Inform 7 release 6M62

If I’ve understood correctly, the Mac build now at should address the Mac IDE issues that are marked as fixed on Mantis.

Thanks, Emily. I installed the new release, and I can confirm that I’m no longer seeing 1807 (the crash on startup) or 1815 (an issue with indented headings). However, I am seeing 1808 (grey bar floating around the workspace as the window is resized) on 10.6.8.

I created a Mantis account, but it’s not clear to me how to add a note to an existing issue.

The MacIDE issues being mostly resolved is excellent, though if someone could help vlaviano update Mantis it’d be even better.

As for the issue of non-English games being totally broken, which I can concur to ocasionally seeing mentions of in non-English communities and it’s natural for it to be a HUGE problem there and non-existent here, Emily said

I’ve seen you since posting things like this - are they in response to the first of your monthly emails?

I’d find it informative to be able to see your summary of outstanding issues, what would you think about posting the email you send here in the forum? (Minus any sensitive issues, if there are any.)

They are indeed. I thought about doing what you just asked about, then thought “Ah, I’ll just send what people are talking about now. If I prompt people, they’ll come up with suggestions, but that’s what Mantis is for; the idea of this endeavour is to see what comes up naturally, to get a real feel for what people are actually actively in need of”.

I expect to over/under-report for a while, until I get the hang of what’s actually worth reporting! So although in my post I did say “bugs people are finding critical”, I kinda extended my scope a bit. There just isn’t that much that’s critical to report, quantity-wise! It’s just that the stuff there is is crucial to a number of people, so it gets more focus.

I also decided against the dedicated thread I mentioned I could do. There just isn’t enough volume to justify it at this point, maybe in the future.

What follows, in the rant tag, is the e-mail I sent Emily Short, verbatim, minus some basic formatting.

[rant]Hello there! I suppose this has sort of been agreed upon, so I went over the posts made in March and am sending you the things that may concern the I7 dev team that crop up on the board.

I’ll probably also include stuff that isn’t really relevant; in that case, just let me know. I prefer to have too much stuff in than miss out on anything. If something also starts to come up on the board that may be a known issue with no plans to be fixed (like the colour support in Glulx, which involves a redesign of Glulx rather than anything I7 specific)… then I’ll also bring it up. Because if it’s important enough that people keep talking about it, then it may be of a higher-priority than one would think!

Right then.

vlaviano took a very in-depth look at what goes on behind the scenes when the player’s command is “x my arms” and such. It all goes way over my head, and ends with the following:

“TL;DR: “my” doesn’t work in Inform 7 for things that are part of the player, because they aren’t children of the player object. I7 should also consider component_parent == selfobj when parsing “my”.”

The second sentence is the one I thought might be useful to bring to your attention. It doesn’t mean much to me, but looks like a suggestion that no one else commented on, and it had a lot of thought behind it.

(probably definitely in the “over-reporting” cathegory, but I won’t know unless I report!)

It came up and went unanswered that a certain rule, the “carrying requirements rule”, was not coming up when using the test command RULES ALL. A J Mako said “Internal rules apparently aren’t mentioned.”, mikegentry said “Yeah, that’s weird. I don’t know why the carrying requirements rule isn’t declared.”, and that’s as far as it went. Since the RULES ALL command is usually eveyone’s first port of call to find out why things are not working the way they want it to work…

(in this particular case, the OP wanted to reorder some rules so that his custom messages for “You can’t wear a person” actually fired; because by default it never gets to that stage, being stopped by an implicit take “You can’t take that”)

…this small incident may not be much of an issue, or it may be a point in which things could be made clearer in the manual.

Or maybe, you know, the rule actually WAS supposed to list on RULES ALL and it doesn’t and it’s such a small thing no one bothered. :slight_smile:

The relevant last posts are

Upon using “showme me”, mikegentry got a very concise, clear and readable list, except for a long line that said:

“singular-named, proper-named, unmarked for special listing; unlit, inedible, portable, examined, nearby, relevant, disclosed, inflammable; female, awake”.

It didn’t seem to refer to anything, and seemed to mix “properties an object can have, (…) properties a thing can have [and] are properties a person can have.” (matt w). It also caused zero impact on the game overall. But it’s something that made people curious and also went unanswered.

