Bug#793651: RFS: hdump/2.3-1 [ITP] -- Hexadecimal and ASCII dumper for binary files
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA256
Hi Jakub,
I work more in package and correct issues that you point about packaging and upload again to mentors.
http://mentors.debian.net/package/hdump
I send to upstream your review about LSF and more.
Thanks.
[]'s
kretcheu
:x
On 26-07-2015 09:04, Jakub Wilk wrote:
> I don't intend to upload this package, but here's my review:
>
> * Paulo <kretcheu@gmail.com>, 2015-07-25, 21:16:
>> http://mentors.debian.net/debian/pool/main/h/hdump/hdump_2.3-1.dsc
>
> The package description reads:
>
>> Description: Hexadecimal and ASCII dumper for binary files
>
> The synopsis is not a sentence, so no need to start it with a capital letter. (Developer's Reference §6.2.2)
Corrected.
>> Fast and simple hexadecimal/ASCII dumper for binary files,
>
> OTOH, the long description should consist of full sentences. (Developer's Reference §6.2.3)
Correced.
>> written in ANSI C.
>
> That's not something end users care, so I don't think it should belong in the package description.
OK.
>> - Multi-platform (tested on GNU/Linux and Windows).
>
> Again, not relevant for end users.
OK.
>> - Specify the initial byte (-b). Supports hex notation.
>
> Huh, what? There's no such option. (Not that I know what it would be useful for...)
typo, corrected.
>> - Define numbers of bytes (-n). Multiple of the number of columns.
>
> I don't understand what the second sentence is supposed to mean.
I rewrote.
> The patch header reads:
>
>> Description: fix some issues in upstream Makefile to hardening.
>
> That's not very informative. I would have no idea what this patch does without reading the actual diff.
Changed.
> Also, please forward the patch upstream.
done.
> The Makefile passes both -W and -Wextra to gcc, even though -W is just an obsolete alias for -W.
done.
> I think fsprintf might be a typo.
>
> debian/manpage/genallman.sh ignores all errors. (Although you don't use it debian/rules, so meh...)
include genallman.sh to work on d/rules.
> The manpage, like the package description, says something about ANSI C, the mysterious -b option, and "multiple of the number of columns".
corrected.
> man-pages(7) strongly discourages AUTHORS sections, and so do I.
> I find the "for the Debian project (but may be used by others)" part particularly silly.
deleted.
> Upstream changelog reads:
>
>> - Fixed compilation failure due misuse of fsprintf() function.
>
> Lintian says:
> X: hdump: binary-file-built-without-LFS-support usr/bin/hdump
>
> And indeed, the program doesn't support files bigger an 2GB on 32-bit architectures:
>
> $ truncate -s 3G foo
> $ hdump foo
> file not found or not readable
>
> The program ignores read and write errors:
>
> $ hdump README.md > /dev/full
> $ echo $?
> 0
>
> $ hdump /proc/self/mem
> $ echo $?
> 0
>
> If you provide an option, but not the path, it treats the last argument as path:
>
> $ hdump -c 0
> file not found or not readable
>
> You can't dump non-seekable files:
>
> $ echo foo | hdump /dev/stdin
> unable to seek through file
>
>
> [ This review was written for the debian-mentors mailing list (but may be read by others). ;-) ]
Good!
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2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=RCWY
-----END PGP SIGNATURE-----
Reply to: