Skip to content

Conversation

palfrey
Copy link

@palfrey palfrey commented Sep 7, 2025

Mass upgrade to v3 to fix #950

@Eeems
Copy link
Member

Eeems commented Sep 7, 2025

You chose v3.1 instead of v3.2 as it went to stable, which downgrades some packages. Since you are updating from v2 based images, it wouldn't matter as they would still mainly be upgraded. Let's use v3.2 instead so we can avoid the issues we swapped to stable to resolve.

@palfrey
Copy link
Author

palfrey commented Sep 7, 2025

You chose v3.1 instead of v3.2 as it went to stable, which downgrades some packages. Since you are updating from v2 based images, it wouldn't matter as they would still mainly be upgraded. Let's use v3.2 instead so we can avoid the issues we swapped to stable to resolve.

Done

Copy link
Member

@Eeems Eeems left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

All recipes that were changed will need their pkgver's rel (The number after the -) updated to trigger the package build.

@palfrey
Copy link
Author

palfrey commented Sep 7, 2025

This also fixes #871 as a side-effect...

@Eeems
Copy link
Member

Eeems commented Sep 7, 2025

We will need to get a new version of toltecmk pushed out to include upgrading to to resolve the build issues. It probably is worth just bumping absolutely everything when we do to make sure we didn't introduce any new issues. After we get the other PR merged I'll get the toltecmk version pushed, but it may be a bit. I'm not at home and mostly on mobile or a mobile hotspot

@Eeems
Copy link
Member

Eeems commented Sep 8, 2025

0.3.5 should be released with your changes.

recipe_bundle[arch].packages[pkg_name]
for pkg_name in recipe_bundle[arch].packages
]
logger.info("Building %s", name)
Copy link
Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Made it possible to figure out what package was broken!

@palfrey
Copy link
Author

palfrey commented Sep 8, 2025

Yay! With a small set of exceptions (folly, plato, retris) all of which have comments and links to their bug tickets, we have a working build with all the other 2.x packages on 3.x.

@Eeems
Copy link
Member

Eeems commented Sep 9, 2025

I'll take a closer look and do some testing later this week.

@LinusCDE
Copy link
Member

LinusCDE commented Sep 16, 2025

retris, chessmarkable, doomarkable and now plato all got new releases now (based on new libremarkable. features: 0.6.3 » MSRV bump: 0.7.0), so now they all build fine.

See respective releases (linked above) for more as some had the one or other small improvement added as well.

Also all support the env LIBREMARKABLE_FB_DISFAVOR_INTERNAL_RM2FB=1 now, to not use the builtin rm2fb client on the rM 2.

Should resolve:

@palfrey
Copy link
Author

palfrey commented Sep 18, 2025

Latest Chessmarkable certainly I can't build with v3.2, but that's a matter of the image being out of date! Logs I get are

[   DEBUG] toltec.builder: build(): Downloaded moxcms v0.7.5
[   DEBUG] toltec.builder: build(): error: failed to parse manifest at `/root/.cargo/registry/src/index.crates.io-6f17d22bba15001f/moxcms-0.7.5/Cargo.toml`
[   DEBUG] toltec.builder: build():
[   DEBUG] toltec.builder: build(): Caused by:
[   DEBUG] toltec.builder: build(): feature `edition2024` is required
[   DEBUG] toltec.builder: build():
[   DEBUG] toltec.builder: build(): The package requires the Cargo feature called `edition2024`, but that feature is not stabilized in this version of Cargo (1.83.0-nightly (c1fa840a8 2024-08-29)).
[   DEBUG] toltec.builder: build(): Consider trying a more recent nightly release.
[   DEBUG] toltec.builder: build(): See https://doc.rust-lang.org/nightly/cargo/reference/unstable.html#edition-2024 for more information about the status of this feature.

@palfrey
Copy link
Author

palfrey commented Sep 18, 2025

Same problem for retris and plato

@palfrey
Copy link
Author

palfrey commented Sep 18, 2025

Tried deleting the Cargo.lock for those projects, and downgrading to the minimum versions, but you hit tomaka/hlua#226 (among other problems) when trying that.