Finally, the big one. We can start this thread at:

The bugfix site says at that the issue causing I7 not to run on Macs is fixed, though an actual usable Mac version containing this fix hasn’t been released. Teaspoon has been told to contact the Mac IDE maintainer, but no one could tell him how.

Natrium729 says, and this is something that comes up with lots of sighing in non-English communities occasionally, that 6M62 remains “completely broken for non-English projects (the infamous parse tree broken error”. This is a big one. I really do occasionally see people lamenting it in other communities - they’re completely locked out of 6M62.

tinman brings up the issue where the screen is partially obscured when resizing a panel, which persists though it’s considered “resolved”. May also be a matter of contacting the Mac IDE maintainer, I suppose.

—> And this last one is the one you replied to, and the point where I offered to do this, so we’re pretty much caught up <–

Vlaviano brings up the issue where 6M62 crashes on startup, fixed but not shipped and effectively preventing him from using (and testing) 6M62. He makes a good point when he says:

"I realize that the overall impact of a bug is a weighted product of severity * frequency over the entire userbase, and not many people are still running Snow Leopard. For me individually, though, severity is critical and frequency is always. It would have been nice to have seen a bugfix release as soon as that fix went in.

The situation also has addon effects like my withholding bug reports because I can’t repro a bug on the latest release because I can’t run the latest release."

Again, you already replied to this one. I’m just putting it here because I don’t want to miss anything.[/rant]

EDIT - BTW, I’ll see if she replies to vlavliano’s latest post. If she doesn’t I’ll e-mail her in a few days - since it’s a fix that didn’t actually get fixed, I’m not sure it’d be a good thing to wait a whole month.

Thanks for doing this Peter :slight_smile:

And posting it just in the rant tag here could be a good alternative to a new thread.


Yes, Peter, thanks so much!

It works! I can now download extensions! All the extensions!

(I am indescribably happy. Thanks to everyone in the process - Peter, Emily, Toby - for making the fixes. Really looking forward to testing out those conversation extensions now.)

Update: and spoke too soon. It now runs 62, I can see the public library, and I can now (apparently) install extensions but I still can’t get extensions to run. Instead I get the following error message in my Installed Extensions tab:

"Warning. Inform checks the folder of user-installed extensions each time it translates the source text, in order to keep this directory page up to date. Each file must be a properly labelled extension (with its titling line correctly identifying itself), and must be in the right place - e.g. ‘Marbles by Daphne Quilt’ must have the filename ‘Marbles.i7x’ (or just ‘Marbles’ with no file extension) and be stored in the folder ‘Daphne Quilt’. The title should be at most 51 characters long; the author name, 51. At the last check, these rules were not being followed:

Default Messages.i7x - non-folder found where author folders should be"

Also it minimises Inform at random whenever you try looking at the extension stuff.

So…I guess that might be something to send along in another month, Peter. Heigh-ho.

further further update: realised that Default messages was a relic of prior download attempts and unnecessary, and scrapped it. Inform now accepts Extensions, and will even run updated ones (I checked with Object Response Tests, updated for 62, and it works beautifully. Wish I’d had this earlier.)

On the other hand, it still randomly shuts down your project when you try switching tabs. So…still not very usable behaviour. But, it’s lots better than it was.

Thanks all!

I’m on the fence about that. Issues that are not really resolved probably shouldn’t wait that long before they’re reported - although in your case they ultimately were. The big ones, at any rate. Then again, trying to work and having I7 mionimize/shut down all the time is not fun.

Again, I’ll see if Emily noticed this thread, and see if there are more reports piling up here, and probably e-mail her by the end of the week. Or possibly in two weeks, since she did say that that was the minimum comfortable period.

More playing around with it suggests that I can download the extension in 62 and put it in a game, then shut it down and open it up in 38 with no apparent random-close issues and the extension available. I’ll be working with this further, but if it works out then I can use extensions with a fairly straightforward workaround, which is the main thing I wanted 62 for in the first place.

Will keep you updated. Thanks for noticing.

Is this still where I post what I send to Emily? Well, I’ll keep posting here until I’m told to go elsewhere.

