Dialog

lft
Linus Åkesson
Wed 12-Dec-2018 09:35

The "can't reach" message is printed by (refuse $), which is invoked before instead-of. So you can override refuse with a negated rule:

~(refuse [take *]) %% Don't refuse to take this object.

Anonymous
Thu 13-Dec-2018 11:55

Is it possible to suppress implicit actions temporarily.
For example
1.dropping something in a held container first tries to take the object before dropping.

2. putting something into something where the first something is in a held container tries to take before putting in

If I know that the take will always succeed can I suppress the first trying to take reporting.

lft
Linus Åkesson
Thu 13-Dec-2018 22:54

Yes, there are several ways. You can override the before-predicate for that particular case, with a negative rule. For each particular action:

~(before [drop $X])
($X is #in $Y)
($Y is #heldby $Z)
(current player $Z)

But many actions invoke a common predicate, '(ensure $ is held)', and you can add a new rule to that one instead:

(ensure $X is held)
($X is #in $Y)
($Y is #heldby $Z)
(current player $Z)

When the given conditions are true, your rule will succeed, and this prevents the later rule definition in the library from executing.

Eleas
Björn Paulsen
Fri 14-Dec-2018 11:30

The Cloak of Darkness example seems bugged under 0.15. Attempting to go somewhere yields a crash:

Eleas
Björn Paulsen
Fri 14-Dec-2018 11:34

This is odd. I get the same error when running the first example in chapter 5. I'm running the win32 version of dialogc on Windows 10, in case that matters.

Eleas
Björn Paulsen
Fri 14-Dec-2018 11:44

I've now checked each release of the compiler, and found that moving between rooms only works under release 0b/01. Later than that leads to the fatal error described above.

Anonymous
Fri 14-Dec-2018 12:18

Not sure if it's related but I only get the error when using Gargoyle. Frotz doesn't give the error. (latest versions Windows 10)

Eleas
Björn Paulsen
Fri 14-Dec-2018 12:54

Thank you. Can verify. Does anyone know the best way to pass a bug report to one of the maintainers?

lft
Linus Åkesson
Fri 14-Dec-2018 13:47

This looks like a Dialog bug. Thanks for reporting; I'm on it!

lft
Linus Åkesson
Fri 14-Dec-2018 14:36

Fixed in 0c/04.

Anonymous
Tue 18-Dec-2018 12:05

Can I reference an object variable from within the scoring options?
(increase score by (#object variable))

lft
Linus Åkesson
Tue 18-Dec-2018 13:14

As a general rule, queries can't be nested; they don't even have return values.

(Confusingly, rule heads may contain nested expressions, but that is a special case of syntactic sugar.)

But output can be communicated using ordinary parameters. To take the output from one query, and use it as input in another, you would write the queries one after the other. Then you'd use a variable to carry the value.

Assuming you have a predicate ($ has score $) that assigns a score to every object:

(#object has score $X) %% Here, $X becomes bound (i.e. to the output).
(increase score by $X) %% Use the value of $X as input.

mstram
Sun 3-Feb-2019 07:01

The following code crashes the compiler

If I substitute the (else) with (if), I get a more reasonable message : "Error (if) without (then).

** The comment editor seems to strip away the indentation ... does it accept some kind of 'markup' ? (I.e. the second line (else) *was* indented.

Also I had much more in the source file, I isolated the crash down to these two lines.

Assertion failed!

Program: c:\bats\dialogc.exe
File: backend_z.c, Line 3348

Expression: call_lab

--------------------
(story title)
(else)

mstram
Sun 3-Feb-2019 07:15

The actual 'realistic' use case for the above was that I had accidentally "not-escaped" the title i.e.
(story title) ( Else) test file
instead of

(story title) \(Else \) test file

lft
Linus Åkesson
Sun 3-Feb-2019 07:54

Thanks! I'll fix that.

There's currently no tag to retain formatting in the comments; sorry for that. I've been using hardspaces to get around it. I suppose I'll fix that too.

mstram
Sun 3-Feb-2019 08:39

Are global flags (or maybe it's the (now) predicate) designed to work in "open code" (like global variables?

If I put :

(now)(a-global-flag) in "open code", I get :

"Special syntax cannot be redefined"

Mike

lft
Linus Åkesson
Sun 3-Feb-2019 20:54

Anything that starts in the very first column of a line is interpreted as a rule definition, so the compiler thinks you're trying to define a rule for a predicate called '(now)'. But that's special syntax, so its behaviour can't be modified.

