Project

General

Profile

Bug #12494

Buffer overflow read in BitStream | [temperature] failed to read camera while balance information

Added by Const A 5 months ago. Updated 5 months ago.

Status:
Closed: invalid
Priority:
Low
Assignee:
-
Category:
Darkroom
Target version:
Start date:
12/25/2018
Due date:
% Done:

0%

Affected Version:
git master branch
System:
Ubuntu
bitness:
64-bit
hardware architecture:
amd64/x86

Description

I have a problem with new darktable 2.6.0 and nikon d800 files.

When I open a file from d800 I get a message "[temperature] failed to read camera while balance information...". I've already tried to upload the image into pixls.us twice and both times there were no errors and no the file appeared on the site.
darktable_2.7.0_git3.802f28438_amd64.deb [rawspeed] (_DSC5199.NEF) rawspeed::RawImage rawspeed::RawDecoder::decodeRaw(), line 276: void rawspeed::BitStream<Tag, Cache>::fillSafe() [with Tag = rawspeed::MSBBitPumpTag; Cache = rawspeed::BitStreamCacheRightInLeftOut], line 134: Buffer overflow read in BitStream
The file is attached.
How to reproduce:
1) open the file attached
2) switch to darktable mode

actual: get the error
```[rawspeed] (_DSC5199.NEF) rawspeed::RawImage rawspeed::RawDecoder::decodeRaw(), line 276: void rawspeed::BitStream<Tag, Cache>::fillSafe() [with Tag = rawspeed::MSBBitPumpTag; Cache = rawspeed::BitStreamCacheRightInLeftOut], line 134: Buffer overflow read in BitStream
[temperature] failed to read camera white balance information from `_DSC5199.NEF'!
```
expected: I can open the file and can process it in the darktable mode

_DSC5199.NEF.pp3 - the rawtherapee pp3 file (10.7 KB) Const A, 12/25/2018 01:19 PM

_DSC5199.NEF.7z.002 - NEF part 1 (19.6 MB) Const A, 12/25/2018 01:22 PM

_DSC5199.NEF.7z.001 - NEF part 2 (25 MB) Const A, 12/25/2018 01:22 PM


Related issues

Duplicates darktable - Bug #12208: [2/3] Regression: huffmantable is bad for last decode Fixed 05/19/2018

History

#1 Updated by Roman Lebedev 5 months ago

Const A wrote:

I have a problem with new darktable 2.6.0 and nikon d800 files.

Was it previously modified in nikon software, or digikam?

When I open a file from d800 I get a message "[temperature] failed to read camera while balance information...". I've already tried to upload the image into pixls.us twice and both times there were no errors and no the file appeared on the site.
darktable_2.7.0_git3.802f28438_amd64.deb [rawspeed] (_DSC5199.NEF) rawspeed::RawImage rawspeed::RawDecoder::decodeRaw(), line 276: void rawspeed::BitStream<Tag, Cache>::fillSafe() [with Tag = rawspeed::MSBBitPumpTag; Cache = rawspeed::BitStreamCacheRightInLeftOut], line 134: Buffer overflow read in BitStream
The file is attached.
How to reproduce:
1) open the file attached
2) switch to darktable mode

actual: get the error
```[rawspeed] (_DSC5199.NEF) rawspeed::RawImage rawspeed::RawDecoder::decodeRaw(), line 276: void rawspeed::BitStream<Tag, Cache>::fillSafe() [with Tag = rawspeed::MSBBitPumpTag; Cache = rawspeed::BitStreamCacheRightInLeftOut], line 134: Buffer overflow read in BitStream
[temperature] failed to read camera white balance information from `_DSC5199.NEF'!
```

It seems you didn't attach it here.

expected: I can open the file and can process it in the darktable mode

#2 Updated by Const A 5 months ago

Roman Lebedev wrote:

Const A wrote:

I have a problem with new darktable 2.6.0 and nikon d800 files.

Was it previously modified in nikon software, or digikam?

it shouldn't, I've checked with rawtherapee - it works.