@LinusCDE
Copy link
Member

Odd. I was building with the 3.2 image.
But sometimes it seemed to want to update it. No clue how cross does it. Potentially it just bumped rust in that image to latest somehow.

And yeah, seems that going back with hlua would be hard, as before there were compilation errors due to the system library changing some function definitions.

@palfrey
Copy link
Author

palfrey commented Sep 19, 2025

Odd. I was building with the 3.2 image. But sometimes it seemed to want to update it. No clue how cross does it. Potentially it just bumped rust in that image to latest somehow.

And yeah, seems that going back with hlua would be hard, as before there were compilation errors due to the system library changing some function definitions.

toltec-dev/toolchain#43 to get a newer rust with a 3.3 image appears to have sorted all of this out thankfully!

@Eeems
Copy link
Member

Eeems commented Sep 19, 2025

Alright, we have them all building now this will take a while to test as there are quite a few packages, and will require testing on both OS 2.x and 3.x for both rM1 and rM2. If anybody wants to help, feel free to comment with what you've tested. We are only really looking for a quick install/uninstall test along with launching the app to make sure it runs and appears to perform it's basic functions. There is no need for an in depth test, although if you do that, make sure to raise any issues you find on the application's repository, unless it appears to be a packaging issue.

@palfrey
Copy link
Author

palfrey commented Sep 20, 2025

I'm in progress testing things on my rm1 with OS 2.x, and will update this comment as I continue:

  • bash-completion - LGTM
  • bufshot - LGTM
  • calculator - LGTM
  • chessmarkable - LGTM
  • display - LGTM (stub installs fine)
  • doomarkable - LGTM
  • draft - LGTM
  • dumbskull - LGTM
  • evtest - LGTM
  • fbink - LGTM
  • fingerterm - LGTM
  • folly - LGTM
  • fuse - not sure how to test, installs fine
  • genie - LGTM
  • gocryptfs - runs, not really tested
  • harmony - LGTM
  • iago - runs, not really tested
  • keywriter - boots to blank screen (but I'm seeing this with the current version as well)
  • lamp - LGTM
  • libdlib - LGTM
  • libvncserver - LGTM given vnsee works
  • minesweeper - LGTM
  • mmc-utils - seems to run, not sure how tot est
  • nao - LGTM
  • netevent - LGTM
  • netsurf - LGTM
  • plato - LGTM
  • puzzles - LGTM
  • recrossable - LGTM
  • remarkable-splash - LGTM
  • remarkable-stylus - don't have the relevant hardware, installs fine
  • remux - LGTM
  • restream - LGTM
  • retris - LGTM
  • rmfakecloud-proxy - LGTM
  • rmservewacominput - boots, LGTM, not really tested
  • rpncalc - LGTM
  • sas - LGTM
  • templatectl
    Installing this gets the following and it fails to install
Configuring templatectl.
//opt/lib/opkg/info/templatectl.postinst: line 998: warning: here-document at line 21 delimited by end-of-file (wanted `UNIT')
//opt/lib/opkg/info/templatectl.postinst: line 999: syntax error: unexpected end of file

This appears to be in add-bind-mount with the UNIT bit having been indented for some reason?
Not happening on the previous version

  • vnsee - LGTM
  • wikipedia - LGTM
  • wordlet - LGTM
  • zshelf - LGTM

So with the exception of templatectl having a weirdly borked postinst, all good here.

@Eeems
Copy link
Member

Eeems commented Sep 20, 2025

display - is this rm1 relevant?

It's basically a stub on the rM1, but still relevant to test that it installs properly

@palfrey
Copy link
Author

palfrey commented Sep 20, 2025

templatectl issue is a build bug, see toltec-dev/build#60

@palfrey
Copy link
Author

palfrey commented Sep 20, 2025

templatectl install works all fine now

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

3 participants