[rant]Hello there! These are the things that have cropped up on the I7 board at IntFiction last month:

To make this one short: Teaspoon is seeing an issue (which doesn’t seem to have happened to anyone else yet) where switching tabs randomly shuts down the project. He then says “More playing around with it suggests that I can download the extension in 62 and put it in a game, then shut it down and open it up in 38 with no apparent random-close issues and the extension available.”

I PMd Teaspoon to check up on it and here’s the reply I got:

"Well, it turns out that 38 was capable of running extensions all along, and I just hadn’t realised it do to a combination of various factors (did you know you have to click on an inbuilt extension after you’ve written the “Include Such and Such by Such and Such” in your source code to get Inform to process the extension? I was not expecting a point-and-click-adventure with my coding). With this being the case, I’m actually pretty content with running 38 instead of 62.

I suppose the crashing might still be a problem for someone running 62 though, so…possibly the problem is not actually fixed yet.

I could try to figure out the trigger if this is useful for other people, but if it’s just me, I’m good for now."

So I’m not sure this is actually useful to you, but I figured I should at least tell you. After all, it’s a user that is having issues and decided not to use the latest version due to them. But it’s hard not being able to track the issue down.

In this thread, the issue that came up is related to … -of-a-kind. This is officially under review, I see, so I just wanted to let you know it cropped up again.

This is where the recent Mantis report came from. Also, Eleas noticed that ‘regular numbers can be written out as words, which works in all cases other than 0: “zero”, “nought”, “nil” or “null” all fail to be recognized.’ Could this be a small issue that merits a little update for consistency?

Apparently there’s an “if” statement with no body that should cause an error report, and it isn’t, which according to zarf looks like a compiler bug.

vlaviano unearthed It’s filed and being processed, and hasn’t come up elsewhere yet.

vlaviano made a detailed (as usual) study on disambiguation with “referring to part of a person using ‘his’ and ‘her’”. As usual most of it’s greek to me (or should I saw Inform 7, which is almost its own language anyway?), but he ends with a question and a statement no one’s addressed yet:

"- The one indefinite match case is special. Why?

  • The parser’s notion of ownership should include parts of people."

Very simple one. It is felt that the documentation doesn’t really clarify what “person aksed” really is. It tends to create small misunderstandings. … ead#unread

zarf identified this as an actual bug. He also provides some workarounds. Sorry, but I can’t really summarise it as I’m not sure what it actually is.

The inclusion of an I6 line…

To say perform/@ (ph - phrase): (- if (0==0) {ph} -).

…breaks compilation in 6M62 with the “I have no idea what happened” error. Also, further down the thread, there’s a little bit of discussion (didn’t go anywhere much) about deprecated phrases.

This one also goes over my head. It appears there are inconsistencies when using multiple verb clauses in a description (say zarf and otistdog). Sample code was posted which gave otistdog “the impression that [he was] running into behavior that’s half-working (but is intended to be fully working).”

The language thing again! What happened this time is that someone is using an extension which the author cannot keep updated for 6M62 compatibility because he relies on the Italian Language Extension. So this language problem keeps cropping pu.

Very quick one. afeinman was looking for something, which matt w identified as:

“It’s the “prior named object.” (I feel like this used to be in the documentation somewhere, but now it seems to be only in Example 251.)”

I bring this up because the documentation might need updating in this instance. For clarity.[/rant]

Teaspoon, I’ll PM you, but basically Emily would thank you if you were to go into Mantis and add the issue you’re having. The IDE maintainer will then see what they can see.

Guys and gals, it was very, very, very clear in Emily’s e-mail (she didn’t say so outright, but she said “file it on Mantis” so much I felt sorry) that the very best thing you can do to help iron out these issues is to file them. If you see something and it’s a confirmed issue, even if it’s a very minor thing like a documentation typo or an obscure bug that only occurs every second tuesday after 11:23pm except on months that have 30 days… please do file it!

To everyone who already does, many thanks!

@Peter: Thanks for continuing to do this.

I’ve also experienced Teaspoon’s issue. I filed it in Mantis as issue 1895 a while back.

Ah that’s great. :slight_smile: Reproducible bugs are bad, but not half as bad as non-reproducible ones!

