at the one point or 
> another already spent some thoughts on this.

Good idea.  I can't help though.  (New employer has a games division so
doesn't let me work on open source games.)

> Java - I'd say it's more a question of consensus to do it. It's simply 
> difficult to have a 1.5 build environment around forever.

New Java should be 99.9% compatible with old Java so this should be
easy.  We've always been conservative about required version to avoid
losing any players on old hardware or weird OS.  But Java 7 is 4 years
old, which seems plenty long enough.

> Git:
> 
> I assume (hope) that there's no big objections to switch to git. Last 
> days when I worked with Colossus with SVN again, ... it was unbearable 
> to wait and wait and ....
> 
> I suppose it makes more sense to first switch to git, then the working 
> is more efficient. (Yes, I've been working with git-svn, still it's 
> clumsy). I would like to stay on sourceforge (instead of e.g. github), 
> so we can continue to use at least the existing project infrastructure 
> (home page, downloads, developers, tickets, ... ) - at least for now.

Sure, use what makes you more productive.  Converting from svn to git
without losing history shouldn't take more than a couple of hours.

> Maven: Is it possible/useful to split Colossus into modules? Or do we 
> simply stay in one big monolithic build; but how to produce then two 
> (three, if battleland builder is inclcuded) jar files?
>         What type(s) of binary needs to be produced: 1 jar, several 
> jars, war, [ear, ?], ...
> 
> 
> 1) What are candidates for modules?
> 
> Ideally I'd like to see those following as parts, not sure if it's 
> possible:
> 
> a) Common stuff
> b) The client part of playing (including AI?)
> c) The server part of playing
> d) The webserver client
> e) The webserver actual server
> f) Variant resources
> 
>  From that one could pick together to build the following  
> "applications":
> 
> - A full application as now (Colossus.jar), including game server side 
> to host games
> - The "webserver" software/application (ATM ColossusWeb.jar)
>    Note that this ATM does not include game server side logic - it runs 
> each game in an own JVM, though headless, using the "standard" 
> Colossus.jar.
> - ? A thin client just for playing on the public server?
> 
> 
> I take it as granted that we would need to be able to get rid of cyclic 
> dependencies between those which we want to split. I think existing the 
> packages are "quite" much in tree-form already (yes, there's still 
> issues).
> 
> Or first change to a monolithic approach, and split into modules later 
> (i.e., never)?

Things players care about:

1. Is the game fun?
2. Is the game easy to run?  (JWS helps a ton here.  My tech support
burden went way down once we introduced it.)

Things players don't care about:

Maven, modules, wars, etc.

Which is to say that I wouldn't spend a lot of time rearranging the deck
chairs for the sake of rearranging the deck chairs.  If you see direct
benefit, go ahead, but if it's all theoretical, I'd suggest working on
something else.

My opinion is that you need to keep JWS working, for the ability to
either run a standalone game or connect to a server, to make things easy
for players.

I haven't done any jar signing in a while so I'm not sure about the
current state of self-signed certs vs. free certs vs. overpriced certs.
 (What a racket, charging people hundreds of dollars to verify their
identity, then not really doing much work to actually verify their
identity.)  If there are free certs that don't give self-signed
warnings, awesome.

> Last not least, I'd like to see this new Colossus then emerge as 
> Version 1.0 or even 2.0 ....

Sure.  I think the network protocol has been stable for long enough to
deserve a version number, and releases have been rare enough that
maintaining it would not be a huge burden.

-- 
David Ripton    dr...@ri...

On 11/04/15 07:18, John David Galt wrote:
>> On Thu, Jan 16, 2014 at 09:48:46 +0200,
>>     Clemens Katzer <cle...@cl...> wrote:
>>> As we discussed earlier, there seems to be no way around it so a real
>>> certificate is needed (other than running it locally, download, unpack
>>> and then run with the script or doubleclick the jar, which is not really
>>> suitable for the average user).
> On Thu, 16 Jan 2014 07:46:30 -0600, Bruno Wolff III wrote:
>> I don't think "Real" is the right way to describe this. "Blessed" would
>> probably be closer to what is really meant. The top level CA certs
>> are just as self signed as yours are. But yours isn't in the special
>> list of automatically trusted certificates.
> [snip]
>> There are one or two places that will give you certs based on being able
>> to prove control of an email domain, that are in some of the blessed lists.
>> Maybe one of those will work for java web start?
> EFF recently established its own CA which will give you a cert free.
> https://www.eff.org/deeplinks/2014/11/certificate-authority-encrypt-entire-web
>

That sounds pretty cool, but from what I see it is going to be for HTTPS 
only, not for code signing. I didn't dive into their Github site, but 
the website seems very much web centric. If I'm not mistaken JWS needs a 
code signing certificate.

   Peter

> On Thu, Jan 16, 2014 at 09:48:46 +0200,
>    Clemens Katzer <cle...@cl...> wrote:
>> As we discussed earlier, there seems to be no way around it so a real
>> certificate is needed (other than running it locally, download, unpack
>> and then run with the script or doubleclick the jar, which is not really
>> suitable for the average user).

