|
Related Topics:
| Dead. - >.>
Hello! Huge Question... LIS MORRIS SPACESHIP REPAIR TOOL - Hi all! This newsgroup seems to be dead... lets have a try there is this nice spaceship repair tool developed by lis morris. i found out that at least one feature is missing in this it cant restore a killed lis..
creatures 3d - to ANYONE who is still posting here, or even this website, I have had it with waiting for a creatures 3d to come out, so I have decided to make my own, but I can't do it on my own, so I am looking for anyone who is and has had..
Hellooo...? - *walks in* *watches the for a bit* Anyone here? If there is anyone here, am I the only one still playing C1...? Nutter
Hey! The old place is hopping again! =) - Come on DOWN, you lurkers from beyond! Lets *giggles*
|
|
| Author |
Message |
External

Since: Dec 26, 2005 Posts: 5
|
(Msg. 1) Posted: Tue Dec 27, 2005 1:34 pm
Post subject: [oc2e] Serialization? Archived from groups: alt>games>creatures (more info?)
|
|
|
-----BEGIN PGP SIGNED MESSAGE-----
Hash: RIPEMD160
At some point openc2e will clearly have to support serialization. But
what will the on-disk format look like? It's about that time that we
ought to start thinking of such things.
I'm in favor of a .zip or .jar like format (like opendocument and etc
use) with a YAML or XML datastructure for the world data, and genetics,
etc packed into the .zip, but I could be convinced otherwise. The main
reason I'm pushing for a standard format is so we can add fields later
(oc2e is still under heavy development after all ...) without breaking
save compatibility, and also to help with debugging.
Floating point serialization will be tricky though - do endianness
issues crop up with that? It may be possible to just write a hex string
and read it back in at the destination ...
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.2 (GNU/Linux)
iQEVAwUBQ7GJR1yQ4Ir9vYFzAQNLmgf/SvxCBiGWCvwdDNvM9i4Ke48UqsM11eze
mK4vEnYn0znOz/c6X/6V8lja9yZibYlvOaRQ+nBpsj8SG1HDZUgGVXz8QGMLUMBe
fA8wcLbdSD3lLXDXkFEbXELZiJd6r0M4FvWuQRehjmA/DW5XPDX+ZEQSfQpOuyOc
TC71hAwzQ6SnzWLAm637NcxiK4+c+o12g80QsaDOlZ4nXa7h7Gz5Qn3oriEUdRco
6U8GuKknWWTLuY8jgVxLv6b2sUITyoDzt4xK9ByP+vLFO1UIOFFBdQeHTch8uF4A
/HzPxBkEkY7IhnbBNEI466TyxeQ59qZyeD0Mil0JJASE08m+yNQTPQ==
=K0AZ
-----END PGP SIGNATURE----- >> Stay informed about: [oc2e] Serialization? |
|
| Back to top |
|
 |  |
External

Since: Jun 16, 2005 Posts: 22
|
(Msg. 2) Posted: Tue Dec 27, 2005 2:55 pm
Post subject: Re: [oc2e] Serialization? [Login to view extended thread Info.] Archived from groups: per prev. post (more info?)
|
|
|
On Tue, 27 Dec 2005 13:34:50 -0500, bd_ wrote:
> At some point openc2e will clearly have to support serialization. But
> what will the on-disk format look like? It's about that time that we
> ought to start thinking of such things.
I am all in favour of taking the lazy way out and using
boost::serialization. It's sooo easy, and it's not that hard to add/change
fields between versions while maintaining compatibility (although it
does require incrementing the version when you change things and adding
if statements, but we're going to end up doing that anyway, and we can
always just scrap support for too-old versions at intervals). And it does
XML!
It's definitely not as easy as it sounds to implement something yourself,
you have to handle all kinds of issues, eg the most obivous being pointers
(including pointer loops).
> I'm in favor of a .zip or .jar like format (like opendocument and etc
> use) with a YAML or XML datastructure for the world data, and genetics,
> etc packed into the .zip, but I could be convinced otherwise. The main
> reason I'm pushing for a standard format is so we can add fields later
> (oc2e is still under heavy development after all ...) without breaking
> save compatibility, and also to help with debugging.
Yes, storing stuff in a zip file might be nice.. but I'm not sure if it
has real advantages over one big xml file.
> Floating point serialization will be tricky though - do endianness
> issues crop up with that? It may be possible to just write a hex string
> and read it back in at the destination ...
There are varied solutions for handling endianism, without having to
resort to horrible hacks like dumping hex strings.  >> Stay informed about: [oc2e] Serialization? |
|
| Back to top |
|
 |  |
External

