Re: git archive should use vendor extension in pax header

From: fuz@fuz.su <fuz@fuz.su>
Date: Feb 6, 2016, 14:57 UTC
Message-ID: <20160206145726.GA27001@fuz.su>
In-Reply-To: <56B5F3BF.7080601@web.de>

On Sat, Feb 06, 2016 at 02:23:11PM +0100, René Scharfe wrote:

Show 13 quoted lines

> Am 28.01.2016 um 00:45 schrieb fuz@fuz.su:
> >>There is git get-tar-commit-id, which prints the commit ID if it
> >>finds a comment entry which looks like a hexadecimal SHA-1 hash.
> >>It's better than a hex editor at least. :)
> >
> >This is incredibly fuzzy and can get wrong for a pleothora of reasons.
> >I hope you agree though that the situation is suboptimal, git is doing
> >the equivalent of using a custom file format without an easily
> >recognizable magic number.
> 
> It is fuzzy in theory. But which other programs allow writing a
> comment header?  I'm not aware of any, but I have to admit that I
> didn't look too hard.

Well, let's say what happens if the Mercurial folks were to implement the same thing? Suddenly there is a conflict. Yes, of course, right now there might be no program that uses the comment field for its own purpose but such design decisions tend to be not future proof. There is a very good reason why file formats typically have magic numbers and don't just rely on people knowing that the file has a certain type and that is the same reason why git should mark its meta data in a unique fashion.

Show 10 quoted lines

> >>But I'm still interested how you got a collection of tar files with
> >>unknown origin.  Just curious.
> >
> >Easy: Just download the (source) distribution archives of a distribution
> >of choice and try to verify that the tarballs they use to compile their
> >packages actually come from the project's public git repositories.
> 
> OK, that's easier than calculating checksums and comparing them with
> those published by the respective projects, but also less
> trustworthy.

If you have a known trusted archive, you could use it directly, no need for cross-verification. The intent is to be able to check if archives generated by someone from some sources could have plausibly been generated from these sources.

Show 16 quoted lines

> >There are other reasons why an easily detectable git hash might be
> >useful.  For example, file(1) could show that the archive comes from
> >git.  Other utilities could use this to work around git-specific bugs.
> >An unpacker could add corresponding meta-data when unpacking the file.
> 
> file(1) could use the same heuristic as git get-tar-commit-id.
> Something like this would work (the first line is already shipped
> with file):
> 
> 	257	string	ustar\0 POSIX tar archive
> 	>156	string	g
> 	>>512	string	52\ comment=
> 	>>>523	regex	[0-9a-f]{40}	\b, git commit %s
> 
> NB: With Ian Darwin's file you need to use -e tar in order to turn
> off its internal tar test.

Same issues as above.

> I'm very interested in hearing about any git specific bugs.

I don't know any. Bugs tens to be known only after 1000s of buggy archives have been published (just as with some GNU tar bugs). It's great to have a way to detect that the archive might be affected by a bug so you know that you need to work around it.

Show 24 quoted lines

> >>>>>It would be much more useful if git created a
> >>>>>custom key. As per POSIX suggestions, something like this would be
> >>>>>appropriate:
> >>>>>
> >>>>>     GIT.commit=57ca140635bf157354124e4e4b3c8e1bde2832f1
> >>>>
> >>>>This would be included in addition to the comment in order to avoid
> >>>>breaking existing users, I guess.
> >>>
> >>>Good point.  I'm not sure how many user use the comment header at all.
> >>
> >>Apart from git get-tar-commit-id I don't know any program for
> >>extracting pax comments.  And I don't know how widely used that is,
> >>but I assume there is *someone* out there, extracting commit IDs
> >>with it.
> >
> >Neither do I.  But remember, POSIX explicitly specifies that programs
> >that parse pax file must ignore pax comments so an unpacker that
> >interpretes the content of such a comment in any way is in violation of
> >the pax specification.
> 
> Almost right: The spec says that *pax* shall ignore comments.  Which
> is good -- we can use this field to transport anything without pax
> complaining.

The intent of the committee is that comments shall be ignored by all software that processes tar files. Of course you can put metadata into the comments, but that is just a perversion of the file format. POSIX explicitly states that unknown keys in extended headers are to be ignored by extractors, so using these should be fine, too. But you are right, some research needs to be done as to how different archivers deal with unexpected keys in pax headers.

