Porting Subversion to VMS - VMS

This is a discussion on Porting Subversion to VMS - VMS ; I am attempting (with what little knowledge of c that I have) to port Subversion to OpenVMS. I have looked but can't see where anyone has done this already. I currently have three customers that use Netbeans for Java development ...

+ Reply to Thread
Results 1 to 13 of 13

Thread: Porting Subversion to VMS

  1. Porting Subversion to VMS

    I am attempting (with what little knowledge of c that I have) to port
    Subversion to OpenVMS. I have looked but can't see where anyone has done
    this already. I currently have three customers that use Netbeans for
    Java development and they would like the source code to be "in a safe
    place" which they all agree is their OpenVMS system. I could install
    Subversion on a Linux system (they all have them) and back it up to
    OpenVMS, but it would be preferable that the source stay on the OpenVMS
    system.

    I know there is a port of the CVS client to OpenVMS, but CVS appears to
    be replaced by Subversion and Subversion has potential pluses as it can
    work with apache Apache. I'll (hopefully) tackle that later.

    Most of the code compiles and the parts that don't seem to be related to
    the Berkeley DB which I don't plan on using.

    Subversion uses something called "Apache Portable Runtime" which would
    seem to isolate the Subversion code from the real OS file & network
    specifics.

    If I get this done, I will make an Alpha and Itanium version available.

    Does anyone have any experience with Apache Portable Runtime on VMS or
    is there anyone who would care to help?

    Jeff Coffield
    www.digitalsynergyinc.com

  2. Re: Porting Subversion to VMS

    Jeffrey H. Coffield a écrit :
    > I am attempting (with what little knowledge of c that I have) to port
    > Subversion to OpenVMS.


    Have you noticed that Mercurial is available for Vms ?
    http://vmspython.dyndns.org/Download...ationMercurial

    Do you absolutely need Subversion or a control version tool available on
    Vms ?
    If you go to
    http://hg.vmspython.dyndns.org/vmspython/
    you can see the recent activity

    Regards

    Gérard

  3. Re: Porting Subversion to VMS

    palmipede wrote:
    > Jeffrey H. Coffield a écrit :
    >> I am attempting (with what little knowledge of c that I have) to port
    >> Subversion to OpenVMS.

    >
    > Have you noticed that Mercurial is available for Vms ?
    > http://vmspython.dyndns.org/Download...ationMercurial
    >
    > Do you absolutely need Subversion or a control version tool available on
    > Vms ?
    > If you go to
    > http://hg.vmspython.dyndns.org/vmspython/
    > you can see the recent activity
    >
    > Regards
    >
    > Gérard


    Wow, I missed that big time. I'll have a look.

    Thanks,
    Jeff

  4. Re: Porting Subversion to VMS

    Jeffrey H. Coffield wrote:
    > I am attempting (with what little knowledge of c that I have) to port
    > Subversion to OpenVMS. I have looked but can't see where anyone has done
    > this already. I currently have three customers that use Netbeans for
    > Java development and they would like the source code to be "in a safe
    > place" which they all agree is their OpenVMS system. I could install
    > Subversion on a Linux system (they all have them) and back it up to
    > OpenVMS, but it would be preferable that the source stay on the OpenVMS
    > system.
    >
    > I know there is a port of the CVS client to OpenVMS, but CVS appears to
    > be replaced by Subversion and Subversion has potential pluses as it can
    > work with apache Apache. I'll (hopefully) tackle that later.
    >
    > Most of the code compiles and the parts that don't seem to be related to
    > the Berkeley DB which I don't plan on using.


    Only the client code will compile, the server code requires the Berkely
    DB. (The last time I looked)

    If you only compile the client, you will have to host the repository on
    another system.

    The SleepyCat source, which seems to be what LINUX uses for the Berkely
    DB, also seems to have VMS specific code in it, but I am not sure how
    well it works, if at all.

    > Subversion uses something called "Apache Portable Runtime" which would
    > seem to isolate the Subversion code from the real OS file & network
    > specifics.
    >
    > If I get this done, I will make an Alpha and Itanium version available.
    >
    > Does anyone have any experience with Apache Portable Runtime on VMS or
    > is there anyone who would care to help?