Since: Dec 26, 2005 Posts: 5
|
(Msg. 3) Posted: Tue Dec 27, 2005 2:55 pm
Post subject: Re: [oc2e] Serialization? [Login to view extended thread Info.] Archived from groups: per prev. post (more info?)
|
|
|
-----BEGIN PGP SIGNED MESSAGE-----
Hash: RIPEMD160
On 2005-12-27, Alyssa Milburn <fuzzie.DeleteThis@warpedgames.com> wrote:
> On Tue, 27 Dec 2005 13:34:50 -0500, bd_ wrote:
>> At some point openc2e will clearly have to support serialization. But
>> what will the on-disk format look like? It's about that time that we
>> ought to start thinking of such things.
>
> I am all in favour of taking the lazy way out and using
> boost::serialization. It's sooo easy, and it's not that hard to add/change
> fields between versions while maintaining compatibility (although it
> does require incrementing the version when you change things and adding
> if statements, but we're going to end up doing that anyway, and we can
> always just scrap support for too-old versions at intervals). And it does
> XML!
>
> It's definitely not as easy as it sounds to implement something yourself,
> you have to handle all kinds of issues, eg the most obivous being pointers
> (including pointer loops).
Hm, I'll have to look into boost::serialization. I wouldn't expect too
many pointer loops if we make use of UNID exclusively...
>
>> I'm in favor of a .zip or .jar like format (like opendocument and etc
>> use) with a YAML or XML datastructure for the world data, and genetics,
>> etc packed into the .zip, but I could be convinced otherwise. The main
>> reason I'm pushing for a standard format is so we can add fields later
>> (oc2e is still under heavy development after all ...) without breaking
>> save compatibility, and also to help with debugging.
>
> Yes, storing stuff in a zip file might be nice.. but I'm not sure if it
> has real advantages over one big xml file.
Well, for genetics and etc. Though come to think of it c2e has them
unpacked ...
>
>> Floating point serialization will be tricky though - do endianness
>> issues crop up with that? It may be possible to just write a hex string
>> and read it back in at the destination ...
>
> There are varied solutions for handling endianism, without having to
> resort to horrible hacks like dumping hex strings.
I was more thinking of fp accuracy - if we can dump the exponent and
manitessa (sp?) seperately as integers that would be ideal.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.2 (GNU/Linux)
iQEVAwUBQ7GL/1yQ4Ir9vYFzAQOLHQgAgi8XRVNosBDWKaw/Qql6wUdwh9vg3MqK
gqD+Xdpt5LEZDKB1ZkdgWY+QAszTXmTyLSrinT1YN/xVYldY173T1tryC1ZsZopw
PeDFG08oELMnXY+YAYS1C+2eaYUQMVFjFfhXkNaYkwrSwAiY5bpkvJjJ9ordrd0B
L5kYNPxDXxk8cIWbEmGHkuQyfWWnsoQdOnTi0j5i1IvD3Kd8lv9hAXS2q3xm7sRa
aTWM/cLY4G5MKAFQchXtVcR0G+NnkeKKCQMa7aSmmdz9DZM26n7a5UTC5yE/mcuO
e/YCDL15a9VQCc2ef2hUwKPLOjrtWUWR5F6clpCJ+XMKovnk1JDH+Q==
=Anl3
-----END PGP SIGNATURE----- >> Stay informed about: [oc2e] Serialization? |
|
| Back to top |
|
 |  |
External