Nothing much happened this last month, which is good! That, and/or I’ve started to filter things somewhat better. Below is what I sent to Emily.

[rant]Hey there! Not much has been happening on IntFiction (I7 board), and I’ve learned what isn’t worth reporting, so this is a short one (no pun intended!)

This is something that was on mantis already: It just came up again, so I’m letting you know.

It’s not really a bug, as Graham explains, but it IS subject to some confusion. The sort of thing that, when explained, makes you go “Oh, of course! How obvious!”, but until someone does, you’re staring at the screen in frustration wondering why it doesn’t work when it looks so right to you!

Amazingly, the issue…

…managed to come up twice in three threads!

And, related,, a post in which a change was necessary for Computers to work in Counterfeit Monkey. I’m unsure whether this means you have to update the version of Computers in the Public Library, but anyway, I’m just the messenger, so I’m just messaging! (Note: Since I added this last one to my log, angstsmurf has updated the extension. Just letting you know - I’m not sure which version is in the public library, and which version you’d WANT in the public library)[/rant]

Another monthly report! Notably, this month was mostly about unclear documentation, the failings of UserVoice as a useful system, and a couple of issues which make more sense on an abstract, conjectural level than they do on the practical level, i.e., people actually using and mastering them.

[rant]Another month, another report! Incidently, your recent post about immigration was very interesting, thank you again for sharing. Cheers!

This is about the order of evaluation (the sort of thing you use parenthesis to avoid nightmares about). matt w didn’t have enough points for UserVoice (which seems to be a common complaint, undermining UserVoice’s utility I guess), so he points out in the thread that the documentation could at least be made clearer. This is supported by other users, as people seem to be struggling to follow the logic behind the “usual conventions” - as Juhana says, simply put, “The usual convention isn’t clearly defined.”.

The issue is, of course, in phrase evaluations. Arithmetic evaluations seem to function as expected.

This one went a bit over my head. During some optimising of text generation (specifically making it so that it would correctly produce a/an when applicable) matt w started running against some problems. One of them was filed as a bug, but I’m unsure that that covers the whole thing.

I can’t sum this one up, sorry. If you could take a look for yourself it’d be much appreciated.

There are some issues that go beyond the bug report as well–there’s the weird spacing, and whatever the effects of preloading the text substitution were, and, well, just in general I’d like to know how indefinite articles work. The bug report is mostly about the issue two posts up from that, I think.

“Item described” = “self in I6”. This is still confusing (zarf to call it “messy”, which is certainly an alarm bell ringing), and many feel would benefit from at least clearer documentation.

Another bit in the documentation that could be made clearer. Nothing major, though.

These two cover the same area: having a command that’ll stop a timed event. One of the posts even has code that will do that.

I’m reporting these because to have the same issue brought up twice in a month (and for someone to actually code a hack to make it happen) means that there is some interest in the user’s practical side. I understand the reasoning for the way it works, but in practical terms, people seem to want/need it.

One of the threads also covers the UserVoice issue I brought up earlier, where it might be falling short of its utility.

This is on uservoice, but there’s clearly interest: this is an active discussion on how best to rework a phrase that’s sometimes confusing. As Draconis says, On Uservoice there’s been the suggestion to change the phrasing “a random hat worn by the player” when the author knows there can be only one: instead, “the hat worn by the player” is more readable, and could make more efficient code (why loop through all hats if you don’t need to?).
Really active discussion there, so it’s like with the timed event thing: even though it works the way it is and there’s a reason behind it, people are struggling and actively trying to change the behaviour. Might be worth taking a look at, therefore.[/rant]

EDIT - Bleh, I left something in there that shouldn’t be. :slight_smile: matt w, I’m sure you’ll recognise a little non-sequitur because you PMd it to me! Copy-pasting…

Here’s an issue I’m having in 6M62, which I believe might be a Windows IDE specific problem: The section, chapter, etc. headings aren’t being formatted properly (at all) in the Extension tab, including the examples. I thought this was only my extension at first, but no, it seems to be all of them.

Has anyone else run into this?

This was just filed:

It’s all platforms, not just Windows. I don’t see anything about it in the I7 changelog, so I assume it’s a mistake.