If you would like the global flag to be initially set, define a rule with an empty rule body:

(a-global-flag)

This is analogous to how the initial values of other kinds of dynamic properties are defined. For instance:

(#box is #heldby #player)
(#box is open)
(current player #player)

mstram
Sun 3-Feb-2019 23:24

Ok, thanks, that works ... but ...

The "first-column-global" (fcg ? :)) is not showing up in the debugger's @dynamic output.

The flag *does* work though (of course),if I enter
(a-global-flag), giving the expected "Query succeeded:(a-global-flag)

The @dy(namic) cmd seems to be only finding global flags that are referenced inside a rule :

(test)
__ (now)(a-global-flag)


Mike

lft
Linus Åkesson
Mon 4-Feb-2019 14:02

That is correct. A predicate is only considered dynamic if it appears in a (now)-statement somewhere. So, if there's a (now) (a-global-flag) somewhere, then (a-global-flag) is regarded as a boolean variable that can be toggled at runtime. And its initial value is determined by the rule definitions for it. Otherwise, it behaves more like a boolean-valued function, and the rule definitions make up its function body.

It's a matter of terminology. Something that can change is dynamic, but something that cannot possibly change is not.

mstram
Wed 6-Feb-2019 10:12

This crashes both the windows and linux-64 compilers, version
0d02_0_19, and 0e01_0_20,but NOT the 0c05_0_17 version

#gnue
(name *) gnue
(dict *) plane
(vehicle *)


Again it's only a code fragment, but I was trying to split a larger program into multiple files.

Mike

lft
Linus Åkesson
Wed 6-Feb-2019 19:10

Thank you! I can reproduce it. Amusingly, I can get the program to compile by adding a room:

(room #x)

But it's still a compiler bug, of course, and I'll fix it as soon as possible.

lft
Linus Åkesson
Wed 13-Feb-2019 21:01

Fixed in 0e/02.

Anonymous
Sun 19-May-2019 10:00

I think there may be a bug with (fungibility enabled appearance $ $ $)

I noticed that when I had (fungibility enabled) an explicit (appearance $ $ $) rule set on an object did not fire until the object had been moved. After a bit of digging I came to the conclusion that the problem was with stdlib line 3788. What seems to be happening is that when nothing's been touched the list being handled by (fungibility-enabled appearance $ $ $) starts with an object with no appearance, which means that a rule fails on the first object and it never recurses.

If I add an (or) so that it reads

(if)($Ninc = 1)(then)
(appearance $Head $Rel $Loc)
(or)

it works. I'm not sure that adding an (or) simply to prevent failure is good style of course ...

Paul

lft
Linus Åkesson
Tue 21-May-2019 19:59

Thank you! This was indeed a bug, and I've added the '(or)' in library version 0.24. I think it's a reasonably clean fix. The alternative would have been to add a default rule definition for '(appearance $ $ $)' with a blank rule body, but such a change would also have affected stories that do not make use of fungibility, so I decided against it. For now, at least.

Anonymous
Wed 5-Jun-2019 22:21

I can't seem to build this. backend.c and backend_z.h seem to be missing from the tarball.

lft
Linus Åkesson
Wed 5-Jun-2019 23:16

Whoops! Should be fixed now in 0g/02, which is otherwise identical.

Anonymous
Sun 23-Jun-2019 16:52

That got me going, thanks. Have you ever considered putting this in a public repository so that others can contribute, too?


I really like the idea of this system: it's a nice sweet spot between writing and programming.

Anonymous
Thu 15-Aug-2019 18:33

I have published a mostly-complete conversion of Pick Up the Phone Booth and Die to Dialog.

https://github.com/jpcompton/Dialog-diversions

Please enjoy dubious code examples, make improvements, or be inspired to greater heights!

Anonymous
Mon 16-Sep-2019 11:52

I just tried a hello world example and segfaulted the compiler (Dialog, release 0h/02, library 0.30, prebuilt for Linux x64)

$ cat test.dg
(story ifid) B0E2A7D1-45F7-46A1-8348-53ADA71F2CD8
(program entry point)
We'll find out how this works, slowly.