On Thu, 16 Jan 2014 07:46:30 -0600, Bruno Wolff III wrote:
> I don't think "Real" is the right way to describe this. "Blessed" would 
> probably be closer to what is really meant. The top level CA certs 
> are just as self signed as yours are. But yours isn't in the special 
> list of automatically trusted certificates.
[snip]
> There are one or two places that will give you certs based on being able 
> to prove control of an email domain, that are in some of the blessed lists. 
> Maybe one of those will work for java web start?

EFF recently established its own CA which will give you a cert free.
https://www.eff.org/deeplinks/2014/11/certificate-authority-encrypt-entire-web

Hey folks,

making new builds with existing codebase becomes more and more 
difficult. Had to go back to an ancient WinXP desktop to be able to use 
svnkit and jarsigning...
Perhaps it's really time to port something to today's world.

So I'd like to share my thoughts on this with you, and collect 
suggestions especially from those who have experience with this type of 
legacy porting. I would expect that some of you have at the one point or 
another already spent some thoughts on this.


Aspects to consider:

Java 1.5 vs. 1.7
SVN vs. Git
Ant vs. Maven
Java web start ?

And in which order....


~~~~~~~~~~

Java - I'd say it's more a question of consensus to do it. It's simply 
difficult to have a 1.5 build environment around forever.

~~~~~~~~~

Git:

I assume (hope) that there's no big objections to switch to git. Last 
days when I worked with Colossus with SVN again, ... it was unbearable 
to wait and wait and ....

I suppose it makes more sense to first switch to git, then the working 
is more efficient. (Yes, I've been working with git-svn, still it's 
clumsy). I would like to stay on sourceforge (instead of e.g. github), 
so we can continue to use at least the existing project infrastructure 
(home page, downloads, developers, tickets, ... ) - at least for now.

~~~~~~~~~~

Maven: Is it possible/useful to split Colossus into modules? Or do we 
simply stay in one big monolithic build; but how to produce then two 
(three, if battleland builder is inclcuded) jar files?
        What type(s) of binary needs to be produced: 1 jar, several 
jars, war, [ear, ?], ...


1) What are candidates for modules?

Ideally I'd like to see those following as parts, not sure if it's 
possible:

a) Common stuff
b) The client part of playing (including AI?)
c) The server part of playing
d) The webserver client
e) The webserver actual server
f) Variant resources

 From that one could pick together to build the following  
"applications":

- A full application as now (Colossus.jar), including game server side 
to host games
- The "webserver" software/application (ATM ColossusWeb.jar)
   Note that this ATM does not include game server side logic - it runs 
each game in an own JVM, though headless, using the "standard" 
Colossus.jar.
- ? A thin client just for playing on the public server?


I take it as granted that we would need to be able to get rid of cyclic 
dependencies between those which we want to split. I think existing the 
packages are "quite" much in tree-form already (yes, there's still 
issues).

Or first change to a monolithic approach, and split into modules later 
(i.e., never)?


2) How does that effect the deployment => web start?


It seems that maven projects normally create many jar files, and build 
ear or war files out of that. But how would people play with that?

- can they locally run such an *ar file with a "normal" java? Or do 
they need some tomcat whatever?
- can war files be used with Java web start?


Is there any alternatives - how is such stuff handled nowadays in other 
projects?

~~~~~~~~~~~~~

Last not least, I'd like to see this new Colossus then emerge as 
Version 1.0 or even 2.0 ....

Opinions?


Thanks,
Clemens

On Sun, Feb 08, 2015 at 17:50:47 +0200,
  Clemens Katzer <cle...@cl...> wrote:
>
>I have now committed a change (r5369) which, I think, resolves this. 
>It was, after all, almost trivial. (Where's the pitfall?)
>
>( I had looked into that earlier, and commented somewhere "this is not 
>that easy". I guess somehow tried to rearrange the whole 
>awardPointd/acquire/reinforce logic. Much more straightforward is to 
>move resurrectImmortals() after all those stuff, into finishBattle()).
>
>Tested it once locally.
>
>Problem is, I had made 20+ changes since last version (0.14.0, r5331), 
>and can hardly remember what they are all about.... so, not so easy to 
>just make new public version. The game jar on public server, however, 
>is r5358 ;-)
>
>Hm... I could make a 14.1 or 14.0.1 in a branch....

I don't have any strong preference. If a new version of some sort is put 
out, I'll give it a try.

I have now committed a change (r5369) which, I think, resolves this. It 
was, after all, almost trivial. (Where's the pitfall?)

( I had looked into that earlier, and commented somewhere "this is not 
that easy". I guess somehow tried to rearrange the whole 
awardPointd/acquire/reinforce logic. Much more straightforward is to 
move resurrectImmortals() after all those stuff, into finishBattle()).

Tested it once locally.

Problem is, I had made 20+ changes since last version (0.14.0, r5331), 
and can hardly remember what they are all about.... so, not so easy to 
just make new public version. The game jar on public server, however, is 
r5358 ;-)

Hm... I could make a 14.1 or 14.0.1 in a branch....