Since: Jun 16, 2005 Posts: 22
|
(Msg. 4) Posted: Tue Dec 27, 2005 2:55 pm
Post subject: Re: [oc2e] Serialization? [Login to view extended thread Info.] Archived from groups: per prev. post (more info?)
|
|
|
On Tue, 27 Dec 2005 18:46:59 +0000, bd_ wrote:
> Hm, I'll have to look into boost::serialization. I wouldn't expect too
> many pointer loops if we make use of UNID exclusively...
No, but it's still something to bear in mind, and my limited experience
with boost::serialization has shown that I haven't actually had to care
about it, although I have no idea how it would interact with AgentRef etc.
The other serialization libraries out there seem to be inferior .. such as
Kinnison's vaguely-c2e-like (which might be useful for other
reverse-engineering people to note) serialization code which lives in
CommonC++ nowadays.
> Well, for genetics and etc. Though come to think of it c2e has them
> unpacked ...
Well, we'll want to use PRAY for the exported creatures, to maintain
compatibility with the existing in-world agents, I should imagine..
It might be nice to have some kind of external tool which bundled up the
serialized world with the necessary extra resources (eg, genomes) which it
uses, into a zip file, for online distribution and the such, but I don't
think it belongs in the engine, and I'm not sure there's much point at all
to it (but then, this is somewhat the point of using agc for discussion,
to see if anyone pipes up and says "yes, I want that!", I guess).
> I was more thinking of fp accuracy - if we can dump the exponent and
> manitessa (sp?) seperately as integers that would be ideal.
I think my point is more "this isn't even vaguely important because all
the existing serialization libraries handle it fine and we can just
snaffle it from one of them". >> Stay informed about: [oc2e] Serialization? |
|
| Back to top |
|
 |  |
External

Since: Dec 26, 2005 Posts: 5
|
(Msg. 5) Posted: Tue Dec 27, 2005 2:55 pm
Post subject: Re: [oc2e] Serialization? [Login to view extended thread Info.] Archived from groups: per prev. post (more info?)
|
|
|
-----BEGIN PGP SIGNED MESSAGE-----
Hash: RIPEMD160
On 2005-12-27, Alyssa Milburn <fuzzie RemoveThis @warpedgames.com> wrote:
> On Tue, 27 Dec 2005 18:46:59 +0000, bd_ wrote:
>> Hm, I'll have to look into boost::serialization. I wouldn't expect too
>> many pointer loops if we make use of UNID exclusively...
>
> No, but it's still something to bear in mind, and my limited experience
> with boost::serialization has shown that I haven't actually had to care
> about it, although I have no idea how it would interact with AgentRef etc.
With AgentRef we can just delegate to the actual agent, I think.
> The other serialization libraries out there seem to be inferior .. such as
> Kinnison's vaguely-c2e-like (which might be useful for other
> reverse-engineering people to note) serialization code which lives in
> CommonC++ nowadays.
>
>
>> Well, for genetics and etc. Though come to think of it c2e has them
>> unpacked ...
>
> Well, we'll want to use PRAY for the exported creatures, to maintain
> compatibility with the existing in-world agents, I should imagine..
>
> It might be nice to have some kind of external tool which bundled up the
> serialized world with the necessary extra resources (eg, genomes) which it
> uses, into a zip file, for online distribution and the such, but I don't
> think it belongs in the engine, and I'm not sure there's much point at all
> to it (but then, this is somewhat the point of using agc for discussion,
> to see if anyone pipes up and says "yes, I want that!", I guess).
>
>> I was more thinking of fp accuracy - if we can dump the exponent and
>> manitessa (sp?) seperately as integers that would be ideal.
>
> I think my point is more "this isn't even vaguely important because all
> the existing serialization libraries handle it fine and we can just
> snaffle it from one of them".
Fair enough. I'll go have a look at the boost docs for now, see how
compatibility works, exactly.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.2 (GNU/Linux)
iQEVAwUBQ7GO41yQ4Ir9vYFzAQM9PQf/bmd15Re4syl/6ivBbl9D8jbji0YIi8xj
tY/dttMisn2+wTBVpJfIrQ09ev5NnnLyUL42YHEX5JZVMsYiNjbky4nA41ab98we
rQT28nYI2J9VH3TFcleg+hgM9ogzfs4qSA1GFinqup9wxxaQvpBg0HZ93qS++yYC
fRFHN5xeUlWIRtYDT/YR/Hw9DqiIq/icotJO5hin+Bhl9GfV2bAS2AzqUf04V1Lq
CybdwqKzJHfhuu5uBs721wFjR09znYnX1xD6Lq3/nHaCharbY1QOOT93+45axtwL
uMCbnQbNL6YFDeRyiLYjNBN6N5HxdYlInGHloKYwq6lYXjJV51Y/PQ==
=r6LH
-----END PGP SIGNATURE----- >> Stay informed about: [oc2e] Serialization? |
|
| Back to top |
|
 |  |
External

Since: Dec 27, 2005 Posts: 3
|
(Msg. 6) Posted: Thu Dec 29, 2005 7:55 am
Post subject: Re: [oc2e] Serialization? [Login to view extended thread Info.] Archived from groups: per prev. post (more info?)
|
|
|
On 2005-12-27, Alyssa Milburn <fuzzie.TakeThisOut@warpedgames.com> wrote:
>
> [snip]
>
> Yes, storing stuff in a zip file might be nice.. but I'm not sure if it
> has real advantages over one big xml file.
>
Looks nicer when you pick it up from outside the engine.
--
- nornagon >> Stay informed about: [oc2e] Serialization? |
|
| Back to top |
|
 |  |
External

Since: Sep 28, 2005 Posts: 50
|
(Msg. 7) Posted: Thu Dec 29, 2005 8:55 am
Post subject: Re: [oc2e] Serialization? [Login to view extended thread Info.] Archived from groups: per prev. post (more info?)
|
|
|
|
|
| Back to top |
|
 |  |
External

Since: Dec 26, 2005 Posts: 6
|
(Msg. 8) Posted: Thu Dec 29, 2005 8:06 pm
Post subject: Re: [oc2e] Serialization? [Login to view extended thread Info.] Archived from groups: per prev. post (more info?)
|
|
|
-----BEGIN PGP SIGNED MESSAGE-----
Hash: RIPEMD160
On 2005-12-29, Jeremy Apthorp <nornagon.DeleteThis@localhost.localdomain> wrote:
> On 2005-12-27, Alyssa Milburn <fuzzie.DeleteThis@warpedgames.com> wrote:
>>
>> [snip]
>>
>> Yes, storing stuff in a zip file might be nice.. but I'm not sure if it
>> has real advantages over one big xml file.
>>
>
> Looks nicer when you pick it up from outside the engine.
>
Howso? And, I was actually thinking of using boost::serialization's
builtin pseudo-text format (ASCII, but some unreadable internal format).
It's portable between systems, and doesn't demand that I scatter ugly
BOOST_SERIALIZATION_NVP() macros everywhere
PS, your from: header's a bit wrong
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.2 (GNU/Linux)
iQEVAwUBQ7SIF1yQ4Ir9vYFzAQPrEwgAtnz0vwSrsOLkL/N3Y1HcEZey3v3XXyEm
8KkjLpYsVrBKbaY05cNlA1OM59w4HRD+iTytfbMLKhK4wUz5e9V/LfePxO0SFgb1
G+ACrzldwxNmpcanBrLHXsVoRM1kI4/zSo+Okvpuhp10pomvsNnXPbYFJW0+jNYb
DG02DEYaWqnlmFPX2IMHDp1dln4dkluKNjIK17KVdUnZMX4SaVwlMf+QvoM6ptnc
G3zOErP5so5bMOYL2U/x4DzW52rdg0HtD5dWKhczyuZw+CebN2tlOWVCwLy21X8S
QtEN36mfL1ntXARZm7QGQFlyY45b1eYDLk60GhxdRuzycRQVDLYM/A==
=zbp7
-----END PGP SIGNATURE----- >> Stay informed about: [oc2e] Serialization? |
|
| Back to top |
|
 |  |
External