$ valgrind dialogc test.dg
==14539== Memcheck, a memory error detector
==14539== Copyright (C) 2002-2017, and GNU GPL'd, by Julian Seward et al.
==14539== Using Valgrind-3.13.0 and LibVEX; rerun with -h for copyright info
==14539== Command: dialogc test.dg
==14539==
==14539== Invalid read of size 1
==14539== at 0x12A4FA: addstr_escape.constprop.1 (in /usr/local/bin/dialogc)
==14539== by 0x12AE3D: emit_blorb (in /usr/local/bin/dialogc)
==14539== by 0x129FF1: backend_z (in /usr/local/bin/dialogc)
==14539== by 0x10D487: main (in /usr/local/bin/dialogc)
==14539== Address 0x0 is not stack'd, malloc'd or (recently) free'd
==14539==
==14539==
==14539== Process terminating with default action of signal 11 (SIGSEGV)
==14539== Access not within mapped region at address 0x0
==14539== at 0x12A4FA: addstr_escape.constprop.1 (in /usr/local/bin/dialogc)
==14539== by 0x12AE3D: emit_blorb (in /usr/local/bin/dialogc)
==14539== by 0x129FF1: backend_z (in /usr/local/bin/dialogc)
==14539== by 0x10D487: main (in /usr/local/bin/dialogc)
==14539== If you believe this happened as a result of a stack
==14539== overflow in your program's main thread (unlikely but
==14539== possible), you can try to increase the size of the
==14539== main thread stack using the --main-stacksize= flag.
==14539== The main thread stack size used in this run was 8388608.
==14539==
==14539== HEAP SUMMARY:
==14539== in use at exit: 613,930 bytes in 657 blocks
==14539== total heap usage: 1,170 allocs, 513 frees, 1,450,866 bytes allocated
==14539==
==14539== LEAK SUMMARY:
==14539== definitely lost: 7,600 bytes in 50 blocks
==14539== indirectly lost: 206,966 bytes in 102 blocks
==14539== possibly lost: 0 bytes in 0 blocks
==14539== still reachable: 399,364 bytes in 505 blocks
==14539== suppressed: 0 bytes in 0 blocks
==14539== Rerun with --leak-check=full to see details of leaked memory
==14539==
==14539== For counts of detected and suppressed errors, rerun with: -v
==14539== ERROR SUMMARY: 1 errors from 1 contexts (suppressed: 0 from 0)
Segmentation fault (core dumped)

Anonymous
Mon 16-Sep-2019 11:57

This seems to be linked to the (default) zblorb format. The others work fine.

lft
Linus Åkesson
Wed 18-Sep-2019 10:44

Thanks for reporting it!

This appears to happen for small input (i.e. without the standard library), in combination with the zblorb format. I'll fix it tonight. Meanwhile, either switch to the z8 format or include the standard library.

lft
Linus Åkesson
Thu 19-Sep-2019 06:01

This is now fixed in 0h/03.

Anonymous
Mon 23-Sep-2019 12:13

Using dialogc-0h03:

The following program crashes the compiler, regardless of the output format. It works on version 0d02.

(program entry point)
(now) ~($ has parent $)

Error messages:
backend_z.c:2010: generate_code: Assertion `ci->oper[0].value != DYN_HASPARENT' failed.
and
backend_aa.c:1350: compile_routines: Assertion `ci->oper[0].value != DYN_HASPARENT' failed.

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

The following three programs crash the compiler when compiling for the Aa-machine but work fine on the Z-machine:

(program entry point)
(#a = $X)

Error message:
backend_aa.c:752: encode_dest: Assertion `0' failed.

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

Note: I have been unable to minimize this program further because removing any line will cause the crash to disappear.

(program entry point)
($x = 3)
($y = 3)
($ = $)
([#a #b] = [#a #b])
([#a #b] = [#b #a])
([1 [@a] 3] = [1 [@a] 3])
([1 [@a #b] 3] = [2 [@a #b] 3])
([1 [@a #b] 3] = [1 [@a #b] 3])
([1 2 3 4] = [$a $b $c $d])

Error message:
backend_aa.c:3151: opersize: Assertion `aao.value <= 0x3f' failed.

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

And a similar program, also as simplified as possible:

(program entry point)
($x = 3)
($y = 3)
($ = $)
([#a] = [])
([#a #b] = [#a #b])
([#a #b] = [#b #a])
([1 [#a #b] 3] = [1 [#a #b] 3])
([1 [#a #b] 3] = [2 [#a #b] 3])
([1 [#a #b] 3] = [1 [#a #b] 3])
([1 [#a #b] 3] = [1 [#a] 3])

Error message:
backend_aa.c:3151: opersize: Assertion `aao.value <= 0x3f' failed.

lft
Linus Åkesson
Mon 23-Sep-2019 21:36

Thanks! I'll look into these. The first two are corner cases that I hadn't thought of; they should be easy to fix.

The last two happen because the compiler is running out of registers for holding temporary values. Fixing this involves a more substantial implementation effort that I've been putting off. But it needs to be done, of course.

Anonymous
Sat 5-Oct-2019 00:31

Any chance of linking to the source repo, to accept patches. I've found a couple of nits in the documentation I could fix.