-Clemens


On 2015-02-03 16:23, Bruno Wolff III wrote:
> On Tue, Feb 03, 2015 at 15:05:23 +0200,
>  Clemens Katzer <cle...@cl...> wrote:
>>
>>Does it matter whether the other legion contained Titan (=all other
>>legions to be eliminated) ? I guess not, because the cleanup of other
>>legions happens as part of the player die'ing.
>
> Not for this one.
>
>>But if it did not involve the Titan, and now there is another
>>engagement still there, the resurrect has to happen before 
>> that,right?
>
> Yes the resurrection is per engagement. So the lord or demilord would
> be available for subsequent engagements in the same turn.
>
>>I.e. then when one engagement is completed (incl. summon+recruit), 
>> but
>>it does not wait until all engagements are done.
>>
>>Right?
>
> Correct.

On Tue, Feb 03, 2015 at 15:05:23 +0200,
  Clemens Katzer <cle...@cl...> wrote:
>
>Does it matter whether the other legion contained Titan (=all other
>legions to be eliminated) ? I guess not, because the cleanup of other
>legions happens as part of the player die'ing.

Not for this one.

>But if it did not involve the Titan, and now there is another
>engagement still there, the resurrect has to happen before that,right?

Yes the resurrection is per engagement. So the lord or demilord would be 
available for subsequent engagements in the same turn.

>I.e. then when one engagement is completed (incl. summon+recruit), but
>it does not wait until all engagements are done.
>
>Right?

Correct.

On 2015-02-03 05:19, Bruno Wolff III wrote:
> This might be related to bug 703 as it also involves clean up being 
> done
> before a battle was over. Characters tha come back after being slain,
> don't do so until the engagement in which they have slain has been
> resolved.
>
> Tonight I had a game where I was offered a chance to take a warlock
> reinforcent after slaying all of the attacking characters, even 
> though
> the only warlocks availble were ones slain in this battle.

So technically we should call the "recurrectImmortals()" after that 
particular battle has been completed; i.e. it is already in 
summon/recruit-after battle phase and now does advanceBattlePhase (but 
before advancePhase()).

Does it matter whether the other legion contained Titan (=all other 
legions to be eliminated) ? I guess not, because the cleanup of other 
legions happens as part of the player die'ing.

But if it did not involve the Titan, and now there is another 
engagement still there, the resurrect has to happen before that,right?

I.e. then when one engagement is completed (incl. summon+recruit), but 
it does not wait until all engagements are done.

Right?

BR;
-Clemens

This might be related to bug 703 as it also involves clean up being done 
before a battle was over. Characters tha come back after being slain, 
don't do so until the engagement in which they have slain has been 
resolved.

Tonight I had a game where I was offered a chance to take a warlock 
reinforcent after slaying all of the attacking characters, even though 
the only warlocks availble were ones slain in this battle.

Yes,

would perhaps be a good idea, but for several reasons I believe it 
would not fit there, and anyway I have just made a new contract for a 
physical machine (for various other private reasons).

But thanks for the idea!

Thx,
Clemens



On 2014-03-06 01:42, Barrie Treloar wrote:
> https://www.openshift.com/products/pricing
>
> I've bumped into openshift.
> I haven't yet done enough reading but it looks like you might be able
> to push colossus over there and save some money.
>
> I'll keep reading.
>
> 
> ------------------------------------------------------------------------------
> Subversion Kills Productivity. Get off Subversion & Make the Move to
> Perforce.
> With Perforce, you get hassle-free workflows. Merge that actually 
> works.
> Faster operations. Version large binaries.  Built-in WAN optimization 
> and the
> freedom to use Git, Perforce or both. Make the move to Perforce.
> 
> http://pubads.g.doubleclick.net/gampad/clk?id=122218951&iu=/4140/ostg.clktrk
> _______________________________________________
> Colossus-developers mailing list
> Col...@li...
> https://lists.sourceforge.net/lists/listinfo/colossus-developers

https://www.openshift.com/products/pricing

I've bumped into openshift.
I haven't yet done enough reading but it looks like you might be able
to push colossus over there and save some money.

I'll keep reading.

On Thu, Jan 16, 2014 at 09:48:46 +0200,
   Clemens Katzer <cle...@cl...> wrote:
>
>As we discussed earlier, there seems to be no way around it so a real
>certificate is needed (other than running it locally, download, unpack
>and then run with the script or doubleclick the jar, which is not really
>suitable for the average user).

I don't think "Real" is the right way to describe this. "Blessed" would 
probably be closer to what is really meant. The top level CA certs 
are just as self signed as yours are. But yours isn't in the special 
list of automatically trusted certificates.

>I will add a notification about it in the Public Game Server that if
>one wants to play Colossus one should refrain from updating Java as long
>as this is not solved.

That may not be that great. Depending on how old their java instance is, 
there may be real attacks going on against it.