    If you unpack the sources for the VMS Apache product, you will find a
    version of the Apache Portable Runtime that has been ported to VMS already.

    i got the client side stuff to build last year on 8.2. The binary does
    not seem to work with 8.3 though. Also I did not have an SVN server to
    test against, and the firewall was blocking the open source ones that I
    wanted to reference.

    I built the DAV access module, but it did not work.

    If I start over, I would use the things that I have learned about in
    building the components of GTK+ 2.x which are available at
    ftp.encompasserve.org/gnv, along with the patches needed for GNV 2.1 and
    earlier.

    GNU tar on GNV 2.1 seems to be broken, I have not checked to see if
    putting the old GNV kit will work.

    I am currently working on getting GNU TAR 1.19 to run on VMS as I am
    starting to find tarballs that the existing VMSTAR and GNUTAR can not
    handle.

    -John
    wb8tyw@qsl.network
    Personal Opinion Only

  5. Re: Porting Subversion to VMS

    In article ,
    "Jeffrey H. Coffield" wrote:

    > I am attempting (with what little knowledge of c that I have) to port
    > Subversion to OpenVMS. I have looked but can't see where anyone has done
    > this already. I currently have three customers that use Netbeans for
    > Java development and they would like the source code to be "in a safe
    > place" which they all agree is their OpenVMS system. I could install
    > Subversion on a Linux system (they all have them) and back it up to
    > OpenVMS, but it would be preferable that the source stay on the OpenVMS
    > system.
    >
    > I know there is a port of the CVS client to OpenVMS, but CVS appears to
    > be replaced by Subversion and Subversion has potential pluses as it can
    > work with apache Apache. I'll (hopefully) tackle that later.
    >
    > Most of the code compiles and the parts that don't seem to be related to
    > the Berkeley DB which I don't plan on using.
    >
    > Subversion uses something called "Apache Portable Runtime" which would
    > seem to isolate the Subversion code from the real OS file & network
    > specifics.
    >
    > If I get this done, I will make an Alpha and Itanium version available.
    >
    > Does anyone have any experience with Apache Portable Runtime on VMS or
    > is there anyone who would care to help?


    The possibility of porting Subversion to VMS has been discussed before,
    on the ITRC forum at least. See
    <http://forums12.itrc.hp.com/service/...do?threadId=10
    61102>.

    As its name implies, the APR is part of Apache and if you download the
    sources to SWS from the HP site, you'll get a VMS port of the APR.
    Whether it's complete enough and up-to-date enough to form the basis of
    a port of the current Subversion sources is something that would need
    investigating.

    Subversion is not quite as hot an item as it was a year or two ago
    since other cqpable distributed VCS's have come along, but it was
    designed with portability in mind and is pretty ubiquitous and it would
    certainly be nice to have a good port available on VMS. You'll have
    decisions to make about whether and how to support the different VMS
    file formats, how filenames that are illegal on ODS-2 will be handled
    (or whether you'll even support VMS-syntax filenames), etc.

    Someone has mentioned Mercurial, which has a good reputation and is now
    used by the Mozilla project and, I believe, several Sun projects. The
    fact that it's already available on VMS is a plus. There is also
    Bazaar , which, like Mercurial, is written in
    Python and should be rather portable. Git has its adherents
    (); it was specifically written for Linux kernel
    maintenance and is not very portable. It makes liberal use of fork()
    and most of its commands are implemented as shell scripts; I suppose it
    has some potential for running under GNV's bash.

    --
    Posted via a free Usenet account from http://www.teranews.com


  6. Re: Porting Subversion to VMS

    Jeffrey H. Coffield wrote:
    > I know there is a port of the CVS client to OpenVMS, but CVS appears to
    > be replaced by Subversion


    SVN has realy taken off well.

    But there are still a lot of CVS users out there. I would even
    estimate that there are more CVS users than SVN users, but that will
    likely change within the next few years.

    But anyway - CVS is still a usable alternative.

    Arne

  7. Re: Porting Subversion to VMS

    palmipede wrote:
    > Jeffrey H. Coffield a écrit :
    >> I am attempting (with what little knowledge of c that I have) to port
    >> Subversion to OpenVMS.

