Fails to navigate zip/jar/war archives while extensions file intend to allow it

[azazar]

MC fails to open/navigate into ZIP, JAR, and WAR archives that lack the corresponding file extension (.zip, .jar, .war), even though mc.ext.ini contains rules ([zip-by-type], [jar-war-by-type]) intended to handle such files by content-type detection using the file utility.

Cause

1. mc.ext.ini contents do not match actual output of the file utility and should be completely revised.
2. file returns error message when called by MC.

$ file -z -b -S -L test
ERROR:[gzip: Stdin has more than one entry--rest igno] (data)

[zyv]

1. mc.ext.ini contents do not match actual output of the file utility and should be completely revised.

The entries are created (mainly) on the basis of what file does on Ubuntu. However, different systems might have different file utilities with different versions of the database installed. So "should completely revised" is just LLM rubbish. Of course, we'd accept changes backed up by examples. You didn't use our templates, though, so there is no way to know which version of mc you are using and on which operating system.

[azazar]

No. It's not rubbish. It is safe to assume that updates to file utility get into most Linux distros with little delay. But anyway. That's quite easy to check. LLM made a script to compare it's output across various distros. Results are below.

file MC-like output matrix across distros

Generated: 2026-04-27 14:45:39 UTC

Distro	Image	file version	file -v output	binary.bin (file -b -S -L)	json.json (file -b -S -L)	multi-member-no-ext (file -b -S -L)	multi-member.gz (file -b -S -L)	plain.txt (file -b -S -L)	sample-jar-no-ext (file -b -S -L)	sample-war-no-ext (file -b -S -L)	sample-zip-no-ext (file -b -S -L)	sample.jar (file -b -S -L)	sample.tar (file -b -S -L)	sample.tgz (file -b -S -L)	sample.war (file -b -S -L)	sample.zip (file -b -S -L)
almalinux-10	almalinux:10	file-5.45	file-5.45 magic file from /etc/magic:/usr/share/misc/magic	Targa image data - RGBA (1027-1541) 3340 x 3854 x 16 +2312 +2826 - 1-bit alpha - right	JSON text data	gzip compressed data, last modified: Mon Apr 27 14:29:49 2026, max compression, original size modulo 2^32 9	gzip compressed data, last modified: Mon Apr 27 14:29:49 2026, max compression, original size modulo 2^32 9	ASCII text	Java archive data (JAR)	Java archive data (JAR)	Zip archive data, at least v2.0 to extract, compression method=deflate	Java archive data (JAR)	POSIX tar archive	gzip compressed data, was "sample.tgz", last modified: Mon Apr 27 14:29:49 2026, max compression, original size modulo 2^32 10240	Java archive data (JAR)	Zip archive data, at least v2.0 to extract, compression method=deflate
almalinux-9	almalinux:9	file-5.39	file-5.39 magic file from /etc/magic:/usr/share/misc/magic	data	JSON data	gzip compressed data, last modified: Mon Apr 27 14:29:49 2026, max compression, original size modulo 2^32 9	gzip compressed data, last modified: Mon Apr 27 14:29:49 2026, max compression, original size modulo 2^32 9	ASCII text	Java archive data (JAR)	Java archive data (JAR)	Zip archive data, at least v2.0 to extract	Java archive data (JAR)	POSIX tar archive	gzip compressed data, was "sample.tgz", last modified: Mon Apr 27 14:29:49 2026, max compression, original size modulo 2^32 10240	Java archive data (JAR)	Zip archive data, at least v2.0 to extract
amazonlinux-2023	amazonlinux:2023	file-5.39	file-5.39 magic file from /etc/magic:/usr/share/misc/magic	data	JSON data	gzip compressed data, last modified: Mon Apr 27 14:29:49 2026, max compression, original size modulo 2^32 9	gzip compressed data, last modified: Mon Apr 27 14:29:49 2026, max compression, original size modulo 2^32 9	ASCII text	Java archive data (JAR)	Java archive data (JAR)	Zip archive data, at least v2.0 to extract	Java archive data (JAR)	POSIX tar archive	gzip compressed data, was "sample.tgz", last modified: Mon Apr 27 14:29:49 2026, max compression, original size modulo 2^32 10240	Java archive data (JAR)	Zip archive data, at least v2.0 to extract