Anonymous
Fri 11-Oct-2019 01:39

Is there a GitHub project for this? I noticed a couple of tiny documentation bugs that I would submit a PR for.

Also, what's the trajectory here: do you know what "feature complete" looks like, and what's the journey there look like?

I've been dabbling in IF for quite some time; As a coder (I work in Clojure) I prefer the Dialog syntax over Inform (as nifty as Inform7 can be).

Also your forums are broken; database error, can't register.

HLS

lft
Linus Åkesson
Sun 20-Oct-2019 21:35

It's on the todo list. My own repo is intertwined with the development of Tethered, so I have to create a new, public repo with a clean history.


There's no detailed long-term plan, but I think the bulk of the functionality is already in place. The Å-machine needs to run on vintage hardware, as promised, and I'm currently working on that. I have some ideas for new bells and whistles, but the core Dialog language feels fairly stable, and will probably evolve in a backwards-compatible way. But I don't promise this until we get to version 1, which will probably happen when (if) people start to release stuff made in Dialog.


Huh, I haven't been able to reproduce this. Is it still broken from your end?

hlship
hlship@gmail.com
Fri 1-Nov-2019 23:47

Thanks for the update.

I was able to register on the forum this time, so whatever it was, resolved itself.

hlship
hlship@gmail.com
Sat 2-Nov-2019 00:12

So, out of an abundance of hubris, I'm trying to bring Threaded Conversation to Dialog, despite:

- insufficient experience w/ TC in Inform7
- not fully fluent in Inform7
- new to Dialog
- very limited time to work on this!


I'm struggling with scope.

Quips are the basic thing in TC - they represent a exchange between the player and the NPC, and include predicates that govern who can say the quip, and how quips directly or indirectly follow each other (plus much more in TC).

So a player command might be ASK BARMAID ABOUT RUMORS and this should be understood as [discuss #whether-rumors-tell-truly with #barmaid].

But my code, as it stands, is failing:

> ask bar maid about rumors
The whether the rumors tell truly isn't here.

So that's a scoping issue.


What's confusing to me is that, when tracing, it looks like it found it:

| | | | | FOUND (understand [ask bar maid rumors] as [discuss #whether-rumors-tell-truly with #barmaid]) /usr/local/share/dialog-if/stdlib.dg:4710

(I haven't put in any code to restrict quips in any way, that comes soon)

but then Dialog continues on from there:

| | | | | | | | | ENTER (parse [bar maid] as object [1] [13] [1] 0) /usr/local/share/dialog-if/stdlib.dg:4394
| | | | | | | | | QUERY (allowing parse errors) /usr/local/share/dialog-if/stdlib.dg:4395
| | | | | | | | ENTER (split [maid rumors] anywhere into [maid | $] and $) /usr/local/share/dialog-if/stdlib.dg:5273
| | | | | | | | QUERY *(split [rumors] anywhere into $ and $) /usr/local/share/dialog-if/stdlib.dg:5274
| | | | | | | | | ENTER (split [rumors] anywhere into [rumors | $] and $) /usr/local/share/dialog-if/stdlib.dg:5273
| | | | | | | | | QUERY *(split [] anywhere into $ and $) /usr/local/share/dialog-if/stdlib.dg:5274
| | | | | | ENTER (understand [ask bar maid rumors] as [ask $ about $]) /usr/local/share/dialog-if/stdlib.dg:1916

So despite finding a good match, it seems to wander off based on the built-in ask action (?).

Later in the trace it feels like it's back to my discussing action, but something (even with the trace, it's hard to follow) deciding that #whether-rumors-tell-truly is not reachable.

I tried adding

((quip $) in is scope)

to make all quips in scope, but that doesn't seem to help.


My code so far (cobbled from stdlib.dg and guess work and experimentation):


(rewrite [ask | $Words] into [ask | $TailWords])
(split $Words by [about] into $Person and $MoreWords)
(append $Person $MoreWords $TailWords)

(understand [ask | $Text] as [discuss $Quip with $Person])
*(split $Text anywhere into $Someone and $QuipWords)
*(understand $Someone as single object $Person preferably animate)
*(understand $QuipWords as questioning quip $Quip)

(understand $Words as questioning quip $Quip)
*(understand $Words as quip $Quip)
(questioning quip $Quip)

(understand $Words as quip $Quip)
(filter $Words into $Filtered)
(nonempty $Filtered)
(determine object $Quip)
*(quip $Quip)
%% TODO: In scope for conversation ...
(from words)
*(dict $Quip)
(matching all of $Filtered)


So I thought I'd check to see if I'm even headed in the right direction on this.

