I7: Memory issues
weevil
story to compile to z8 as opposed to glulx. The problem is that I am
overflowing the maximum memory for for Z-machine format. I had to
increase my MAX_STATIC_DATA to 132000 when it was Glulx to get it to
work, and that's too big for Z8. I'm not sure what to do. Maybe
there's something I did wrong in my code that made it to where I had
to increase the memory to that large a size. I have some long text
lists in it, but it doesn't seem like they are that long...
Does anyone know of any common coding errors in inform that could
cause max_static_data to need to be raised really high to get the game
to work?
If I use the RULES ALL command in game, what do I look for that would
suggest an infinite loop?
Andrew Plotkin
> I wanted to try out parchment tonight, but I can't seem to get my
> story to compile to z8 as opposed to glulx. The problem is that I am
> overflowing the maximum memory for for Z-machine format. I had to
> increase my MAX_STATIC_DATA to 132000 when it was Glulx to get it to
> work, and that's too big for Z8. I'm not sure what to do. Maybe
> there's something I did wrong in my code that made it to where I had
> to increase the memory to that large a size. I have some long text
> lists in it, but it doesn't seem like they are that long...
> Does anyone know of any common coding errors in inform that could
> cause max_static_data to need to be raised really high to get the game
> to work?
MAX_STATIC_DATA is where I6 array data is stored (in the I6 compiler)
before being written out to the final game file. So if you need to
crank this up, then you're creating a lot of I6 arrays.
Unfortunately this doesn't pin down the problem, because many
different I7 features are implemented as I6 arrays. Look for:
- very large tables
- many-to-many relations
- heavy use of indexed text and other dynamic allocation
> If I use the RULES ALL command in game, what do I look for that would
> suggest an infinite loop?
A sequence of rules that get called over and over and over in a single
turn? :)
--Z
--
"And Aholibamah bare Jeush, and Jaalam, and Korah: these were the borogoves..."
*
weevil
> Here, weevil <wileywg...@gmail.com> wrote:
> > I wanted to try out parchment tonight, but I can't seem to get my
> > story to compile to z8 as opposed to glulx. The problem is that I am
> > overflowing the maximum memory for for Z-machine format. I had to
> > increase my MAX_STATIC_DATA to 132000 when it was Glulx to get it to
> > work, and that's too big for Z8. I'm not sure what to do. Maybe
> > there's something I did wrong in my code that made it to where I had
> > to increase the memory to that large a size. I have some long text
> > lists in it, but it doesn't seem like they are that long...
> > Does anyone know of any common coding errors in inform that could
> > cause max_static_data to need to be raised really high to get the game
> > to work?
>
> MAX_STATIC_DATA is where I6 array data is stored (in the I6 compiler)
> before being written out to the final game file. So if you need to
> crank this up, then you're creating a lot of I6 arrays.
>
> Unfortunately this doesn't pin down the problem, because many
> different I7 features are implemented as I6 arrays. Look for:
>
> - very large tables
> - many-to-many relations
> - heavy use of indexed text and other dynamic allocation
>
I have a lot of fairly long text lists. It'll ve a shame to have to
dump them.
> > If I use the RULES ALL command in game, what do I look for that would
> > suggest an infinite loop?
>
> A sequence of rules that get called over and over and over in a single
> turn? :)
I guess I meant to ask, how would it show that? All I get is a list of
rules.
Andrew Plotkin
> On Nov 11, 10:20�am, Andrew Plotkin <erkyr...@eblong.com> wrote:
> >
> > Unfortunately this doesn't pin down the problem, because many
> > different I7 features are implemented as I6 arrays. Look for:
> >
> > - very large tables
> > - many-to-many relations
> > - heavy use of indexed text and other dynamic allocation
>
> I have a lot of fairly long text lists. It'll ve a shame to have to
> dump them.
There may be other ways to accomplish what you want. We did manage to
make these things work before dynamic lists were adding to I7.
On the other hand, Z-machine support might not be worth the hassle.
(I realize that my incomplete Glulx Javascript app is your stumbling
block.)
> > > If I use the RULES ALL command in game, what do I look for that would
> > > suggest an infinite loop?
> >
> > A sequence of rules that get called over and over and over in a single
> > turn? :)
>
>
> I guess I meant to ask, how would it show that? All I get is a list of
> rules.
Is the list infinite?
Erik Temple
> On Nov 11, 10:20 am, Andrew Plotkin <erkyr...@eblong.com> wrote:
>> Unfortunately this doesn't pin down the problem, because many
>> different I7 features are implemented as I6 arrays. Look for:
>>
>> - very large tables
>> - many-to-many relations
>> - heavy use of indexed text and other dynamic allocation
>>
>
> I have a lot of fairly long text lists. It'll ve a shame to have to
> dump them.
You may just need to shorten them. I assume these are lists used to create
random objects?
>> > If I use the RULES ALL command in game, what do I look for that would
>> > suggest an infinite loop?
>>
>> A sequence of rules that get called over and over and over in a single
>> turn? :)
>
>
> I guess I meant to ask, how would it show that? All I get is a list of
> rules.
If you see the same rule or set of rules repeated over and over again--and
it shouldn't be--then you have a problem. However, if it's truly an
infinite loop, you probably wouldn't have a chance to enter any input at
all after the loop has begun (since the game remains trapped in the loop
and can't restart the turn sequence.) Note that a rule has to have an
explicit name to show up in RULES ALL output.
If you are trying to debug something specific and rules all isn't helping,
I find it's also useful to turn on activities and actions debugging. (I
think you just type ACTIVITIES and ACTIONS.) Actions can be especially
useful because it indicates when you are repeating through a list and can
tell you precisely when an error is occurring within that loop. (Note that
this isn't the same as the infinite loop that might result from problems
with your rulebooks or activities, though.)
--Erik
weevil
>On the other hand, Z-machine support might not be worth the hassle.
>(I realize that my incomplete Glulx Javascript app is your stumbling
>block.)
I guess it depends on how far off in the future parchment support for
glulx is. Unlike most of the people doing IF, the presentation layer
is as important to me as the game- this isn't a game to be played with
a standalone interpreter, it's going to only exist as a playable
webapp. I was hoping I could have it up and running by the end of the
year.
> Is the list infinite?
>
Aha, no.
Aaron A. Reed
up to your expectations, but the staff is working on it. Also, it's
amazingly roomy in here... Here... Ere...
Eliuk Blau
Hahahahahaha! =D