Since: Dec 30, 2005 Posts: 1
|
(Msg. 9) Posted: Thu Dec 29, 2005 8:55 pm
Post subject: Re: [oc2e] Serialization? [Login to view extended thread Info.] Archived from groups: per prev. post (more info?)
|
|
|
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
bd_ wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: RIPEMD160
>
> At some point openc2e will clearly have to support serialization. But
> what will the on-disk format look like? It's about that time that we
> ought to start thinking of such things.
I vote for XML.
Whatever is chosen, please avoid binary stuff, as it can be horribly painful
in some languages. Say, reading binary structures in Perl is rather icky,
and while VB has some fairly decent support for it, it has some really
annoying problems.
Besides, XML is human editable, which is a nice thing for development.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.2 (GNU/Linux)
iD8DBQFDtH93vCkUtBccqkoRAjFAAKCVNXO4mOOTYxHUjdCH9Tb2/bvLSQCgv86w
iFVVGLx0vV8BHlrk3Dv4J5I=
=+flj
-----END PGP SIGNATURE----- >> Stay informed about: [oc2e] Serialization? |
|
| Back to top |
|
 |  |
External

Since: Jun 16, 2005 Posts: 22
|
(Msg. 10) Posted: Thu Dec 29, 2005 8:55 pm
Post subject: Re: [oc2e] Serialization? [Login to view extended thread Info.] Archived from groups: per prev. post (more info?)
|
|
|
On Fri, 30 Dec 2005 01:29:38 +0100, Vadim wrote:
> Whatever is chosen, please avoid binary stuff, as it can be horribly painful
> in some languages. Say, reading binary structures in Perl is rather icky,
> and while VB has some fairly decent support for it, it has some really
> annoying problems.
The thing is, things shouldn't be messing with the serialized data anyway.
And we're stuck with the binary formats for the existing c2e stuff. And
XML is a lot slower to output/parse than binary data.
We should be sure to dump out some kind of basic information with exported
worlds/creatures to make tools able to grab the kind of basic information
they want, but I think that's going to be PRAY-type..
> Besides, XML is human editable, which is a nice thing for development.
That *is* a reason to go for XML, during development, yes. Which is why
boost::serialization is nice, we can switch it between the two modes. >> Stay informed about: [oc2e] Serialization? |
|
| Back to top |
|
 |  |
External

Since: Jun 16, 2005 Posts: 22
|
(Msg. 11) Posted: Thu Dec 29, 2005 8:55 pm
Post subject: Re: [oc2e] Serialization? [Login to view extended thread Info.] Archived from groups: per prev. post (more info?)
|
|
|
On Thu, 29 Dec 2005 20:06:31 -0500, bd_ wrote:
> Howso? And, I was actually thinking of using boost::serialization's
> builtin pseudo-text format (ASCII, but some unreadable internal format).
> It's portable between systems, and doesn't demand that I scatter ugly
> BOOST_SERIALIZATION_NVP() macros everywhere
But, but, readability when the world goes horribly wrong! >> Stay informed about: [oc2e] Serialization? |
|
| Back to top |
|
 |  |
External

Since: Dec 26, 2005 Posts: 6
|
(Msg. 12) Posted: Thu Dec 29, 2005 10:23 pm
Post subject: Re: [oc2e] Serialization? [Login to view extended thread Info.] Archived from groups: per prev. post (more info?)
|
|
|
-----BEGIN PGP SIGNED MESSAGE-----
Hash: RIPEMD160
On 2005-12-30, Alyssa Milburn <fuzzie.TakeThisOut@warpedgames.com> wrote:
> On Thu, 29 Dec 2005 20:06:31 -0500, bd_ wrote:
>> Howso? And, I was actually thinking of using boost::serialization's
>> builtin pseudo-text format (ASCII, but some unreadable internal format).
>> It's portable between systems, and doesn't demand that I scatter ugly
>> BOOST_SERIALIZATION_NVP() macros everywhere
>
> But, but, readability when the world goes horribly wrong!
But, but, gdb!
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.2 (GNU/Linux)
iQEVAwUBQ7SoLlyQ4Ir9vYFzAQP4ZAf5Af/Mx5UwAzFc7RiPPTFS6DJCSg3tgAjL
bzI0DU06FK2xNpp0XAVy3WmhqSIZj3ZAy7WBnbC8ioWBCj5DZIfatnUTsjJ5OuQs
EBD6pNcKec4g6t85DmaifuXYZmASO0k0P0/5fG3DQYuT4bJwu43hogku6Proswkg
1FRYB0yWcbsA1qXm/oI7RNTEXJ2heR5jvzDpIjAXb6i0E7mglFHseR4Bb7/YsgBt
L71lshr6fDDWGbB4z375fdaEaZPbtN+YJTI/zKMDQo+M2JmPXh/iG82zzp+ilGuH
mXtRk9ms/eUHgdu42vuLgNR7FqM9PSktV/lOYVoyATrGO465aptgpA==
=hpOi
-----END PGP SIGNATURE----- >> Stay informed about: [oc2e] Serialization? |
|
| Back to top |
|
 |  |
External

