Crossfire Archive
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

No Subject

________=22Re=3A_CF=3A_Another_client_bug=2C_update=22_=28Sep_24=2C__4=3A2?=
 =?iso-8859-1?Q?4pm=29?=
References: <199809241424.QAA04022@chapelle.eed.ericsson.se>
X-Mailer: Z-Mail (3.2.0 06sep94)
To: Raphael.Quinet@eed.ericsson.se (=?iso-8859-1?Q?_Rapha=EBl_Quinet?=),
 crossfire@ifi.uio.no
Subject: Re: CF: Another client bug, update
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii

Raphael Quinet wrote:
> If I remember correctly, you were asking to fix a bug with some bags
> that should never have been created in the first place (bags holding
> too many items).  If this is really the case, then I think that the
> correct solution is to remove these strange bags from the maps where
> they appear in, and not to change the code so that it allows this kind
> of objects to exist.  If would tend to fix this bug by adding some
> code in the server that silently enforces an upper limit to the weight
> and number of objects that a bag can carry, regardless of what is
> specified in the archetype.  Maybe the server could automatically
> remove these illegal containers from the player's inventory, or
> replace them by some standard bags.

 I thought about adding code to the server which would arbitrary put an upper
limit on either the weight reduction or weight limit in the bag.  But this
opens up a whole new can of worms (should sanity checks for other items be put
in place?  experience for monster?  power of weapons?  etc).  However, the item
in question is not on standard maps, so is not something I can remove/adjust.
 If those maps do get folded into the standard distribution, that bag would be
modified accordingly.


> I disagree.  We could keep on supporting the old X interface forever,
> and that would prevent any serious rewrite of the internals of the
> server.  Maybe the client is not as stable as you would like it to be,
> but then the best way to solve the problems is to submit patches to
> the code.  The client/server separation has already produced a very
> good result: the GTK client.  Once GTK themes are integrated, this
> will give us a very cool CrossFire client.  That would not have been
> possible with the builtin X interface.

 This is largely why I chose to seperate them.  It was going to happen sooner
or later.  If a bug chunk of code is going to be removed, it makes sense to
remove it sooner vs later simply because that is a bug chunk of code that no
longer has to be maintained.

 At some point, there was going to be a set of releases which are not
especially desirable from a end user aspect - in that case, they stick to the
last 'good' release, in this case 0.94.3.

 Some further comments that did not get addressed above:

> I do. Beta does NOT mean that the developers don't have to test it at
> all. These bugs are too obvious, they prove that only minimal testing
> has been done (if anything at all). This seems to be normal, though ---
> I reported another bug some time ago (concerning very full bags),
> and Mark told me that he'll not fix it, which was frustrating enough
> (since it means I'll have to fix it when more people start using the
> client. Luckily there is no danger of this happening right now).

 I have played extensively with the client without the errors.  However, there
is also a tradeoff - I have limited time to work on crossfire.  If I spend 40
hours doing extremely thorough testing, that means that is 40 hours that is not
spent improving the code or doing other bug fixes.  So there is some definately
tradeoffs here.

 AS for not fixing the container bug - I don't remember saying it.  There will
be at least a partial fix in 0.95.0.  However, there is still some limit where
if the number of objects in the container gets beyond a certain point, there is
no good solution (however, container handling for the client does need to get
improved).  But that number of items is only likely to happen with the infinite
storage containers.

>
> Also, we've had relatively stable crossfire versions in the past
> (except for the well known teleporter crash there are not that many
> crashes). With the new client things become extremely unstable (to
> avoid the word "unplayable"), so it is definitely a bad idea to
> discontinue support for the builtin X interface now.

 If the current version is stable, keep on using it then.  Many have argued
there is no good reason to get rid of the X11 code, because it works pretty
good as is - in same cases, that is true, in other cases, that is not true.


> What else should I wait for before being allowed to be frustrated?

 Keep using 0.94.3.

>
> The problem on my side is the lack of time to fix the worst bugs
> myself. Maybe I'll find some time when I'm done studying and before I
> start working; maybe I'll do other things then (esp. since I won't
> be able to run a public server anymore).

 Then you know how I feel - I have limited time, so I have to look at things
which are in the long term goal and which best uses the time available.
 Removing the X11 code from the server is a whole mess of stuff/bugs I no
longer need to worry about, and it had to happen sooner or later.

 Question:  With the standard X11 in the server in 0.94.3, what bugs are there
that makes the game unplayable with that old interface?  If none, then I see
very little reason to complain - just stick at that until the new client/server
has been demonstrated to be reliable enough for you.

 I think there is a very general idea in the software industry at large that
you always needs the latest and greatest version, even if there is nothing in
that version that you will ever use.  The same is true for crossfire - if
0.94.3 works for you, keep on using it.



-- 

-- Mark Wedel
mark@pyramid.com
Follow-Ups: