I like a good jaunt with IDApro as much as the next RE, but my question is what does ASIF do that Qcow2 doesn't?
My other question is why does it take so long to copy an app out of a dmg and into /Applications. Like, just change some pointers to pointers to data on disk and shit.
Additionally, while I don't know much about APFS, I don't think it would be beneficial to point the extracted app to blocks that are also part of the dmg file, i.e. some copying has to happen anyway.
A perfect theoretical filesystem can still have subjective user configurable choices though right? Like case sensitivity, UTF normalization, checksum hash function, extra copies of data/metadata to store for redundancy/healing, etc (as well as compression/encryption). I think ZFS is a pretty strong real world example of a CoW FS, but you can still set a lot of different properties between sub-fs and then need to copy when you go between them to get the structural changes.
Disk images are supposed to function as if they're attached storage I think, and have different properties from what FS you're running on boot or your home folder (which themselves can be different, I run my home folder on my main Mac off a NAS via iSCSI). I'm not sure any underlying FS would avoid a copy operation there in general?
Of course, but that works only if the files are already in the same partition. A dmg is a virtual image, even if it's stored in the same partition, once mounted it acts like another partition.
Speaking of images (media file), I found out that iCloud double-counts storage space for rotated photos, so I need to subscribe to a higher tier just because the iPhone often can't get orientation right, especially for food photos.
Would be nice if they moved to storing diff sidecars.
Small self-advertisement: as an alternative to dissect.cstruct, a fun side-project of mine (C parser + C interpreter in Python) can do a very similar thing:
I have to admit that using C syntax as a string to parse something from Python is definitely a choice. I'm not even sure I would use C structs to lay things out in C…
I asked an LLM to rewrite it for me using the Python built-in struct module, and it gave me this:
import sys
import struct
from collections import namedtuple
# Bake the layout once into a reusable, precompiled object.
HEADER = struct.Struct(">4sIIIQQ16sQQIIII")
# struct only knows positions, not names — pair it with a namedtuple
# to recover the named-field access that cstruct gives you for free.
Header = namedtuple("Header", [
"magic", "field4", "field8", "fieldC",
"field10", "field18", "field20",
"field30", "field38",
"field40", "field44", "field48", "field4C",
])
with open(sys.argv[1], "rb") as fh:
header = Header._make(HEADER.unpack(fh.read(HEADER.size)))
print(header)
To me, this seems significantly less readable... less Pythonic, even. The printed output is also less readable.
No I think Python's struct module is also really bad. My point is if you are making a new DSL for laying out arbitrary formats why not do something better than what we have
Author here, this is a valid point but there are also valid reasons to choose C structures.
The larger framework that this is a part of is primarily targeted towards people working in cybersecurity, not software engineers. Cybersecurity people are very often not great software engineers and there is a high throughput of “throwaway” scripts, or “make a quick hacky change”. C is commonly already well understood, a bespoke DSL usually is not and requires a learning step. You can “hit the ground running”, so to say.
And, as a bonus, creating, say, a filesystem implementation is now often as easy as copy/pasting existing C structure definitions, either from the original source (which is usually C) or from reversing tools such as IDA/Ghidra.
There’s no right or wrong way in my opinion, just preferences.
I became a fan of Kaitai Struct [0] when doing some amateur sleuthing last year. It has a web-based IDE [1] for writing and testing structure definitions against hex dumps, and can generate binary parsers in Python (and many other languages) right from the Web IDE.
I would assume dissect.cstruct was written for interopt with c programs using C structs, or to use formats documented as C structs. Not as a greenfield tool for arbitrary formats.
C structs seem less bad than python structs, so why not use them? Especially why write a struct parser and create a DSL for it, when there's already one that you can use that uses a well known DSL you might already understand.
OK so what's your alternative then? It's easy to say you don't like something but the onus is on to show there's something actually better.
The library used in the author's post seems perfectly readable to me, enough that it didn't even register until I read your comment. Could it be tweaked slightly to not use C syntax? Sure, but it's still going to need roughly the same pattern of identifier + type (including size). Types in C are straightforward so long as you don't have functions/pointers (which have the "inside out" problem, but they're not needed for binary encodings), so you're going to be looking at pretty trivial changes to syntax. Certainly not enough to warrant this level of quibbling.
I still think it proves my point: your original objection was about the syntax being C-like and, as I predicted, the differences in syntax in your idea (where the type goes, colon vs positional, etc.) are all trivialities that don't affect usability.
What's better about your idea is that it's actual Python code rather than being embedded in a string. Maybe that was your point originally and I misunderstood.
Yeah I don't think this is too unusual, I'm not surprised that someone else has come up with it too. My point is that if you write it in Python you can get things like type hints, and use language features to make things nicer. Obviously there isn't much you can do to skip the part where you have to list fields and their types but using a string of C syntax seems kind of bad in any language that's not C work working with it directly.
But yeah, while I think the `cstruct` helper function to describe a binary data layout in Python is more elegant than the builtin alternatives, it would have been much less painful to just go with a minimal C command line program (or any other programming language where a struct directly maps to memory). Python and most other scripting languages have been built for manipulating text data, but suck when working with binary data.
Worth noting ASIF's compression tradeoff also affects Spotlight indexing —
since the content is opaque until mounted, you lose searchability on
unmounted disk images that you'd get with a regular folder structure.
More info on how ASIF differs from the decades-old sparseimage format: https://news.ycombinator.com/item?id=44259132
I like a good jaunt with IDApro as much as the next RE, but my question is what does ASIF do that Qcow2 doesn't?