Show 13 quoted lines

> >>>Maybe there should also be an
> >>>GIT.original-name key.
> >>
> >>What would it be used for?
> >
> >In case an export substition changes the file name so the implementation
> >can verify that the original file could plausibly have been substituted
> >into the current name.  Also for the case where multiple files
> >substitute into the same name to tell which file git should check
> >equivalency with.
> 
> Stupid question: Could you please provide an example?  The only
> possibility for name changes that I'm aware of is using --prefix.

I was under the impression that export substitutions also apply to file names but it seems like I'm wrong about this.

Show 21 quoted lines

> >>>An option GIT.export-ignore is not required.  Instead it would be more
> >>>useful to have a special file type G (for git) with the convention that
> >>>the file name .gitattributes means “attributes that apply to this git
> >>>archive.”
> >>
> >>That would be a non-standard extension.  Archivers would extract
> >>these as regular files.  Storing a list of excluded paths (in
> >>GIT.exclude or so) might be a better idea.
> >
> >No, that's not a good idea as pax headers are interpreted as “attributes
> >pertaining to a file.”  A file doesn't have the attribute that other
> >files have been omitted.  Making this a special file type is useful as
> >it allows archivers that don't implement git extensions to recover this
> >information in a useful way (after all, the .gitattributes file took
> >part in creating the archive) and, more importantly, reserves a file
> >type for future git extensions.
> 
> We can interpret our own keywords as we see fit.  Other programs
> will ignore them (or at most print a warning).  There are precedents
> for global headers pertaining to the whole archive, e.g.
> SCHILY.archtype of star by Jörg Schilling.

Since a tar archive is semantically a concatenation of individual file records, it makes sense that each file has the attribute “this file has star extensions.”

Thinking about the problem a bit more and discussion with the aforementioned Jörg Schilling we came to the conclusion that the best way to deal with an “file omitted” attribute is to attach it to the directory that would normally contain the omitted file.

> Letting archivers extract meta data as regular files is annoying to
> those that are not interested in it.  Extended headers themselves
> (type g) are bad enough already in this regard for those stuck with
> old tar versions.

I think we can safely assume that systems support pax headers 15 years after they have been standardized. I was actually unable to find a non-historical version of a serious archiver that claims to support tar archives but doesn't support pax headers.

Show 21 quoted lines

> >>>The GIT.path option holds the paths that are being archived. It is a bit
> >>>tricky to get right.  The intent of POSIX pax headers is that each key
> >>>is an attribute that applies to a series of files.  In the case of a
> >>>global header, each key applies until it is overridden with a new
> >>>header or with a local header.  A GIT.path key should only apply to the
> >>>files that correspond to this path operant to git archive.  Thus, a new
> >>>GIT.path should be written frequently.  There should always be at least
> >>>one GIT.path.
> >>
> >>That's for the optional path parameters of git archive, right?  A
> >>list of included paths (GIT.include) would be simpler and should
> >>suffice, no?
> >
> >No.  Again: An attribute in a pax header pertains a file.  It's metadata
> >attached to a file, not metadata attached to the whole archive, even when
> >part of a global header.  Thus each file should have attached what path
> >operand it came from.  A file doesn't have the attribute what other path
> >operands git received, only the path operand that caused the inclusion of
> >that one file is an attribute of the file.
> 
> Not an issue; we can make our own rules for our own keywords.

Well, yes, but they should still stick to the semantic concept POSIX imposes for extended headers: headers pertain to files and the only difference between a g header and an x header is that the former applies until it is revoked by a new g header or overridden by an x header. Not sticking to this concept can lead to weird problems with programs that modify tar archives (like GNU-tar) and is not future proof. Better stick to the standard.

Show 7 quoted lines

> >>Users can always go back to the original format.  At least I don't
> >>expect this new format becoming the default too quickly.
> >
> >Sure thing.  If this is going to be implemented, I would add options
> >to choose what / what style of metadata to include.
> 
> Alright.  (An environment requiring these options sounds scary, though.)

Always remember: https://xkcd.com/1172/

> René

Yours, Robert Clausecker

-- 
()  ascii ribbon campaign - for an 8-bit clean world 
/\  - against html email  - against proprietary attachments