For TC managing the scope is pretty essential; TC is about a balance of revealing what you can talk about in the current point of the current conversation without giving away other possible conversation quips that don't (yet) apply. So ultimately, whether Dialog should recognize 'rumors' as referring the quip depends on who the player is conversing with, what quips have already been performed, and whether a potential quip follows (directly or indirectly) some previously performed quip.

Anonymous
Sat 15-Aug-2020 13:57

Am I missing something? The third example in Chapter4 Items doesn't appear to work. As shown the amethyst is not in the bowl. Adding (* is #in #bowl) to the amethyst seems to fix this. All of the options for multiple stones now work. Latest windows version.

Anonymous
Sun 16-Aug-2020 17:01

Okay, does this example need updating? Changing ((stone to (*(stone in line 39 seems to fix this.

lft
Linus Åkesson
Thu 20-Aug-2020 10:10

Thanks for reporting this! There was a change in the language definition a while back, and this example should have been updated. I'll fix it.

Anonymous
Thu 22-Oct-2020 10:37

Is there any plans for a swedish translation of the library?

lft
Linus Åkesson
Tue 27-Oct-2020 14:35

Not at the moment, because I don't want to lock down the design of the library too early, and I don't want to coordinate major design changes across multiple versions.

But feel free to make a translation of the current version, if you like! And make sure to check out the German library by Mikawa.

Anonymous
Thu 10-Dec-2020 12:51

The identical objects example at the end of standard library chapter 11 doesn't seem to work as described.

After running, the example commands do not work. For example get boxes gives the (I only understood you as far as wanting to take something.) message.

Thanks

Anonymous
Fri 11-Dec-2020 08:46

Is adding these the solution?

(plural dict (box $)) boxes
(plural dict (marble $)) marbles

There was a library change between 14 and 15

(plural dict $O)
(plural name $O)

became

(plural dict $)
(fail)

Anonymous
Fri 15-Jul-2022 17:41

Have you finished working on Dialog?

lft
Linus Åkesson
Tue 19-Jul-2022 13:54

Not stopped as such, but it's been a long time since I made a new release. I needed a break from IF after the very intense work that went into The Impossible Bottle, and during that break other projects and interests floated to the surface. But I have made some unreleased improvements to the Dialog compiler since then, so in that sense the project is still ongoing. I'd like to get TIB up and running on the c64, which requires some tool changes. I'm very happy to see that people are using and discussing Dialog.

Anonymous
Wed 10-Aug-2022 07:17

You have clearly demonstrated by your work such as Impossible Bottle that Dialog if very capable. As it stands now, is it close to full release or can we expect further beta releases in the future? In particular, can we expect future releases to cause deprecations in the current code? Thank you.

lft
Linus Åkesson
Fri 19-Aug-2022 12:47

Thanks!

I would say that most of the language is stable, but I do have a few changes in mind that would break compatibility.

The first is to remove the inline style-changing predicates, e.g. '(bold)', because style classes already cover this functionality. Removing the inline versions will simplify the language definition and the virtual machines.

The second is to extend the range of integers to include larger numbers and negative numbers. This will break the current behaviour where e.g. '($ minus $ into $)' fails if the answer would be negative.

Third, I'd like to add built-ins for analyzing and synthesizing words that have a mandatory part and an ending. When that's available, removable word endings can be implemented entirely in library code, and thus removed from the language itself.

I'm also pondering if maybe number should be a subtype of word (so all numbers are also words). This would in my opinion make the language definition more elegant, but it would definitely break old code.

So I don't want to leave beta just yet. But it may be possible to write games that will work with future versions of the language, if the above changes are taken into account.

Anonymous
Sun 21-Aug-2022 17:40

Very nice! Thank you for your thoughtful reply. I am looking forward to future releases.

Anonymous
Fri 26-Aug-2022 16:59

Hi Linus,

Thank you so much for working on and sharing Dialog, it got me back to IF again, and I really love how the language itself makes you think out of the box when compared to a considered-to-be standard programming language.

In case you missed it, I've posted a possibly incorrect unification behaviour regarding lists in the Dialog category over at IntFiction, which currently inhibits lists from being more dynamic by being able to apply items to them multiple times.

Regarding numbers, I like that they are regarded as rule head parameters at the moment, they are particularly useful when used as output for queries.

Cheers!

Anonymous
Sun 29-Oct-2023 23:10

In a thread involving implementing a "Talk To" rule to Dialog, you mentioned you would consider adding it to the library.

Any further thoughts?

Thank you.

Can't compile on Windows (in MSYS) because this uses termios.h and ioctl.h. Definitely not a good decision.