379c3c1738
PSD: return error if a complete line is read/decompressed
2025-11-17 14:38:07 +01:00
6519d2930e
Update version to 6.21.0
2025-11-14 18:25:25 +01:00
ffcfa49f13
Update dependency version to 6.20.0
2025-11-07 19:26:04 +01:00
d3d4bea2f1
PSD: limit memory usage on corrupted files
2025-11-07 15:36:31 +01:00
cfbeb397a8
Fix assert on broken data
...
Change the isnan test to !isfinite. This way we include infinite numbers
too as not accepted
2025-11-05 23:16:03 +01:00
1c6ce2437e
PSD: minor fixes while reading Image resource Section
2025-11-03 08:27:17 +01:00
3d9f1c4e9e
GIT_SILENT: Bump kf ecm_set_disabled_deprecation_versions. Make sure that it compiles fine without kf 6.19 deprecated methods
2025-11-02 16:07:02 +01:00
e7d01c286f
IFF: fix crash on malformed files
2025-10-31 09:07:40 +01:00
dfda863841
iff: Fix crash on malformed files
...
oss-fuzz/455510771
2025-10-28 00:50:06 +01:00
3d4d74fc02
PSD: improve sections size checks
2025-10-24 08:39:58 +02:00
326e56029a
TGA: ignoring large metadata on dev area
2025-10-23 14:54:52 +02:00
e8e5b091bc
Check device before read
2025-10-23 12:18:59 +02:00
aeb2988eec
Fix possible infinite loop when reading a broken jp2
2025-10-23 11:59:39 +02:00
abfda2ce69
On demand buffer allocation on PCHG decompression
2025-10-23 07:58:58 +00:00
d6d67c7b8b
Fix possible read overflow with malformed data
2025-10-23 08:23:57 +02:00
54c94764d1
Fix wrong palette check
2025-10-21 19:33:12 +02:00
7512874620
Use internal FP32 conversion instead of FP16
2025-10-18 05:43:08 +02:00
146f637385
Fix read test failing on aarch64
2025-10-17 13:43:24 +02:00
436310b4f0
Remove Qt version checks now that 6.8.0 is the minimum required
2025-10-16 20:20:35 +02:00
d8a8bcb68e
GIT_SILENT Upgrade Qt6 version requirement to 6.8.0.
...
See https://community.kde.org/Frameworks/Policies
2025-10-16 19:10:25 +02:00
2b80a0a55a
Fix assert on broken data
...
The nan eventually ends up in qRound inside Qt code.
That asserts because it doesn't know what to do with a nan
2025-10-14 23:55:59 +02:00
8e164474b4
CI: Try newer libjxl for ossfuzz
2025-10-13 11:34:15 +02:00
4c52480b73
tga: Reduce Warning to Debug
...
It happens that apps have an unknown source and ask if they can read it,
we should not show a weird warning to the user
2025-10-12 15:14:41 +02:00
19faa55a6d
Fix crash on malformed files
2025-10-10 09:03:18 +02:00
b9bfeedbc1
Fix crash on malformed files
2025-10-06 01:04:40 +02:00
3c8539d53d
Update version to 6.20.0
2025-10-05 15:16:13 +02:00
d332ad717b
Update dependency version to 6.19.0
2025-10-03 15:04:37 +02:00
1ca7baed98
Fix assert on broken data
...
The nan eventually ends up in qRound inside Qt code.
That asserts because it doesn't know what to do with a nan
2025-10-02 01:02:08 +02:00
f1b0c9f0ec
add const so we know that QImage is not being modified
2025-10-01 19:15:10 +02:00
b83f4c0231
Add color space check during read test
2025-09-27 11:12:50 +02:00
a457e5ddcb
Limits the max RAW size to 300000x300000 pixels
2025-09-21 15:13:17 +02:00
f28cd98661
Limits the max DDS size to 300000x300000 pixels
2025-09-21 11:28:45 +02:00
05c3a1afe6
Fix compilation failure with Qt 6.7
2025-09-20 13:37:52 +08:00
fda751c641
Switch all plugins to QLoggingCategory
2025-09-19 10:00:26 +02:00
a4e18734bd
Resolution calculations performed by functions
...
Added functions for dpi <-> ppm transformations and used in all plugins.
2025-09-17 12:22:44 +02:00
14286a6ab0
Add a CI job for oss-fuzz
2025-09-13 10:15:10 +00:00
945fd6f0ce
GIT_SILENT: Bump kf ecm_set_disabled_deprecation_versions. Make sure that it compiles fine without kf 6.18 deprecated methods
2025-09-13 08:53:25 +02:00
c36b4e2350
Use std::lround instead of qRound
...
qRound will assert if the resulting integer is out of range, by using
lround we can ask if the rounding range failed
2025-09-11 13:50:50 +02:00
95f0d15e14
RAW: Disable broken stream protection
...
It seems that on BSD the definition introduced with MR !384 crashes Telegram.
2025-09-11 11:52:21 +02:00
56c8bc7323
Add checks on the seek return value
2025-09-09 22:20:34 +02:00
08e178f098
Fix Null-dereference READ
2025-09-09 21:15:19 +02:00
6881e3111b
HDR: Limits the header to the first 128 lines
...
The HDR header is a set of text information spread across a few lines.
According to the specifications I have, the parameters seem to be a total of 7 / 8 but there could be more. This patch limits the header to 128 lines of maximum 1024 bytes.
Closes #40
2025-09-09 13:59:14 +00:00
463da81fad
IFF: support for PCHG chunk
...
Highlights:
- Adds support for a new palette changer chunk. Some test cases attached to #38 .
- Fixes the reading of ILBMs with the mask (test case: [cyclone.iff](/uploads/d8734d2155fd0d21f7b003b37e0d1259/cyclone.iff)).
- Adds support for HAM5 encoding.
- Adds more test cases created using [HAM Converter](http://mrsebe.bplaced.net/blog/wordpress/ ).
- Adds support for Atari STE RAST chunk outside FORM one (test case: [fish.iff](/uploads/c461cf4b6a1423cec60fbce645d9fd07/fish.iff)).
NOTE: I contacted Sebastiano Vigna, the author of the PCHG chunk specifications, and he provided me with:
- Some images to test the code (but I can't include them in the test cases).
- Permission to use [his code](https://vigna.di.unimi.it/amiga/PCHGLib.zip ) without restrictions: Huffman decompression was achieved by converting `FastDecomp.a` via AI.
Closes #38
2025-09-08 15:39:50 +00:00
8036b1d032
Fix assert when read corrupted floats
2025-09-08 11:01:06 +02:00
71cc137254
Update version to 6.19.0
2025-09-08 10:23:10 +02:00
7858c4eeec
Update dependency version to 6.18.0
2025-09-07 14:25:38 +02:00
eae41980b2
dds: Fix assert when reading broken data
...
oss-fuzz testcase 6027629841154048
2025-09-06 22:39:16 +00:00
3bf2281610
GIT_SILENT: Use Qt CamelCase includes
2025-09-04 08:32:52 +02:00
59089855fa
Unified maximum pixel value for large image plugins
2025-08-30 09:17:58 +02:00
5a067130af
Removed the conversion of the entire image to a compatible format when saving from the PCX, PIC, and RGB plugins.
...
For example, the PCX plugin converts all RGB images to RGB(A)32. So if you try to save a 1 GiB RGB888 image, it will be converted to RGB32 image so, you need additional 1.25GiB of ram. The conversion now occurs line by line, significantly saving memory.
2025-08-30 09:17:40 +02:00