>As a recap, the problem with the certificate is not the money; it is
>that bureaucratically-wise I can't get one at the moment.

There are one or two places that will give you certs based on being able 
to prove control of an email domain, that are in some of the blessed lists. 
Maybe one of those will work for java web start?

hei-ho,

> Today colossus asked me to update java as I started the game.
> Once it was done it told me that my security settings are blocking 
> it.
> Has anyone else been unable to play today?



Looks like the security tightening with Java has struck.

As we discussed earlier, there seems to be no way around it so a real 
certificate is needed (other than running it locally, download, unpack 
and then run with the script or doubleclick the jar, which is not really 
suitable for the average user).

I will add a notification about it in the Public Game Server that if 
one wants to play Colossus one should refrain from updating Java as long 
as this is not solved.



As a recap, the problem with the certificate is not the money; it is 
that bureaucratically-wise I can't get one at the moment.

( I checked e.g. ComodoSSL. For a certificate to be granted, I need a 
domain which is registered in my name, with my address, and my phone 
number (they will call back there to confirm). (And I am fairly certain 
that other certificate authorities will have pretty much the same 
authentication requirements). The domain play-colossus.net is hosted by 
Strato, but they only deal with German addresses, thus the contract is 
in my name with my sisters address and phone number. I have already for 
a while planned to change the hosting provider (Metzner) where I can 
then use my real address etc.; I would use also a physical server 
instead of virtual, because at the moment it feels that frequently my 
virtual server (which has 0.01 load 99% of the time) sometime does not 
get the CPU for even 1-3 seconds. Having a physical server with 4 cores 
(the smallest they have) will also give me possibilities to do a lot of 
other things I've been dreaming about. For example, AIs could be run on 
2 of the 4 cores. Drawback is, that such a server costs 50€ instead of 
25€ as I have now, or 10€ or so as a virtual server with same specs 
as my current one nowadays would cost. The real problem is, that this 
involves paperwork  ;-)))     And of course installing the new server 
from scratch; though going to CentOS 6 is overdue anyway.

(( well the 50€ instead of 25€ is somewhat an issue as my wife is 
not really thrilled about me spending that much then regularly, but 
well, I think I can live with it :->  As said, there's additional 
advantages to that change. :))
)


*sigh*

But now, as there is a real reason "to act now" chances are improving 
that I will do something in near future :)

Thx,
Clemens

Notice the times of Tommy/DLMartin battle and the 3 battles you have with me.  Mine were earlier.

 

Your description of events ended with Tommy at 8:03:49 and me at 7:44:43.  But I outlived Tommy.   Plus, "Colossus, Hang Log 2" is 181 pages vs. 173 pages for "Colossus, Hang Log."

 

I made "Hang Log" immediately before Tommy withdrew.   I made "Hang Log 2" immediately before DLMartin withdrew.

 

Jeff Matthews

MATTHEWS | EASLEY | CHANEY

13430 Northwest Freeway, Suite 990

Houston, Texas 77040

Ph. (713) 223-4000

Fax (281) 589-9000

 

 

From: Jeff Matthews [mailto:je...@xe...] 
Sent: Monday, December 09, 2013 9:24 AM
To: 'Clemens Katzer'
Cc: col...@li...
Subject: Re: [Colossus-developers] 2nd Hang, same game

 

I think all is correct as you describe, except as to my battle (the 2nd hang):

 

If my client auto-fled, that would be something odd of which I have never heard.  Plus, it was my Titan legion.  

 

The game did not play on after that.  DLMartin withdrew and went to bed.

 

Regarding the first hang, DLMartin attacked TommyDice.   DL and I both were looking at battleboards.  Tommy was not.  Tommy withdrew and went to bed.

 

After Tommy withdrew, I lasted maybe 4 more turns.  I do not recall the number of battles after the withdrew.  It could have been 4, but I seem to doubt it because I was on my last leg and only had 2 legions when Tommy withdrew.

 

I did copy the entire log files.

 

You should have "Colossus, Hang Log" and "Colossus, Hang Log 2."  Notice that the 2nd file is the 2nd hang, and it has 181 pages, whereas the first has 173 pages.

 

The hex on which it hung this final time was DLMartin attacking my titan legion in the Desert (Hex 7).  No battleboard for either of us.  I was not asked to concede or not concede.

 

 

 

Jeff Matthews

MATTHEWS | EASLEY | CHANEY

13430 Northwest Freeway, Suite 990

Houston, Texas 77040

Ph. (713) 223-4000

Fax (281) 589-9000

 

 

-----Original Message-----
From: Clemens Katzer [mailto:cle...@cl...] 
Sent: Monday, December 09, 2013 8:39 AM
To: je...@xe...
Cc: col...@li...
Subject: Re: [Colossus-developers] 2nd Hang, same game

 

 

Hm.

 

 

Looks like this was game 37778 with tommydice, Forthright, dlmartin and you; started ~ 06:30 AM Germany  or, I guess, 11:30 PM (US east coast).

 

You were playing as Jeff M color Orange. Right?

