PyAI: How would we go about adding new commands into PyAI's logic, if we had the technology to add them into the game with a plugin?

PyDAT:
Under units -> AI, 'Sounds' is listed in the box that encases the fields.



Under orders, changing the 'highlight' field will not update the displayed icon unless you click off and back on to the entry you edited.

PyTBL:
I may have requested it before, but can we get a search box on the bottom left, similar to PyDAT? Would be much easier to access than the find function, and you could still keep the find function for settings (like `case sensitive`, `wrap`, and so on).

Speaking of which, does `case sensitive` even do anything? If I type in `ion cannon` without it ticked, it still doesn't send me to the Ion Cannon entry because of capitalization.

Post has been edited 1 time(s), last time on Nov 7 2017, 2:26 am by Pr0nogo.

PyDAT: When set to small window sizes, a lot of data fields are obscured with no scroll bar present to indicate that more data exists. Possibly true of other PyMS programs.

Doesn't look like PyLO works anymore. Throws this error whenever I try to open any files.



edit: redownloading fixed this issue??? no idea why

Post has been edited 1 time(s), last time on Nov 24 2017, 12:16 am by Pr0nogo.

I'm going to start this by saying sorry for not responding in a long time. I got busy, and then put on a new project at work which has destroyed any motivation to do technical things when I get home. It's getting better, so I might have the motivation to put in some work soon. For now I'm still collecting the issues as tickets on github.


That is a "crash", I just capture it to provide the traceback and then attempt to continue the program if possible, though in this case the app is too unstable to continue so it closes. I thought I had fixed this crash, but I don't know if I'll be able to figure it out without getting on a windows machine and debugging it.


Sorry this didn't make it in before I got busy, I hope it will be the first feature once I get around to fixing any critical bugs I can.


Np, glad its fixed (thanks Neiv for helping pinpoint the commit that introduced the issue!)


Unfortunately I left this too long and the pastebin link has expired (you should be able to choose an unlimited expiry next time if you don't mind, just in case ) I assume you probably don't have the script anymore? Also I think you missed the last line of the error traceback which should include the actual error itself, so I don't think I can do much.


Hmm, that seems very odd. Well probably too late to investigate At least its working for you now


Yep, I'll look into it, thanks! (ticket)


It seems like there is no guarantee of the ordering of the files returned to me on different OS's and OS versions, so I'll probably need to sort them to ensure a specific ordering. (ticket)


Thanks, will look into it (ticket)


Its not very intuitive. I wish I had made things more data driven so its easier to tweak without changing the code (or even coded it better ), and I may look into updating some stuff for that (ticket). Here is a basic outline:

In Libs/AIBIN.py:
1) Add a "descriptive name" to the `labels` list
2) Add a "short name" to the `short_labels`
3) Add its list of parameters to the `parameters`
You must ensure that the names and parameters are at the same index in all 3 lists

In PyAI.pyw:
Add the tooltip help to a section in the `cmds`

This might not be complete, but I can help once/if someone actually does it.


Hmm I thought I fixed that. Thanks, will look into it (ticket)


Will look into it (ticket)


Yeah I can do that (ticket)


I'll have to look into it (ticket)


Yeah I think PyDAT is the biggest offender, so I'll look into it (ticket). If you have the same issue with other programs just report em!


This was an issue trying to show the preview, but for some reason it didn't have the overlay grp properly loaded. I'll look into it (ticket). If it happens again, you can remove PyLO.txt from Settings or edit one of the overlay settings to fix it without needing a redownload.


Thanks for the reports guys, keep em coming, I'll get to them when I can!

Congrats on your resurrection! Glad to see you'll be working on PyMS again. Holiday season is crazy enough without your work life making it even more stressful, so I hope that comes to a close for you soon.

I had a few thoughts as to how some programs could be updated but of course I never wrote any of them down. I'll try and compile a short list and get back to you on that.

I had an issue earlier today where unused blocks caused StarCraft to freeze after the AI was run. Removing the unused blocks or reactivating them solved the issue, but as far as I'm aware the code should just never be called if there is no command to do so. Any idea if this is a PyAI issue? If you need me to test it more, just let me know what steps I should take.

