1. Forum
  2. Usenet
  3. LINUX.DEBIAN.BUGS.DIST

  • Bug#240992: This is a feature request rather than a bug...

    From Thomas Viehmann@1:229/2 to All on Wed Aug 18 20:30:16 2004
    From: [email protected]

    This is an OpenPGP/MIME signed message (RFC 2440 and 3156)
    package gnucash
    severity 240992 wishlist
    retitle 240992 Please link against libktoblzcheck1
    thanks

    Hi.

    This is a feature request, I'm retiteling it and adjusting the severity. GnuCash's HBCI module can do checking of account numbers if compiled
    against libktoblzcheck (by adding libktoblzcheck1-dev to the build-dependencies).
    Sounds like a useful and easy thing to me, I actually tested it, too.

    Kind regards

    Thomas
    --
    Thomas Viehmann, <http://thomas.viehmann.net/>

    -----BEGIN PGP SIGNATURE-----
    Version: GnuPG v1.2.5 (GNU/Linux)
    Comment: GnuPG key at <http://thomas.viehmann.net/>

    iD8DBQFBI5uFriZpaaIa1PkRAtd+AJwJCOC1eaw+iOMJrKoBggyhLlQxXwCeKASd 3WhcyMN+jPSSt5o6EQkRNOg=
    =dQ/C
    -----END PGP SIGNATURE-----

    --- SoupGate-Win32 v1.05
    * Origin: you cannot sedate... all the things you hate (1:229/2)
  • From Jeroen van Wolffelaar@1:229/2 to Thomas Viehmann on Wed Aug 18 21:30:09 2004
    From: [email protected]

    On Wed, Aug 18, 2004 at 08:10:13PM +0200, Thomas Viehmann wrote:
    GnuCash's HBCI module can do checking of account numbers if compiled
    against libktoblzcheck (by adding libktoblzcheck1-dev to the build-dependencies).
    Sounds like a useful and easy thing to me, I actually tested it, too.

    I might be mistaken, but isn't it so that currently, if you accidently
    have libktoblzcheck1-dev installed, the resulted binary is linked
    against it, and otherwise, it isn't?

    Then, isn't it true that this package either
    1) needs to build-depends on libktoblzcheck1-dev
    2) needs to build-conflicts on libktoblzcheck1-dev
    3) needs to be ./configured to not link to libktoblzcheck1-dev, whether
    it's available or not?

    Now the build may be significantly different depending on the builder's installed
    packages.

    --Jeroen

    --
    Jeroen van Wolffelaar
    [email protected] (also for Jabber & MSN; ICQ: 33944357) http://Jeroen.A-Eskwadraat.nl

    -----BEGIN PGP SIGNATURE-----
    Version: GnuPG v1.2.5 (GNU/Linux)
    Comment: Signed by Jeroen van Wolffelaar <[email protected]>

    iD8DBQFBI6jKl2uISwgTVp8RAk6RAJ45yX1xuA+KqOZBeiKFL56DLmHwAgCbBqWS qza+mVFnSsUkPdetTxRZRCU=
    =inkn
    -----END PGP SIGNATURE-----

    --- SoupGate-Win32 v1.05
    * Origin: you cannot sedate... all the things you hate (1:229/2)
  • From Thomas Bushnell BSG@1:229/2 to Jeroen van Wolffelaar on Wed Aug 18 22:00:13 2004
    From: [email protected]

    severity 240992 normal
    thanks

    Jeroen van Wolffelaar <[email protected]> writes:

    I might be mistaken, but isn't it so that currently, if you accidently
    have libktoblzcheck1-dev installed, the resulted binary is linked
    against it, and otherwise, it isn't?

    Then, isn't it true that this package either
    1) needs to build-depends on libktoblzcheck1-dev
    2) needs to build-conflicts on libktoblzcheck1-dev
    3) needs to be ./configured to not link to libktoblzcheck1-dev, whether
    it's available or not?

    Yes, I believe you are correct. Number (2) is a foolish idea; the
    package should either use the library, and adopt (1), or it should
    not, and use (3).

    That means that a maybe-yes-maybe-no is inconsistent and a bug in the
    package. So since I'm the new maintainer, I'm bumping it back up to
    normal severity.

    Thomas


    --
    To UNSUBSCRIBE, email to [email protected]
    with a subject of "unsubscribe". Trouble? Contact [email protected]

    --- SoupGate-Win32 v1.05
    * Origin: you cannot sedate... all the things you hate (1:229/2)
  • From Thomas Viehmann@1:229/2 to Jeroen van Wolffelaar on Wed Aug 18 22:10:18 2004
    From: [email protected]

    This is an OpenPGP/MIME signed message (RFC 2440 and 3156)
    Hi.

    Jeroen van Wolffelaar wrote:
    I might be mistaken, but isn't it so that currently, if you accidently
    have libktoblzcheck1-dev installed, the resulted binary is linked
    against it, and otherwise, it isn't?
    That's 100% correct.

    Then, isn't it true that this package either
    1) needs to build-depends on libktoblzcheck1-dev
    2) needs to build-conflicts on libktoblzcheck1-dev
    3) needs to be ./configured to not link to libktoblzcheck1-dev, whether
    it's available or not?
    Given this bug report I'd say some people prefer 1). I'm not sure about
    how much of a requirement this is.

    Now the build may be significantly different depending on the builder's installed
    packages.
    Many packages show this type of behaviour and this seems to be (at least quietly) accepted. Also, it's not that big a problem if you routinely
    build packages in a chroot. Here, the difference is pretty well
    localized in the HBCI module, so most people will hardly notice.

    Kind regards

    Thomas

    P.S.: Thomas Bushnell just wrote me a mail that he's maintaining
    gnucash, so he'll probably do stuff or other. Or maybe not. That's the
    fate of the low priority bugs on addons to big packages...
    --
    Thomas Viehmann, <http://thomas.viehmann.net/>

    -----BEGIN PGP SIGNATURE-----
    Version: GnuPG v1.2.5 (GNU/Linux)
    Comment: GnuPG key at <http://thomas.viehmann.net/>

    iD8DBQFBI7RXriZpaaIa1PkRAsJVAJ9FzFCDYZF8m2Q25ramCR6a4AK5zgCgh2nx nINAO73uRGytbkiEHSO6Ptc=
    =XHEZ
    -----END PGP SIGNATURE-----

    --- SoupGate-Win32 v1.05
    * Origin: you cannot sedate... all the things you hate (1:229/2)