When I open a file from d800 I get a message "[temperature] failed to read camera while balance information...". I've already tried to upload the image into pixls.us twice and both times there were no errors and no the file appeared on the site.
darktable_2.7.0_git3.802f28438_amd64.deb [rawspeed] (_DSC5199.NEF) rawspeed::RawImage rawspeed::RawDecoder::decodeRaw(), line 276: void rawspeed::BitStream<Tag, Cache>::fillSafe() [with Tag = rawspeed::MSBBitPumpTag; Cache = rawspeed::BitStreamCacheRightInLeftOut], line 134: Buffer overflow read in BitStream
The file is attached.
How to reproduce:
1) open the file attached
2) switch to darktable mode

actual: get the error
```[rawspeed] (_DSC5199.NEF) rawspeed::RawImage rawspeed::RawDecoder::decodeRaw(), line 276: void rawspeed::BitStream<Tag, Cache>::fillSafe() [with Tag = rawspeed::MSBBitPumpTag; Cache = rawspeed::BitStreamCacheRightInLeftOut], line 134: Buffer overflow read in BitStream
[temperature] failed to read camera white balance information from `_DSC5199.NEF'!
```

It seems you didn't attach it here.

expected: I can open the file and can process it in the darktable mode

Could you check the attached file please, I've uploaded it again.

#3 Updated by Const A 5 months ago

_DSC5199.NEF.pp3

#4 Updated by Const A 5 months ago

the NEF file
  1. d1d68380623afc3349f97ab98b75c8e1 _DSC5199.NEF
    7zr x _DSC5199.NEF.7z.001

#5 Updated by Roman Lebedev 5 months ago

  • Duplicates Bug #12208: [2/3] Regression: huffmantable is bad for last decode added

#6 Updated by Roman Lebedev 5 months ago

  • Status changed from New to Duplicate
  • Affected Version changed from 1.6.0 to git master branch

Ah, yes, that bug remains.

#7 Updated by Const A 5 months ago

Roman Lebedev wrote:

Ah, yes, that bug remains.

Could you send me a git branch url after when it be fixed, I'll try to build it locally and check on the several broken NEF's I have.

And it seems that the old versions of Darktable ( less than 1.2 ) is not affected by the bug.

#8 Updated by Roman Lebedev 5 months ago

  • Status changed from Duplicate to Closed: invalid

Const A wrote:

Roman Lebedev wrote:

Ah, yes, that bug remains.

Actually no, that really does not look like that bug.

In this file it happens not for the last pixel, but ~200 (!) pixels before that.
So best case scenario would be that it needs at least 200 bits to decode (in reality - more),
for which it would need to read at least ~25 bytes.

So while i can not comment on other software, or older darktable versions,
this very strongly looks like corrupted and/or truncated raw file.
Sorry. Perhaps you have some older copy of this file in backups?

Could you send me a git branch url after when it be fixed, I'll try to build it locally and check on the several broken NEF's I have.

And it seems that the old versions of Darktable ( less than 1.2 ) is not affected by the bug.

#9 Updated by Const A 5 months ago

Hi.

Actually no, that really does not look like that bug.

I've checked the files and you are completely right. The problem is the broken file. I've made a test using two cards in the same time and found that one of the card contains corrupted files but for second card was all OK. So it's my fault and it was hardware problem.
Thank you so much for your help and support.

I hope the ticket should be closed.

Roman Lebedev wrote:

Const A wrote:

Roman Lebedev wrote:

Ah, yes, that bug remains.

Actually no, that really does not look like that bug.

In this file it happens not for the last pixel, but ~200 (!) pixels before that.
So best case scenario would be that it needs at least 200 bits to decode (in reality - more),
for which it would need to read at least ~25 bytes.

So while i can not comment on other software, or older darktable versions,
this very strongly looks like corrupted and/or truncated raw file.
Sorry. Perhaps you have some older copy of this file in backups?

Could you send me a git branch url after when it be fixed, I'll try to build it locally and check on the several broken NEF's I have.

And it seems that the old versions of Darktable ( less than 1.2 ) is not affected by the bug.

Also available in: Atom PDF