    >
    > Have you noticed that Mercurial is available for Vms ?
    > http://vmspython.dyndns.org/Download...ationMercurial
    >
    > Do you absolutely need Subversion or a control version tool available on
    > Vms ?
    > If you go to
    > http://hg.vmspython.dyndns.org/vmspython/
    > you can see the recent activity


    And related to the original poster - there are a NetBeans
    plugin.

    And since the NetBeans team is switching to Mercurial themselves,
    then it is a good assumption that it will work !

    Arne

  8. Re: Porting Subversion to VMS

    Arne Vajhøj wrote:
    > palmipede wrote:
    >> Jeffrey H. Coffield a écrit :
    >>> I am attempting (with what little knowledge of c that I have) to port
    >>> Subversion to OpenVMS.

    >>
    >> Have you noticed that Mercurial is available for Vms ?
    >> http://vmspython.dyndns.org/Download...ationMercurial
    >>
    >> Do you absolutely need Subversion or a control version tool available
    >> on Vms ?
    >> If you go to
    >> http://hg.vmspython.dyndns.org/vmspython/
    >> you can see the recent activity

    >
    > And related to the original poster - there are a NetBeans
    > plugin.
    >
    > And since the NetBeans team is switching to Mercurial themselves,
    > then it is a good assumption that it will work !
    >
    > Arne


    Many thanks to palmipede. I have installed Mercurial on OpenVMS and it
    seems to work. It is in NetBeans 6 which is part of the reason I missed
    it when I started. All our production Java is done in NB 5.5.1 but we
    are starting to evaluate NB 6. I like that Mercurial is written in
    Python as I use Python with Blender.

    Jeff

  9. Re: Porting Subversion to VMS

    Have you looked at Theirry Uso's port at http://vmsfree.free.fr/freen/index.php?s=suv
    >


  10. Re: Porting Subversion to VMS

    Galen wrote:
    > Have you looked at Theirry Uso's port at http://vmsfree.free.fr/freen/index.php?s=suv


    Interesting !

    But it is "superversion" not "subversion".

    :-)

    Arne

  11. Re: Porting Subversion to VMS

    In article <4iUfj.22893$Ux2.3521@attbi_s22>,
    "John E. Malmberg" wrote:

    > Only the client code will compile, the server code requires the Berkely
    > DB. (The last time I looked)
    >
    > If you only compile the client, you will have to host the repository on
    > another system.
    >
    > The SleepyCat source, which seems to be what LINUX uses for the Berkely
    > DB, also seems to have VMS specific code in it, but I am not sure how
    > well it works, if at all.


    I'm not up to date with version numbers or anything, but the VMS port of
    htdig contains Berekeley DB.

    --
    Paul Sture

    Sue's OpenVMS bookmarks:
    http://eisner.encompasserve.org/~stu...bookmarks.html

  12. Re: Porting Subversion to VMS

    P. Sture wrote:
    > "John E. Malmberg" wrote:
    >> Only the client code will compile, the server code requires the Berkely
    >> DB. (The last time I looked)
    >>
    >> If you only compile the client, you will have to host the repository on
    >> another system.
    >>
    >> The SleepyCat source, which seems to be what LINUX uses for the Berkely
    >> DB, also seems to have VMS specific code in it, but I am not sure how
    >> well it works, if at all.

    >
    > I'm not up to date with version numbers or anything


    Neither am I...

    > but the VMS port of htdig contains Berekeley DB.


    Indeed. In order to independent of version incompatibilities, the htdig
    maintainers packaged a Berkeley DB v2.6.4 with ht://Dig 3.1.6 (which
    is the version I ported to VMS a few years ago). Couldn't run the tests,
    though, as I didn't have a TCL installation on VMS. It seems to work
    quite well. But it's probably way out of date.

    cu,
    Martin
    --
    One OS to rule them all | Martin Vorlaender | OpenVMS rules!
    One OS to find them | work: mv@pdv-systeme.de
    One OS to bring them all | http://vms.pdv-systeme.de/users/martinv/
    And in the Darkness bind them.| home: martin.vorlaender@t-online.de

  13. Re: Porting Subversion to VMS

    > But it is "superversion" not "subversion".
    >
    > :-)
    >
    > Arne


    Thanks for the correction, Arne. :-)

    I was looking at superversion a few months ago but did little more
    than install it and start it up once or twice.

    Once it "sank" off my radar, it was easy to confuse its name with
    "subversion". :-)

+ Reply to Thread