It might not jump properly because apparently try_townpoint sometimes malfunctions without expansion commands? Couldn't confirm it, but it was mentioned in some SCAI thread I found. Not sure if PyAI should be able to find out whether it malfunctions anyway, though, so it's not really relevant :V

The script that crashes is here, I still have it - https://pastebin.com/raw/rayHqUWM
Edit: and the used unitdef.txt is here - https://pastebin.com/raw/FtdpepgA
It's set never to expire now.

Only the Chain/Return part of the script seem to crash/corrupt(?) PyAI? I extracted everything used by Slay into a separate script and it worked fine. (chain/returntoc/slay = labels used by the three jumps below "multirun upgrades")
I'm too tired now to look through the script again, so I'm really sorry if it's some obvious mistake I missed previously! ><

Really weird. Doesn't seem to be the case for me, had multiple unused blocks leftover in the past. Can you try using the label, but just making it completely functionally unreachable (such as with some impossible to fill enemyowns_jump)?

Blizzard theoretically uses two scripts with normally unreachable blocks for xZ6, and they obviously don't seem to crash/freeze anything, though with how good Z6 scripts are I wouldn't be surprised if they actually do freeze sometimes :^)

Post has been edited 3 time(s), last time on Dec 3 2017, 11:09 am by Nekron.

Original script reconstructed (only difference from the original is the attack_to coordinates and attack_add counts/types): https://pastebin.com/raw/AaEv6cXn
This script froze the game within 5 seconds of activating. Removing `emitter1` and `emitter6` blocks resolved the issue.

I don't have time to check your request right now so I'll just look into it later.

Yeah, it does indeed either crash the game (try copy-pasting the last block in PB3A, with a different blockname, but delete the goto -> gamecrash) or freeze it (such as in your case); the workaround I presented also doesn't work. Seems pretty weird, historically I had quite a few scripts that have unused blocks leftover back from the SCAI days I'm pretty sure. Too lazy to find concretes right now, it's unlikely I'll search through everything before xmas but once I do I'll make sure to pop in

What's up with this error? Does PyMPQ not support extracting files from old SEMPQs?

Where are the registry entries for opening .mpq files (and the other ones)? I've tried setting file associations on a fresh install of win10 and it worked initially, but after setting it to winmpq, I can't set it back to pympq.

Trying to use PyMPQ gives me a dependency error:
"PyMS has currently no Mac support for MPQ files, thus this program is useless. If you are a Mac user and can help compile and test SFmpq then please Contact me!"
I don't think I can help you, though (using Ubuntu Linux, with BW v. 1.16 via Wine).

There doesn't appear to be a way to change the overlay frame in PyLO - either that or I'm completely missing it.

EDIT: While we're at it, it would be nice to be able to select the palette used for the overlay as well.

Post has been edited 1 time(s), last time on Jan 11 2018, 11:45 pm by Voyager7456.

PyTILE:

Bug - typing an alphabetical character in the Buildable Raw box crashes the program with this crashlog
Request - ctrl+c and ctrl+v to copy/paste minitile flags (e.g. copy painted walkable flags and paste them between tiles)
Request - shift/ctrl click to select multiple groups to edit group flags (e.g. Buildable, Overlay, Width/Height, Text String)


PyAI:

Bug/Request - comments currently break the script, would be nice to have comment support


PyGRP:

Bug - cmdicons are saved at a weird offset, resulting in some parts of the graphic being hidden in-game


All programs:

Request - confirmation on overwriting default files when saving (some programs have this IIRC, notably PyAI does not have this)

I think I have requested these before but if not, I second this very very badly.

PyTILE:

Request - make height color for middle a deeper orange to better contrast against high (yellow)
Request - double-right click on a megatile in the main window also applies minitile flags just like right clicking in the painter window
Request - hotkeys for selecting height field (e.g. ctrl+shift+r for red, ctrl+shift+o for orange, ctrl+shift+y for yellow)
Request - right-click on painter window with height selected cycles through the different height levels (red + right-click = orange, etc)

Hey guys, back with some more updates.