 

Above I had to guess and quiz together. If (in future) you can tell game number and which battles it were, (or, like, Turn 10 dlmartinss turn, land Swamp 123 ) things would be much easier....

 

(and all of you were using rather recent versions, so that rules out my theory of "it's the player with the old version where it hangs." )-:

 

---

 

 

For the first hang, it happened to another player (tommydice, Black ?); for you all worked normally (you got the battle board), right?

So there's not much one can say from your client log. :-(

 

And tommydice quit after approx. 4 minutes, yes.

 

 

12/09-07:59:31: INFO  : Legion Bk05 (Hand) in Brush hex 120 moves to Plains hex 115 entering on LEFT

12/09-07:59:37: INFO  : Phase advances to Fight

12/09-07:59:38: INFO  : A new engagement: Plains hex 110 attacker Bk09 defender Rd01

12/09-07:59:43: INFO  : Bk09 (dlmartin) attacks Rd01 (tommydice) in Plains hex 110

 

12/09-08:03:49: INFO  : Setting isGone to true in CH for 'tommydice' 

(reason: EOF on channel)

12/09-08:03:49: INFO  : Before EOF processing, calling processByteBuffer to handle the 12 bytes that were read before.

12/09-08:03:49: INFO  : Setting isGone to true in CH for 'tommydice' 

(reason: received explit 'disconnect' request from client)

12/09-08:03:49: INFO  : Received explicit 'disconnect' request from Client tommydice - calling 'withdrawIfNeeded'.

12/09-08:03:49: FINE  : Client disconnected without explicit withdraw - doing automatic withdraw (if needed) for player tommydice

12/09-08:03:49: INFO  : Withdrawal for processing client tommydice requested.

 

 

 

> Few turns later, I was attacked. The dialog did not appear to ask me 

> to flee or fight. Neither player received a battleboard.

 

According to server log, Orange was attacked three times and server log says he fled after 3-4 seconds.

 

But you say you didn't get any dialog. So, did your client "automatically" answered for you ?

 

> The dialog did not appear to ask me to flee or fight.

 

So how did the thing continue ? Did game continue just "same as if" 

battle had been resolved ?

In that case, it must be something like "the GUI responding for you with "flee" without asking you.

(for which it would be weird, that it takes 3-4 seconds...)

 

 

Otherwise, what did any of you do to get out of that situation (did you 

or somebody concede that engagement, or totally quit the game, ... ) ?

 

 

12/09-07:24:01: INFO  : A new engagement: DesertAlt hex 107 attacker 

Bk10 defender Or03

12/09-07:24:05: INFO  : Legion Or03 flees from legion Bk10

12/09-07:24:05: INFO  : dlmartin earns 21.0 half-points (116.0 + 21.0 

=> 137.0)

12/09-07:24:05: INFO  : Legion Or03[Cyclops, Gargoyle, Gargoyle] is 

eliminated

12/09-07:24:05: INFO  : Battle completed, result: winner is Bk10 

(player dlmartin)

12/09-07:24:05: INFO  : No more engagements to resolve - advanding 

phase.

 

 

12/09-07:39:40: INFO  : A new engagement: SwampAlt hex 142 attacker 

Bk10 defender Or01

12/09-07:39:43: INFO  : Legion Or01 flees from legion Bk10

12/09-07:39:43: INFO  : dlmartin earns 15.0 half-points (137.0 + 15.0 

=> 152.0)

12/09-07:39:43: INFO  : Legion Or01[Ranger, Lion] is eliminated

12/09-07:39:43: INFO  : Battle completed, result: winner is Bk10 

(player dlmartin)

12/09-07:39:43: INFO  : No more engagements to resolve - advanding 

phase.

 

 

12/09-07:44:40: INFO  : A new engagement: Desert hex 7 attacker Bk06 

defender Or02

12/09-07:44:43: INFO  : Legion Or02 flees from legion Bk06

12/09-07:44:43: INFO  : dlmartin earns 18.0 half-points (152.0 + 18.0 

=> 170.0)

12/09-07:44:43: INFO  : Legion Or02[Ogre, Ogre, Ogre] is eliminated

12/09-07:44:43: INFO  : Battle completed, result: winner is Bk06 

(player dlmartin)

 

 

BTW, there were 4 or so more engagements in your log after the one 

where your Or02 was attacked. Was that so ?

(= so you did not copy-paste from the log window 2nd time right at the 

point where the "no dialog" thing happened).

 

 

BR,

Clemens

 

 

 

On 2013-12-09 09:13, Jeff Matthews wrote:

> At first hang (see previous email), defender quit. Game went on.

> 

> Few turns later, I was attacked. The dialog did not appear to ask me

> to flee or fight. Neither player received a battleboard.

> 

> The complete log (from game beginning) is attached.

> 

> Jeff Matthews

> 

> MATTHEWS | EASLEY | CHANEY

> 

> 13430 Northwest Freeway, Suite 990

> 

> Houston, Texas 77040

> 

> Ph. (713) 223-4000

> 

> Fax (281) 589-9000

 

I think all is correct as you describe, except as to my battle (the 2nd hang):

 

If my client auto-fled, that would be something odd of which I have never heard.  Plus, it was my Titan legion.  