debian-11	debian:11-slim	file-5.39	file-5.39 magic file from /etc/magic:/usr/share/misc/magic	data	JSON data	gzip compressed data, last modified: Mon Apr 27 14:29:49 2026, max compression, original size modulo 2^32 9	gzip compressed data, last modified: Mon Apr 27 14:29:49 2026, max compression, original size modulo 2^32 9	ASCII text	Java archive data (JAR)	Java archive data (JAR)	Zip archive data, at least v2.0 to extract	Java archive data (JAR)	POSIX tar archive	gzip compressed data, was "sample.tgz", last modified: Mon Apr 27 14:29:49 2026, max compression, original size modulo 2^32 10240	Java archive data (JAR)	Zip archive data, at least v2.0 to extract
debian-12	debian:12-slim	file-5.44	file-5.44 magic file from /etc/magic:/usr/share/misc/magic	Targa image data - RGBA (1027-1541) 3340 x 3854 x 16 +2312 +2826 - 1-bit alpha - right	JSON text data	gzip compressed data, last modified: Mon Apr 27 14:29:49 2026, max compression, original size modulo 2^32 9	gzip compressed data, last modified: Mon Apr 27 14:29:49 2026, max compression, original size modulo 2^32 9	ASCII text	Java archive data (JAR)	Java archive data (JAR)	Zip archive data, at least v2.0 to extract, compression method=deflate	Java archive data (JAR)	POSIX tar archive	gzip compressed data, was "sample.tgz", last modified: Mon Apr 27 14:29:49 2026, max compression, original size modulo 2^32 10240	Java archive data (JAR)	Zip archive data, at least v2.0 to extract, compression method=deflate
debian-13	debian:13-slim	file-5.46	file-5.46 magic file from /etc/magic:/usr/share/misc/magic	Targa image data - RGBA (1027-1541) 3340 x 3854 x 16 +2312 +2826 - 1-bit alpha - right	JSON text data	gzip compressed data, last modified: Mon Apr 27 14:29:49 2026, max compression, original size modulo 2^32 9	gzip compressed data, last modified: Mon Apr 27 14:29:49 2026, max compression, original size modulo 2^32 9	ASCII text	Java archive data (JAR)	Java archive data (JAR)	Zip archive data, made by v2.0 UNIX, extract using at least v2.0, last modified Apr 27 2026 17:29:48, uncompressed size 27, method=deflate	Java archive data (JAR)	POSIX tar archive	gzip compressed data, was "sample.tgz", last modified: Mon Apr 27 14:29:49 2026, max compression, original size modulo 2^32 10240	Java archive data (JAR)	Zip archive data, made by v2.0 UNIX, extract using at least v2.0, last modified Apr 27 2026 17:29:48, uncompressed size 27, method=deflate
opensuse-leap-16.0	opensuse/leap:16.0	file-5.46	file-5.46 magic file from /etc/magic:/usr/share/file/magic seccomp support included	Targa image data - RGBA (1027-1541) 3340 x 3854 x 16 +2312 +2826 - 1-bit alpha - right	JSON text data	gzip compressed data, last modified: Mon Apr 27 14:29:49 2026, max compression, original size modulo 2^32 9	gzip compressed data, last modified: Mon Apr 27 14:29:49 2026, max compression, original size modulo 2^32 9	ASCII text	Java archive data (JAR)	Java archive data (JAR)	Zip archive data, made by v2.0 UNIX, extract using at least v2.0, last modified, last modified Sun, Apr 27 2026 17:29:48, uncompressed size 27, method=deflate	Java archive data (JAR)	POSIX tar archive	gzip compressed data, was "sample.tgz", last modified: Mon Apr 27 14:29:49 2026, max compression, original size modulo 2^32 10240	Java archive data (JAR)	Zip archive data, made by v2.0 UNIX, extract using at least v2.0, last modified, last modified Sun, Apr 27 2026 17:29:48, uncompressed size 27, method=deflate
oraclelinux-10	oraclelinux:10	file-5.45	file-5.45 magic file from /etc/magic:/usr/share/misc/magic	Targa image data - RGBA (1027-1541) 3340 x 3854 x 16 +2312 +2826 - 1-bit alpha - right	JSON text data	gzip compressed data, last modified: Mon Apr 27 14:29:49 2026, max compression, original size modulo 2^32 9	gzip compressed data, last modified: Mon Apr 27 14:29:49 2026, max compression, original size modulo 2^32 9	ASCII text	Java archive data (JAR)	Java archive data (JAR)	Zip archive data, at least v2.0 to extract, compression method=deflate	Java archive data (JAR)	POSIX tar archive	gzip compressed data, was "sample.tgz", last modified: Mon Apr 27 14:29:49 2026, max compression, original size modulo 2^32 10240	Java archive data (JAR)	Zip archive data, at least v2.0 to extract, compression method=deflate