My other question is why does it take so long to copy an app out of a dmg and into /Applications. Like, just change some pointers to pointers to data on disk and shit.
> what does ASIF do that Qcow2 doesn't
Mount natively in macOS
> why does it take so long to copy [...] out of a dmg
Compression mostly. DMG contents can optionally be compressed using zlib, lzfse, or slow as molasses bzip2.
Also Gatekeeper.
Additionally, while I don't know much about APFS, I don't think it would be beneficial to point the extracted app to blocks that are also part of the dmg file, i.e. some copying has to happen anyway.
in a perfect theoretical filesystem, copy-on-write means copying is as cheap as moving a file, though uncompressing time makes sense.
A perfect theoretical filesystem can still have subjective user configurable choices though right? Like case sensitivity, UTF normalization, checksum hash function, extra copies of data/metadata to store for redundancy/healing, etc (as well as compression/encryption). I think ZFS is a pretty strong real world example of a CoW FS, but you can still set a lot of different properties between sub-fs and then need to copy when you go between them to get the structural changes.
Disk images are supposed to function as if they're attached storage I think, and have different properties from what FS you're running on boot or your home folder (which themselves can be different, I run my home folder on my main Mac off a NAS via iSCSI). I'm not sure any underlying FS would avoid a copy operation there in general?
Of course, but that works only if the files are already in the same partition. A dmg is a virtual image, even if it's stored in the same partition, once mounted it acts like another partition.
Image as in “filesystem snapshot” not as in “media file”.
Speaking of images (media file), I found out that iCloud double-counts storage space for rotated photos, so I need to subscribe to a higher tier just because the iPhone often can't get orientation right, especially for food photos.
Would be nice if they moved to storing diff sidecars.
Small self-advertisement: as an alternative to dissect.cstruct, a fun side-project of mine (C parser + C interpreter in Python) can do a very similar thing:
https://github.com/albertz/PyCParser/blob/master/demos/disse...
I have to admit that using C syntax as a string to parse something from Python is definitely a choice. I'm not even sure I would use C structs to lay things out in C…
I asked an LLM to rewrite it for me using the Python built-in struct module, and it gave me this:
To me, this seems significantly less readable... less Pythonic, even. The printed output is also less readable.No I think Python's struct module is also really bad. My point is if you are making a new DSL for laying out arbitrary formats why not do something better than what we have
Author here, this is a valid point but there are also valid reasons to choose C structures. The larger framework that this is a part of is primarily targeted towards people working in cybersecurity, not software engineers. Cybersecurity people are very often not great software engineers and there is a high throughput of “throwaway” scripts, or “make a quick hacky change”. C is commonly already well understood, a bespoke DSL usually is not and requires a learning step. You can “hit the ground running”, so to say.
And, as a bonus, creating, say, a filesystem implementation is now often as easy as copy/pasting existing C structure definitions, either from the original source (which is usually C) or from reversing tools such as IDA/Ghidra.
There’s no right or wrong way in my opinion, just preferences.
I became a fan of Kaitai Struct [0] when doing some amateur sleuthing last year. It has a web-based IDE [1] for writing and testing structure definitions against hex dumps, and can generate binary parsers in Python (and many other languages) right from the Web IDE.
[0] https://doc.kaitai.io/user_guide.html
[1] https://ide.kaitai.io/devel/
I would assume dissect.cstruct was written for interopt with c programs using C structs, or to use formats documented as C structs. Not as a greenfield tool for arbitrary formats.
C structs seem less bad than python structs, so why not use them? Especially why write a struct parser and create a DSL for it, when there's already one that you can use that uses a well known DSL you might already understand.
OK so what's your alternative then? It's easy to say you don't like something but the onus is on to show there's something actually better.
The library used in the author's post seems perfectly readable to me, enough that it didn't even register until I read your comment. Could it be tweaked slightly to not use C syntax? Sure, but it's still going to need roughly the same pattern of identifier + type (including size). Types in C are straightforward so long as you don't have functions/pointers (which have the "inside out" problem, but they're not needed for binary encodings), so you're going to be looking at pretty trivial changes to syntax. Certainly not enough to warrant this level of quibbling.
idk just spitballing I would maybe do something like
I'll admit I do really like that.
I still think it proves my point: your original objection was about the syntax being C-like and, as I predicted, the differences in syntax in your idea (where the type goes, colon vs positional, etc.) are all trivialities that don't affect usability.
What's better about your idea is that it's actual Python code rather than being embedded in a string. Maybe that was your point originally and I misunderstood.
Looks like this package works like this: https://harrymander.xyz/dataclasses-struct/
Yeah I don't think this is too unusual, I'm not surprised that someone else has come up with it too. My point is that if you write it in Python you can get things like type hints, and use language features to make things nicer. Obviously there isn't much you can do to skip the part where you have to list fields and their types but using a string of C syntax seems kind of bad in any language that's not C work working with it directly.
in video/image space most code we deal with day to day is still C, lots more rust plugins in gstreamer ecosystem, but 90%+ still C
The article is about disk images though :)
But yeah, while I think the `cstruct` helper function to describe a binary data layout in Python is more elegant than the builtin alternatives, it would have been much less painful to just go with a minimal C command line program (or any other programming language where a struct directly maps to memory). Python and most other scripting languages have been built for manipulating text data, but suck when working with binary data.
Sure, but it's a greenfield Python script
Worth noting ASIF's compression tradeoff also affects Spotlight indexing — since the content is opaque until mounted, you lose searchability on unmounted disk images that you'd get with a regular folder structure.