Signed-off-by: Mattia Rizzolo <mattia@debian.org>

Signed-off-by: Mattia Rizzolo <mattia@debian.org>

Signed-off-by: Mattia Rizzolo <mattia@debian.org>

comparators.java: properly fall back from procyon to javap also when procyon exists but doesn't return any output

Signed-off-by: Mattia Rizzolo <mattia@debian.org>

Gbp-Dch: ignore

Gbp-Dch: Ignore
Signed-off-by: Mattia Rizzolo <mattia@debian.org>

MR: !13


Signed-off-by: Mattia Rizzolo <mattia@debian.org>

Only test[12].fat12 are actual FAT images, test1.fat{16,32} are
truncated to 200 bytes, just enough to pass identification test.

Signed-off-by: Marek Marczykowski-Górecki <marmarek@invisiblethingslab.com>

Signed-off-by: Marek Marczykowski-Górecki <marmarek@invisiblethingslab.com>

Keep the limit on [amd64] to avoid trying to install it in debian...

Gbp-Dch: Ignore
Signed-off-by: Mattia Rizzolo <mattia@debian.org>

Signed-off-by: Mattia Rizzolo <mattia@debian.org>

Temporarily drop Build-Depends and Test-Depends on apktool and oggvideotools, as they are not available in buster at the moment.

Signed-off-by: Mattia Rizzolo <mattia@debian.org>

Signed-off-by: Mattia Rizzolo <mattia@debian.org>

Move the list of optional python packages into setup.py, and have dh_python3 check the generated requires.txt

Signed-off-by: Mattia Rizzolo <mattia@debian.org>

This package Provides:imagemagick, but doesn't ship the "compare" tool.

See: #910652

Signed-off-by: Mattia Rizzolo <mattia@debian.org>

I couldn't find any version of procyon in the archive where the binary
was called procyon-decompiler; as a result this comparator was probably
never run.

Gbp-Dch: Short
Signed-off-by: Mattia Rizzolo <mattia@debian.org>

Thanks: Marek Marczykowski-Górecki for making me notice they weren't actually running.
Signed-off-by: Mattia Rizzolo <mattia@debian.org>

Signed-off-by: Mattia Rizzolo <mattia@debian.org>

comparators.macho: Use `str.startswith(...)` over `str.index(...) == 0` to prevent a traceback if the filename cannot be found on the line. (Closes: #910540)

Signed-off-by: Mattia Rizzolo <mattia@debian.org>

This reverts both 344ae4d6 and
ef5ae496. The mess was caused by me not
actually pushing my change causing duplicated work by Mattia. Mea culpa...

This reverts commit 255cbe59 - there was a
race-condition in fixing this which, I think, stemmed from the fact I didn't
push my fix at the time.

Gbp-Dch: Ignore
Signed-off-by: Mattia Rizzolo <mattia@debian.org>

Gbp-Dch: Ignore
Signed-off-by: Mattia Rizzolo <mattia@debian.org>

the previous jobs failed to install fuse.  some debugging showed that
the reasoon was that /dev is read-only and therefore a chmod(1) call
from fuse's postinst failed.

Gbp-Dch: Ignore
Signed-off-by: Mattia Rizzolo <mattia@debian.org>

Signed-off-by: Mattia Rizzolo <mattia@debian.org>

stretch only has the former

Signed-off-by: Mattia Rizzolo <mattia@debian.org>

Signed-off-by: Mattia Rizzolo <mattia@debian.org>

Signed-off-by: Mattia Rizzolo <mattia@debian.org>

Marek Marczykowski-Górecki authored Sep 18, 2018 and Chris Lamb committed Sep 19, 2018

Non-text files have even less change being json files, so this fuzzy early
elimination is even more useful for them.  In particular, it avoids loading
large files (like ISO images) just to discover it isn't a JSON format.

Signed-off-by: Chris Lamb <lamby@debian.org>

Signed-off-by: Mattia Rizzolo <mattia@debian.org>

MR: !9


Signed-off-by: Mattia Rizzolo <mattia@debian.org>

Signed-off-by: Mattia Rizzolo <mattia@debian.org>

The `cd-iccdump` command has no --version command and, whilst the `colord`
binary does, it relies on the daemon running to return any output.  Therefore
we hackily compare the output (via the line length) of colord 1.4.3:

    "  Profile ID    = 0477fa4bb5ae5ae9a778f5cd72eb45a4"

... versus, for example, colord 1.3.3:

    " Profile ID    = 0x0477fa4b"

We don't massage the output (say, in an `Iccdump.filter` method) as it would
remove the accuracy and, unfortunately, colord 1.4.3 also removes a
somewhat-arbitrary newline too.