oraclelinux-9	oraclelinux:9	file-5.39	file-5.39 magic file from /etc/magic:/usr/share/misc/magic	data	JSON data	gzip compressed data, last modified: Mon Apr 27 14:29:49 2026, max compression, original size modulo 2^32 9	gzip compressed data, last modified: Mon Apr 27 14:29:49 2026, max compression, original size modulo 2^32 9	ASCII text	Java archive data (JAR)	Java archive data (JAR)	Zip archive data, at least v2.0 to extract	Java archive data (JAR)	POSIX tar archive	gzip compressed data, was "sample.tgz", last modified: Mon Apr 27 14:29:49 2026, max compression, original size modulo 2^32 10240	Java archive data (JAR)	Zip archive data, at least v2.0 to extract
rockylinux-9	rockylinux:9	file-5.39	file-5.39 magic file from /etc/magic:/usr/share/misc/magic	data	JSON data	gzip compressed data, last modified: Mon Apr 27 14:29:49 2026, max compression, original size modulo 2^32 9	gzip compressed data, last modified: Mon Apr 27 14:29:49 2026, max compression, original size modulo 2^32 9	ASCII text	Java archive data (JAR)	Java archive data (JAR)	Zip archive data, at least v2.0 to extract	Java archive data (JAR)	POSIX tar archive	gzip compressed data, was "sample.tgz", last modified: Mon Apr 27 14:29:49 2026, max compression, original size modulo 2^32 10240	Java archive data (JAR)	Zip archive data, at least v2.0 to extract
ubuntu-22.04-lts	ubuntu:22.04	file-5.41	file-5.41 magic file from /etc/magic:/usr/share/misc/magic	Targa image data - RGBA (1027-1541) 3340 x 3854 x 16 +2312 +2826 - 1-bit alpha - right	JSON data	gzip compressed data, last modified: Mon Apr 27 14:29:49 2026, max compression, original size modulo 2^32 9	gzip compressed data, last modified: Mon Apr 27 14:29:49 2026, max compression, original size modulo 2^32 9	ASCII text	Java archive data (JAR)	Java archive data (JAR)	Zip archive data, at least v2.0 to extract, compression method=deflate	Java archive data (JAR)	POSIX tar archive	gzip compressed data, was "sample.tgz", last modified: Mon Apr 27 14:29:49 2026, max compression, original size modulo 2^32 10240	Java archive data (JAR)	Zip archive data, at least v2.0 to extract, compression method=deflate
ubuntu-24.04-lts	ubuntu:24.04	file-5.45	file-5.45 magic file from /etc/magic:/usr/share/misc/magic	Targa image data - RGBA (1027-1541) 3340 x 3854 x 16 +2312 +2826 - 1-bit alpha - right	JSON text data	gzip compressed data, last modified: Mon Apr 27 14:29:49 2026, max compression, original size modulo 2^32 9	gzip compressed data, last modified: Mon Apr 27 14:29:49 2026, max compression, original size modulo 2^32 9	ASCII text	Java archive data (JAR)	Java archive data (JAR)	Zip archive data, at least v2.0 to extract, compression method=deflate	Java archive data (JAR)	POSIX tar archive	gzip compressed data, was "sample.tgz", last modified: Mon Apr 27 14:29:49 2026, max compression, original size modulo 2^32 10240	Java archive data (JAR)	Zip archive data, at least v2.0 to extract, compression method=deflate
ubuntu-26.04-lts	ubuntu:26.04	file-5.46	file-5.46 magic file from /etc/magic:/usr/share/misc/magic	Targa image data - RGBA (1027-1541) 3340 x 3854 x 16 +2312 +2826 - 1-bit alpha - right	JSON text data	gzip compressed data, last modified: Mon Apr 27 14:29:49 2026, max compression, original size modulo 2^32 9	gzip compressed data, last modified: Mon Apr 27 14:29:49 2026, max compression, original size modulo 2^32 9	ASCII text	Java archive data (JAR)	Java archive data (JAR)	Zip archive data, made by v2.0 UNIX, extract using at least v2.0, last modified Apr 27 2026 17:29:48, uncompressed size 27, method=deflate	Java archive data (JAR)	POSIX tar archive	gzip compressed data, was "sample.tgz", last modified: Mon Apr 27 14:29:49 2026, max compression, original size modulo 2^32 10240	Java archive data (JAR)	Zip archive data, made by v2.0 UNIX, extract using at least v2.0, last modified Apr 27 2026 17:29:48, uncompressed size 27, method=deflate

