Port of Simon Tatham's Portable Puzzle Collection

[__builtin]

Finally got around to it...

Source is up on Gerrit, all 120,000 lines of it: http://gerrit.rockbox.org/r/1421/

Testers welcome, builds available on request.

[chrisjj]

Testers welcome, builds available on request.

Build for Creative ZEN, please.

[chrisjj]

Build for Creative ZEN, please.

FTR, got it, and discussion is here:

https://www.rockbox.org/irc/log-20161216
https://www.rockbox.org/irc/log-20161217

[__builtin]

Just an update here, the majority of the puzzles have made it into the mainline repository.

I have some concerns, however, about some of the more memory-intensive puzzles running out of memory at runtime on some of the targets with smaller plugin buffers. These targets are the Sansa c200v2 and c100. If you have one of these targets, I'd appreciate it if you could test the following puzzles on them, at various settings (listed by order of increasing size):

Code: [Select]

92 sgt-bridges.rock
92 sgt-galaxies.rock
92 sgt-tracks.rock
96 sgt-singles.rock
100 sgt-net.rock
132 sgt-pearl.rock

[tdb]

Hi - gave it a try on my c200v2

Code: [Select]

92 sgt-bridges.rock
92 sgt-galaxies.rock
92 sgt-tracks.rock
96 sgt-singles.rock
100 sgt-net.rock
132 sgt-pearl.rock

They all give me a fatal error due to low memory - but not only these. After a while most games i tried gave a fatal error.

[__builtin]

Ok, thanks tdb. I'll put together a patch in the coming days that will hopefully "solve" the issue to an extent, though it will break playback in the games in some cases.

[saratoga]

c200v2 and clipv1 have only 2 MB of RAM, and the c200v2 must use a color framebuffer, so memory is very tight.

If you want to test low memory, you could configure a device build with less RAM, or use the SIM and further decrease memory a little (to be safe since the SIM isn't quite the same with allocations).

[__builtin]

Just pushed the fix. Could you please test the newest autobuild and tell me if it works now?

Commit: https://git.rockbox.org/?p=rockbox.git;a=commit;h=637c7414a91c33e2627f99fee7b4c9cbf04abdb4

[tdb]

No errors on my c200 Every game starts and seems to work.   
Using version d787191-170204

A couple of fun and intuitive games but some of them require reading a manual (most games I just started and pressed a few buttons to see if they worked). Screen of the c200 is pretty small and the playing field usually takes up even less than half of the screen so not all games can be played comfortably.
The cursor/pointer is nearly invisible so that doesn't help either

 

 

[__builtin]

Added a new feature for dynamic font sizing in commit e7a35ba. You'll need to extract this zip file to /.rockbox/fonts/: http://download.rockbox.org/useful/sgt-fonts.zip. It lets the puzzle choose among differently-sized fonts at runtime. Without it, the puzzles will fall back to the old method of just hoping whatever UI font was set is of a reasonable size.

[codyteam]

Can i use this on my sansa clip+?

[__builtin]

Can i use this on my sansa clip+?

No, it only supports color displays. You might be able to make it work with greylib, though.

[Rick314]

I am a fan of Tatham's puzzle collection and use it on my Android phone and tablet regularly.  I've also been a Rockbox fan since 2008.  Thank you for your efforts to port the puzzle collection to Rockbox!  But on my SanDisk Sansa c250 v1 running Rockbox v3.14 many of the puzzles are unusable.  For example, see the attached images for the simplest possible Bridges (7x7) and Net (5x5) puzzles.  Solving the Bridges image required using a 5x eye loupe and decoding the numbers in the circles as each was selected, like the "2" that can be seen in the top-center of the image.  It would help to use the whole display area, or ideally add a "zoom" feature so only part of the puzzle is seen and the view window can be panned to see different areas of the puzzle.  But again, thanks for efforts so far.

Edit: I added images of what Bridges looks like on an Android tablet, for a 7x7 and a 15x15 game. Perhaps some of these puzzles just aren't appropriate to include on limited-resolution displays.

[__builtin]

Unfortunately the puzzles were written for PCs in mind, and are simply unsuitable for the extremely small screens we support (c200 has 112x80, I think). As for the zoom feature, it's certainly feasible and I will look into it more when I have time.

EDIT: the rockbox port is constantly being updated with changes from the upstream source, try the dev build to see if those changes improve anything (http://git.tartarus.org/?p=simon/puzzles.git;a=commit;h=84d3fd2bd8d3bd5db8cf61a134a4601307b9b7ea might help with Net, for example).

EDIT 2: looked further into the zoom idea, turns out it would require a complete rewrite of the drawing code.

[Rick314]

Unfortunately the puzzles were written for PCs in mind, and are simply unsuitable for the extremely small screens we support (c200 has 112x80, I think). As for the zoom feature ... turns out it would require a complete rewrite of the drawing code.

I understand and have an alternate suggestion then.  For each puzzle identify a Minimum System Requirement for display resolution and maybe other specs (mouse, track wheel, shift key, etc.).  Use conditional compilation to make it so that if a given Rockbox device doesn't meet the MSR, that puzzle doesn't appear to the user.  Games that can't realistically be played at even their simplest settings won't show up to users.  If in the future drawing code is changed, the Minimum System Requirement changes and more puzzles magically appear on devices where they weren't available earlier.

I think the SanDisk Sansa c200 series has a 132Wx80H pixel display, but the 80 is what matters and yes it is very limiting.  (I have a few pictures and videos on mine, just to appreciate what can be done on such a compact device. But they aren't very attractive.)