 

The game did not play on after that.  DLMartin withdrew and went to bed.

 

Regarding the first hang, DLMartin attacked TommyDice.   DL and I both were looking at battleboards.  Tommy was not.  Tommy withdrew and went to bed.

 

After Tommy withdrew, I lasted maybe 4 more turns.  I do not recall the number of battles after the withdrew.  It could have been 4, but I seem to doubt it because I was on my last leg and only had 2 legions when Tommy withdrew.

 

I did copy the entire log files.

 

You should have "Colossus, Hang Log" and "Colossus, Hang Log 2."  Notice that the 2nd file is the 2nd hang, and it has 181 pages, whereas the first has 173 pages.

 

The hex on which it hung this final time was DLMartin attacking my titan legion in the Desert (Hex 7).  No battleboard for either of us.  I was not asked to concede or not concede.

 

 

 

Jeff Matthews

MATTHEWS | EASLEY | CHANEY

13430 Northwest Freeway, Suite 990

Houston, Texas 77040

Ph. (713) 223-4000

Fax (281) 589-9000

 

 

-----Original Message-----
From: Clemens Katzer [mailto:cle...@cl...] 
Sent: Monday, December 09, 2013 8:39 AM
To: je...@xe...
Cc: col...@li...
Subject: Re: [Colossus-developers] 2nd Hang, same game

 

 

Hm.

 

 

Looks like this was game 37778 with tommydice, Forthright, dlmartin and you; started ~ 06:30 AM Germany  or, I guess, 11:30 PM (US east coast).

 

You were playing as Jeff M color Orange. Right?

 

Above I had to guess and quiz together. If (in future) you can tell game number and which battles it were, (or, like, Turn 10 dlmartinss turn, land Swamp 123 ) things would be much easier....

 

(and all of you were using rather recent versions, so that rules out my theory of "it's the player with the old version where it hangs." )-:

 

---

 

 

For the first hang, it happened to another player (tommydice, Black ?); for you all worked normally (you got the battle board), right?

So there's not much one can say from your client log. :-(

 

And tommydice quit after approx. 4 minutes, yes.

 

 

12/09-07:59:31: INFO  : Legion Bk05 (Hand) in Brush hex 120 moves to Plains hex 115 entering on LEFT

12/09-07:59:37: INFO  : Phase advances to Fight

12/09-07:59:38: INFO  : A new engagement: Plains hex 110 attacker Bk09 defender Rd01

12/09-07:59:43: INFO  : Bk09 (dlmartin) attacks Rd01 (tommydice) in Plains hex 110

 

12/09-08:03:49: INFO  : Setting isGone to true in CH for 'tommydice' 

(reason: EOF on channel)

12/09-08:03:49: INFO  : Before EOF processing, calling processByteBuffer to handle the 12 bytes that were read before.

12/09-08:03:49: INFO  : Setting isGone to true in CH for 'tommydice' 

(reason: received explit 'disconnect' request from client)

12/09-08:03:49: INFO  : Received explicit 'disconnect' request from Client tommydice - calling 'withdrawIfNeeded'.

12/09-08:03:49: FINE  : Client disconnected without explicit withdraw - doing automatic withdraw (if needed) for player tommydice

12/09-08:03:49: INFO  : Withdrawal for processing client tommydice requested.

 

 

 

> Few turns later, I was attacked. The dialog did not appear to ask me 

> to flee or fight. Neither player received a battleboard.

 

According to server log, Orange was attacked three times and server log says he fled after 3-4 seconds.

 

But you say you didn't get any dialog. So, did your client "automatically" answered for you ?

 

> The dialog did not appear to ask me to flee or fight.

 

So how did the thing continue ? Did game continue just "same as if" 

battle had been resolved ?

In that case, it must be something like "the GUI responding for you with "flee" without asking you.

(for which it would be weird, that it takes 3-4 seconds...)

 

 

Otherwise, what did any of you do to get out of that situation (did you 

or somebody concede that engagement, or totally quit the game, ... ) ?

 

 

12/09-07:24:01: INFO  : A new engagement: DesertAlt hex 107 attacker 

Bk10 defender Or03

12/09-07:24:05: INFO  : Legion Or03 flees from legion Bk10

12/09-07:24:05: INFO  : dlmartin earns 21.0 half-points (116.0 + 21.0 

=> 137.0)

12/09-07:24:05: INFO  : Legion Or03[Cyclops, Gargoyle, Gargoyle] is 

eliminated

12/09-07:24:05: INFO  : Battle completed, result: winner is Bk10 

(player dlmartin)

12/09-07:24:05: INFO  : No more engagements to resolve - advanding 

phase.