I have added this to PyDAT on the Units>AI Actions tab. I was not entirely sure about a few things:
1) Since it is based on weapon settings, I assumed there was both a Ground and Air value so calculate both
2) I was not sure what to do if the unit had no weapon, so I defaulted weapon values to 0 (except cooldown as 1 because of the division)
I have no idea if it is actually calculating correctly, so if someone could check and confirm, that would be awesome. And please provide any other feedback you have.

So it turns out the warning was supposed to show up, it was just worded incorrectly. It should have said "everything after this command will not run", because the `wait(1)` after the `goto(ultimate)` will never run. I have fixed the message.

I have update PyMS to always sort when selecting multiple files.

Fixed

Fixed

Fixed

It turned out the logic was invereted, should be fixed now.

Still not sure how this happened, but I put in a check so it shouldn't crash anymore, just won't render the preview.

I'm not really sure what that is unfortunately, it should support exporting from SQMPQs.

The registry entries set by PyMS programs are here. Yeah I don't know what entries WinMPQ is changing but if someone could provide details I would love to fix this. I have create an issue to track this.

That messaging was actually old, PyMPQ does support Mac now, it just doesn't support Linux. I have updated it to be more clear. If you would like to try and compile SFmpq to get MPQ support working for you in PyMS, I would love to help.

I had some weird logic for previewing the overlay. I must admit I don't really remember how .LO* files work, so I was not sure if that logic was bugged or just completely wrong. For now I have removed that logic entirely, so the scrollbar under the preview should now work to allow you to move through the overlay frames.

What are you trying to preview? I have created a general issue for GRP previews in all the programs.


I guess there is a quote limit per post or something, so I have put my answers inline inside the one quote above (highlighted in red)


Thanks for the reports!

Post has been edited 2 time(s), last time on Mar 24 2018, 7:18 pm by poiuy_qwert.

The display looks really nice =)


The ground and air values are separate, if the unit has no weapon, the force is 1, as long as the other force is also 1 or lower, either from the formula giving a very small value or unit having no weapons at all. Otherwise the weaponless value is 0.

(I don't think that the distinction between 0 and 1 matters at all though)

The calculation seems to be pretty much correct, but comparing the values to bw's memory I found some differences:

- The last unit-specific adjustment doesn't round down to an integer (e.g. drone should have 16 flat)
- That specification for unit-specific adjustments had actually a mistake: Inf. terrans and *scourges*, not defilers have 1/16 multiplier. (Defiler's and scourge's unit ids were right next to each other D:)
- Apparently larvae, eggs, cocoons, and lurker eggs are also hardcoded to have 0 value, but this is set before the "no weapon -> 1 force" check. It's a really minor and pointless detail but scarab has 1 air force since it has no air weapon, but those zerg units have 0 force no matter what.
- Carriers and reavers (and hero variants) use interceptor's / scarab's weapon for their strength (but not the child unit's other stats and obviously not their x0 multiplier). Carrier should have 1133 force with default stats and reaver should have 300.
- If the unit has a subunit, its weapon is used for the calculations.

Post has been edited 3 time(s), last time on Mar 24 2018, 9:05 pm by Neiv.

Thanks for all the updates!

No, they break the script and comment out extra stuff. edit: actually they don't seem to do anything anymore, they just get removed.

That works, presumably just ctrl+c with a group selected in the main window and ctrl+v when selecting a different group? Mostly just to paste unbuildable data but obviously useful for other things.

Looks fixed in the latest build after adding the default icons and modified ones. I'll report it again if it comes up.

edit:

PyTILE bug - scrolling in the main window after selecting an item from the minitile painter's dropdown menu also scrolls that menu. Has no effect but is visually distracting and occasionally makes me panic after doing an hour's worth of pathing work and noticing it says 'ramp'.

Post has been edited 3 time(s), last time on Mar 25 2018, 11:33 am by Pr0nogo.

PyTILE:

Just got a crash after scrolling up a lot.