Apprently this had been noticed before, but never got the final fix:
IRC log
```
2021-02-04
ermo: mmm, I must have replaced Ncurses with NcursesW on some Dependencies file (and forgot
to take note of which file I touched)
anyhow, /S/Settings/Scripts/CompatibilityList should have been involved at some point to map
Ncurses -> NcursesW
/System/Settings/Scripts/CompatibilityList 25:Ncurses: NcursesW
paranoidd: roger
did you read my forum post?
I still think the idea I proposed there might have merit in terms of being able to easily introspect
recursive CheckDependencies calls?
just read it. Yeah, that's a good idea! Let me implement it now, one sec
:thumbsup:
paranoidd: also, have you seen the bug I filed re. the kernel package and BIOS-only systems?
on GH
yes, I have. But I didn't have time to look into that one yet
roger
paranoidd: so, /S/S/S/CompatibilityList also has Ncurses: NcursesW here
do you think it might presently be ignored?
It's a pretty decent way of handling temporary renames (until all deps get updated)
it's usually handled by CheckDependencies, but the branch the code is taking when we call it with
--mode=list is probably causing it to skip looking into that file
lemme know when you need me to try again
box is at the ready ;)
ermo: done. Please update your copies of Scripts and Compile
on it
I'll update the thread on the forum to state this change
ah!
on Compile, when we print the resolved dependencies, we're sorting the output alphabetically
let me change that, otherwise this new feature becomes useless
done. Please update Compile again
...
paranoidd: no dice, Ncurses is still ignored...
```
I have written a new recipe
At-Spi2-Core 2.54.0
which is supposed to replaceATK
.Therefore I have added the following to
/S/S/Scripts/CompatibilityList
.In my understanding this is supposed to replace a dependency called
ATK
byAt-Spi2-Core
in a program'sDependency
file.However upon running eg
Compile --debug ATKMM
(which hasATK
in itsDependencies
file),Compile
still sandboxesATK
rather thanAt-Spi2-Core
:Apprently this had been noticed before, but never got the final fix:
IRC log
``` 2021-02-04Just for reference, these are the commits from the day of the discussion: Compile: 6b4c9d1391b752263a241444d6d375f1dab7cf51 3e5865b99c93fa823172c41f1694a36b8d763e9b Scripts: adf1c6b9f9146c1a4f2a85b6a394baed5e791d2f