 

 

12/09-07:39:40: INFO  : A new engagement: SwampAlt hex 142 attacker 

Bk10 defender Or01

12/09-07:39:43: INFO  : Legion Or01 flees from legion Bk10

12/09-07:39:43: INFO  : dlmartin earns 15.0 half-points (137.0 + 15.0 

=> 152.0)

12/09-07:39:43: INFO  : Legion Or01[Ranger, Lion] is eliminated

12/09-07:39:43: INFO  : Battle completed, result: winner is Bk10 

(player dlmartin)

12/09-07:39:43: INFO  : No more engagements to resolve - advanding 

phase.

 

 

12/09-07:44:40: INFO  : A new engagement: Desert hex 7 attacker Bk06 

defender Or02

12/09-07:44:43: INFO  : Legion Or02 flees from legion Bk06

12/09-07:44:43: INFO  : dlmartin earns 18.0 half-points (152.0 + 18.0 

=> 170.0)

12/09-07:44:43: INFO  : Legion Or02[Ogre, Ogre, Ogre] is eliminated

12/09-07:44:43: INFO  : Battle completed, result: winner is Bk06 

(player dlmartin)

 

 

BTW, there were 4 or so more engagements in your log after the one 

where your Or02 was attacked. Was that so ?

(= so you did not copy-paste from the log window 2nd time right at the 

point where the "no dialog" thing happened).

 

 

BR,

Clemens

 

 

 

On 2013-12-09 09:13, Jeff Matthews wrote:

> At first hang (see previous email), defender quit. Game went on.

> 

> Few turns later, I was attacked. The dialog did not appear to ask me

> to flee or fight. Neither player received a battleboard.

> 

> The complete log (from game beginning) is attached.

> 

> Jeff Matthews

> 

> MATTHEWS | EASLEY | CHANEY

> 

> 13430 Northwest Freeway, Suite 990

> 

> Houston, Texas 77040

> 

> Ph. (713) 223-4000

> 

> Fax (281) 589-9000

 

Hm.


Looks like this was game 37778 with tommydice, Forthright, dlmartin and 
you; started ~ 06:30 AM Germany  or, I guess, 11:30 PM (US east coast).

You were playing as Jeff M color Orange. Right?

Above I had to guess and quiz together. If (in future) you can tell 
game number and which battles it were, (or, like, Turn 10 dlmartinss 
turn, land Swamp 123 ) things would be much easier....

(and all of you were using rather recent versions, so that rules out my 
theory of "it's the player with the old version where it hangs." )-:

---


For the first hang, it happened to another player (tommydice, Black ?); 
for you all worked normally (you got the battle board), right?
So there's not much one can say from your client log. :-(

And tommydice quit after approx. 4 minutes, yes.


12/09-07:59:31: INFO  : Legion Bk05 (Hand) in Brush hex 120 moves to 
Plains hex 115 entering on LEFT
12/09-07:59:37: INFO  : Phase advances to Fight
12/09-07:59:38: INFO  : A new engagement: Plains hex 110 attacker Bk09 
defender Rd01
12/09-07:59:43: INFO  : Bk09 (dlmartin) attacks Rd01 (tommydice) in 
Plains hex 110

12/09-08:03:49: INFO  : Setting isGone to true in CH for 'tommydice' 
(reason: EOF on channel)
12/09-08:03:49: INFO  : Before EOF processing, calling 
processByteBuffer to handle the 12 bytes that were read before.
12/09-08:03:49: INFO  : Setting isGone to true in CH for 'tommydice' 
(reason: received explit 'disconnect' request from client)
12/09-08:03:49: INFO  : Received explicit 'disconnect' request from 
Client tommydice - calling 'withdrawIfNeeded'.
12/09-08:03:49: FINE  : Client disconnected without explicit withdraw - 
doing automatic withdraw (if needed) for player tommydice
12/09-08:03:49: INFO  : Withdrawal for processing client tommydice 
requested.



> Few turns later, I was attacked. The dialog did not appear to ask me
> to flee or fight. Neither player received a battleboard.

According to server log, Orange was attacked three times and server log 
says he fled after 3-4 seconds.

But you say you didn't get any dialog. So, did your client 
"automatically" answered for you ?

> The dialog did not appear to ask me to flee or fight.

So how did the thing continue ? Did game continue just "same as if" 
battle had been resolved ?
In that case, it must be something like "the GUI responding for you 
with "flee" without asking you.
(for which it would be weird, that it takes 3-4 seconds...)


Otherwise, what did any of you do to get out of that situation (did you 
or somebody concede that engagement, or totally quit the game, ... ) ?


12/09-07:24:01: INFO  : A new engagement: DesertAlt hex 107 attacker 
Bk10 defender Or03
12/09-07:24:05: INFO  : Legion Or03 flees from legion Bk10
12/09-07:24:05: INFO  : dlmartin earns 21.0 half-points (116.0 + 21.0 
=> 137.0)
12/09-07:24:05: INFO  : Legion Or03[Cyclops, Gargoyle, Gargoyle] is 
eliminated
12/09-07:24:05: INFO  : Battle completed, result: winner is Bk10 
(player dlmartin)
12/09-07:24:05: INFO  : No more engagements to resolve - advanding 
phase.