Since: Jun 16, 2005 Posts: 22
|
(Msg. 13) Posted: Mon Jan 09, 2006 9:55 am
Post subject: Re: [oc2e] Serialization? [Login to view extended thread Info.] Archived from groups: per prev. post (more info?)
|
|
|
On Mon, 09 Jan 2006 14:22:57 +0100, Thomas J. Boschloo wrote:
> Alyssa Milburn wrote:
>> Yeah, that's pretty much just writing the raw data to the files, it's
>> definitely the easiest way.
>
> But that would depend on the endianess of the processor architecture
> wouldn't it?? iow, you select endianess at compile time
Well, you only 'select' endianness in that you choose which processor to
target, generally. But yes.
But, I meant, the standard doesn't do any fancy encoding tricks. It says,
hardware is responsible for making sure the endianness of the bytes is
correct etc, and to dump the rest in .. big-endian, I think/assume. But
that's just a question of a little byteswapping. >> Stay informed about: [oc2e] Serialization? |
|
| Back to top |
|
 |  |
External

Since: Jul 06, 2005 Posts: 1641
|
(Msg. 14) Posted: Mon Jan 09, 2006 9:55 am
Post subject: Re: [oc2e] Serialization? [Login to view extended thread Info.] Archived from groups: per prev. post (more info?)
|
|
|
Alyssa Milburn wrote:
> On Mon, 09 Jan 2006 14:22:57 +0100, Thomas J. Boschloo wrote:
>
>>Alyssa Milburn wrote:
>>
>>>Yeah, that's pretty much just writing the raw data to the files, it's
>>>definitely the easiest way.
>>
>>But that would depend on the endianess of the processor architecture
>>wouldn't it?? iow, you select endianess at compile time
>
>
> Well, you only 'select' endianness in that you choose which processor to
> target, generally. But yes.
>
> But, I meant, the standard doesn't do any fancy encoding tricks. It says,
> hardware is responsible for making sure the endianness of the bytes is
> correct etc, and to dump the rest in .. big-endian, I think/assume. But
> that's just a question of a little byteswapping.
Byte (octet) swapping is the whole point of endianess isn't it
And saving 16 bit words to disk in an intel environment will result in
byte swapping, which is exactly what you don't want here if you follow
RFC 1014!
Kind Regards,
Thomas
--
Gothika: "How can you trust someone who thinks you are crazy" >> Stay informed about: [oc2e] Serialization? |
|
| Back to top |
|
 |  |
External

Since: Jun 16, 2005 Posts: 22
|
(Msg. 15) Posted: Mon Jan 09, 2006 9:55 am
Post subject: Re: [oc2e] Serialization? [Login to view extended thread Info.] Archived from groups: per prev. post (more info?)
|
|
|
On Mon, 09 Jan 2006 14:50:27 +0100, Thomas J. Boschloo wrote:
> And saving 16 bit words to disk in an intel environment will result in
> byte swapping, which is exactly what you don't want here if you follow
> RFC 1014!
It won't result in byte swapping, since that's how it is inside the
processor, and in memory, etc, it'll just result in a different endianness
to the big-endian 'standard', which you might want to fix by byte swapping.
The thing is, I'm not convinced of the point of RFC 1014, it could be
summarized as saying "write the raw standard numbers, making sure that the
hardware copes with byte endianness, and you swap the data to big-endian
if necessary". >> Stay informed about: [oc2e] Serialization? |
|
| Back to top |
|
 |  |
|