[zyv]

No. It's not rubbish.

In some previous edits of your ticket, there were claims about patterns being incorrect, on the basis of which the conclusion followed that they have to be completely rewritten. This is rubbish, because the LLM (or you) didn't understand the quoting requirements of mc.ext.ini, which are, unfortunately, not trivial indeed; see #4502 for the reference.

It is safe to assume that updates to file utility get into most Linux distros with little delay.

As I already explained, there are different implementations of file and versions of databases beyond Linux. We support macOS, FreeBSD, Solaris, and AIX, to name only major platforms, not even mentioning HP-UX and others.

But anyway. That's quite easy to check. LLM made a script to compare its output across various distros. Results are below.

Nope, it's not easy at all. I'm not sure what I should learn from this table.

First, you only checked a few file versions and specifically on Linux. See above why this is not sufficient.

Second, you don't seem to realize how the patterns work. I think we would agree that if you have the correct extension, then the patterns do not matter. However, if you don't, then things become "interesting".

The output of file (with and without -z) depends not only on the version of the tool and the database, but also on specific archives, because the patterns are programmed in a way that they can give different results depending on the file contents. Check this issue for a sample: #4861. There are also some example files attached to other issues.

If you want to improve the patterns, you need to not only take into account different file and database versions, but also check it on various test files, which were known to produce different results in the past.

---

The bottom line is as I said, I absolutely do not object to improving the patterns based on empirical examples and a correct understanding of the system. My suggestion would be to pick a specific pattern first, like jar-war-by-type, collect example files, run tests on different platforms and file implementations, get a list of file outputs, and on the basis of this list decide how the pattern can be amended for better coverage.

What I do object to is the approach of "let's rewrite everything without even having understood how it works and what the problems are, and surely everything will be better". This is a typical LLM attitude, unfortunately and unsurprisingly... after all, it's been trained on human output.

[azazar]

In some previous edits of your ticket, there were claims about patterns being incorrect, on the basis of which the conclusion followed that they have to be completely rewritten. This is rubbish, because the LLM (or you) didn't understand the quoting requirements of mc.ext.ini, which are, unfortunately, not trivial indeed; see #4502 for the reference.

I'm sorry for that AI slop. I didn't expect LLM to escape using MCP from Copilot Chat. I've rewritten that post in a few minutes, once I noticed that.

[zyv]

I'm sorry for that AI slop. I didn't expect LLM to escape using MCP from Copilot Chat. I've rewritten that post in a few minutes, once I noticed that.

That's fine. I was just explaining what I was referring to as rubbish to avoid any misunderstandings.

[azazar]

If you want to improve the patterns, you need to not only take into account different file and database versions, but also check it on various test files, which were known to produce different results in the past.

I could also extract the patterns actually used by various OS distributions and use them to create tests for the supplied .ini file, so that the .ini file itself is covered by tests - if that would be a useful contribution.

[zyv]

If you want to improve the patterns, you need to not only take into account different file and database versions, but also check it on various test files, which were known to produce different results in the past.

I could also extract the patterns actually used by various OS distributions and use them to create tests for the supplied .ini file, so that the .ini file itself is covered by tests - if that would be a useful contribution.

That would be an extremely useful contribution!

Unfortunately, we don't currently have any tests for the extension file. You'd have to think of some low-footprint and portable way to hook it into our test harness. At which level to do that is a non-trivial question, and definitively some good thinking is required to make a reasonable decision.

Probably one valid idea could be to write a data-driven check test, because it would not only test the patterns themselves, but also the whole loading / escaping / matching machinery and so on, which used to cause a lot of trouble in the past.