12/09-07:39:40: INFO  : A new engagement: SwampAlt hex 142 attacker 
Bk10 defender Or01
12/09-07:39:43: INFO  : Legion Or01 flees from legion Bk10
12/09-07:39:43: INFO  : dlmartin earns 15.0 half-points (137.0 + 15.0 
=> 152.0)
12/09-07:39:43: INFO  : Legion Or01[Ranger, Lion] is eliminated
12/09-07:39:43: INFO  : Battle completed, result: winner is Bk10 
(player dlmartin)
12/09-07:39:43: INFO  : No more engagements to resolve - advanding 
phase.


12/09-07:44:40: INFO  : A new engagement: Desert hex 7 attacker Bk06 
defender Or02
12/09-07:44:43: INFO  : Legion Or02 flees from legion Bk06
12/09-07:44:43: INFO  : dlmartin earns 18.0 half-points (152.0 + 18.0 
=> 170.0)
12/09-07:44:43: INFO  : Legion Or02[Ogre, Ogre, Ogre] is eliminated
12/09-07:44:43: INFO  : Battle completed, result: winner is Bk06 
(player dlmartin)


BTW, there were 4 or so more engagements in your log after the one 
where your Or02 was attacked. Was that so ?
(= so you did not copy-paste from the log window 2nd time right at the 
point where the "no dialog" thing happened).


BR,
Clemens



On 2013-12-09 09:13, Jeff Matthews wrote:
> At first hang (see previous email), defender quit. Game went on.
>
> Few turns later, I was attacked. The dialog did not appear to ask me
> to flee or fight. Neither player received a battleboard.
>
> The complete log (from game beginning) is attached.
>
> Jeff Matthews
>
> MATTHEWS | EASLEY | CHANEY
>
> 13430 Northwest Freeway, Suite 990
>
> Houston, Texas 77040
>
> Ph. (713) 223-4000
>
> Fax (281) 589-9000

> on the colossus-checkin mailing list.

Correction: It's about the colossus colossus-developers mailing list, 
of course.

BR, C.


On 2013-12-05 10:20, Clemens Katzer wrote:
> Hi all,
>
> I've now got several bounces for mails sourcefourge has sent to 
> members
> on the colossus-checkin mailing list.
>
>  From the bounce it's not clear who of the subscribers it is.
>
> If anything of the following is familiar to you, please contact me:
>
>
> <somebody>@jmso.lv
> Banco do Brasil <somebody2>@bb.com
> chavesse.....
>
>
> Thx,
> Clemens
>
>
> 
> ------------------------------------------------------------------------------
> Sponsored by Intel(R) XDK
> Develop, test and display web and hybrid apps with a single code 
> base.
> Download it for free now!
> 
> http://pubads.g.doubleclick.net/gampad/clk?id=111408631&iu=/4140/ostg.clktrk
> _______________________________________________
> Colossus-developers mailing list
> Col...@li...
> https://lists.sourceforge.net/lists/listinfo/colossus-developers

Hi all,

I've now got several bounces for mails sourcefourge has sent to members 
on the colossus-checkin mailing list.

 From the bounce it's not clear who of the subscribers it is.

If anything of the following is familiar to you, please contact me:


<somebody>@jmso.lv
Banco do Brasil <somebody2>@bb.com
chavesse.....


Thx,
Clemens

yes, I know. One creates a key, create a signing request and gets a 
signed key back or something....

Point being, in the whole procedure I haven't seen any "send it 
somewhere". (whatever one sends).

Thx,
Clemens

On 2013-11-12 18:31, Bruno Wolff III wrote:
> On Tue, Nov 12, 2013 at 17:26:38 +0200,
>    Clemens Katzer <cle...@cl...> wrote:
>>
>>>The CA just signs the public key.
>>
>>of course. But roughly looking over it, I didn't see any "send it to
>>them and they sign it" step.
>
> At work we use comodo to sign our keys and for certifcates we send
> the CSR (typically, though there are other options) which doesn't 
> include
> the private key.
>
> 
> ------------------------------------------------------------------------------
> November Webinars for C, C++, Fortran Developers
> Accelerate application performance with scalable programming models. 
> Explore
> techniques for threading, error checking, porting, and tuning. Get 
> the most
> from the latest Intel processors and coprocessors. See abstracts and 
> register
> 
> http://pubads.g.doubleclick.net/gampad/clk?id=60136231&iu=/4140/ostg.clktrk
> _______________________________________________
> Colossus-developers mailing list
> Col...@li...
> https://lists.sourceforge.net/lists/listinfo/colossus-developers

On Tue, Nov 12, 2013 at 17:26:38 +0200,
   Clemens Katzer <cle...@cl...> wrote:
>
>>The CA just signs the public key.
>
>of course. But roughly looking over it, I didn't see any "send it to 
>them and they sign it" step.

At work we use comodo to sign our keys and for certifcates we send 
the CSR (typically, though there are other options) which doesn't include 
the private key.