Bug #3942
Corrupt encoding of container version 2 GRF
| Status: | Closed | Start date: | 2012-04-26 | |
|---|---|---|---|---|
| Priority: | Normal | Due date: | ||
| Assignee: | - | % Done: | 100% |
|
| Category: | - | |||
| Target version: | - |
Description
Related to #3912. One revision later, in which the only change was to pass GRFCodec the "-g 2" argument for inclusion of 32 bpp sprites, the sprites in the GRF have become completely mangled (see the screenshot). Decoding the attached GRF, compiled with r930, gives me this output and produces two zero-byte PNG files.
grfcodec -d britrains.grf -h 5000 -p 1
Found grf container version 2
Decoding:
Sprite 0 at F, 0% done
Error: not enough data to perform regular decoding for sprite 379
Associated revisions
Fix #3942: Decoding of regular encoded 32bpp sprites failed.
History
#1 Updated by oberhuemer about 1 year ago
- File Unnamed, 7th Jan 20000.png added
- File britrains.grf added
NML does the better job yet again.
#2 Updated by oberhuemer about 1 year ago
GRFCodec shows the same decoding errors and produces the same empty PNGs even with the latter NewGRF.
#3 Updated by oberhuemer about 1 year ago
The same error appears when decoding OpenGFX+ Trains r448, with only the sprite number changed to 346. I hope this doesn't mean GRFCodec's support for decoding (and encoding) container version 2 is completely broken.
#4 Updated by frosch about 1 year ago
- Status changed from New to Closed
- % Done changed from 0 to 100
Applied in changeset 705e415b886b.