Re: Any filename problems going from VAX/ODS-2 to Alpha/ODS-5/case blind? - VMS

This is a discussion on Re: Any filename problems going from VAX/ODS-2 to Alpha/ODS-5/case blind? - VMS ; From: "Richard B. Gilbert" > AEF wrote: > > That's why you should always use F$PARSE (or its system services > > equivalent) whenever possible. Well, duh. > When would it not be possible to use F$PARSE? It's normally _possible_, ...

+ Reply to Thread
Results 1 to 6 of 6

Thread: Re: Any filename problems going from VAX/ODS-2 to Alpha/ODS-5/case blind?

  1. Re: Any filename problems going from VAX/ODS-2 to Alpha/ODS-5/case blind?

    From: "Richard B. Gilbert"

    > AEF wrote:
    > > That's why you should always use F$PARSE (or its system services
    > > equivalent) whenever possible.


    Well, duh.

    > When would it not be possible to use F$PARSE?


    It's normally _possible_, but when you inherit code which says things
    like:

    [...]
    #if defined(VAXC) || defined(VMS)
    # define PATH_SEP ']'
    # define PATH_SEP2 ':'
    # define SUFFIX_SEP ';'
    # define NO_MULTIPLE_DOTS
    # define Z_SUFFIX "-gz"
    [...]

    then you can expect it to be annoying and/or difficult to convert it to
    use $PARSE instead of its own name-parsing code.

    > If you don't use F$PARSE,
    > there is an excellent possibility that someone can construct a legal
    > file spec that you cannot parse correctly.


    Someone _can_, but it may be pretty unlikely.

    > Legal file specs, even if
    > you limit them to ODS-2 file specs, can get pretty weird. You have to
    > remember that VMS inherited a lot of cruft from DEC operating systems
    > that only the old timers remember!


    Assuming that most folks won't actually be using ";" or ":" or "]" in
    file names (or "<>" to delimit a directory, or ...), then much of such
    sub-ideal code can be expected to work (for most people, most of the
    time).

    ------------------------------------------------------------------------

    Steven M. Schweda sms@antinode-org
    382 South Warwick Street (+1) 651-699-9818
    Saint Paul MN 55105-2547

  2. Re: Any filename problems going from VAX/ODS-2 to Alpha/ODS-5/case blind?

    On Aug 20, 10:32 am, s...@antinode.org (Steven M. Schweda) wrote:
    > From: "Richard B. Gilbert"
    >
    > > AEF wrote:
    > > > That's why you should always use F$PARSE (or its system services
    > > > equivalent) whenever possible.

    >
    > Well, duh.


    Just emphasizing the point and trying to make it perfectly clear and
    pound the point home in a (perhaps futile) effort to minimize the
    changes of anyone skipping F$PARSE without good reason.

    Duh!

    >
    > > When would it not be possible to use F$PARSE?

    >
    > It's normally _possible_, but when you inherit code which says things
    > like:
    >
    > [...]
    > #if defined(VAXC) || defined(VMS)
    > # define PATH_SEP ']'
    > # define PATH_SEP2 ':'
    > # define SUFFIX_SEP ';'
    > # define NO_MULTIPLE_DOTS
    > # define Z_SUFFIX "-gz"
    > [...]
    >
    > then you can expect it to be annoying and/or difficult to convert it to
    > use $PARSE instead of its own name-parsing code.
    >
    > > If you don't use F$PARSE,
    > > there is an excellent possibility that someone can construct a legal
    > > file spec that you cannot parse correctly.

    >
    > Someone _can_, but it may be pretty unlikely.
    >
    > > Legal file specs, even if
    > > you limit them to ODS-2 file specs, can get pretty weird. You have to
    > > remember that VMS inherited a lot of cruft from DEC operating systems
    > > that only the old timers remember!

    >
    > Assuming that most folks won't actually be using ";" or ":" or "]" in
    > file names (or "<>" to delimit a directory, or ...), then much of such
    > sub-ideal code can be expected to work (for most people, most of the
    > time).


    Yep. My program doesn't like [0,0], but I thought users could get by
    without it, and the code uses \ as a shortcut for the MFD anyway. I
    suppose I could special case it for my next release, if I do one.

    >
    > ------------------------------------------------------------------------
    >
    > Steven M. Schweda sms@antinode-org
    > 382 South Warwick Street (+1) 651-699-9818
    > Saint Paul MN 55105-2547


    Chill out, dude. Go peruse sci.physics if you want to see what a
    *REALLY* bad newsgroup is like.


    AEF


  3. Re: Any filename problems going from VAX/ODS-2 to Alpha/ODS-5/caseblind?

    Steven M. Schweda wrote:
    > From: "Richard B. Gilbert"
    >
    >>AEF wrote:
    >>
    >>>That's why you should always use F$PARSE (or its system services
    >>>equivalent) whenever possible.

    >>

    >
    > Well, duh.
    >
    >
    >>When would it not be possible to use F$PARSE?

    >
    >
    > It's normally _possible_, but when you inherit code which says things
    > like:
    >
    > [...]
    > #if defined(VAXC) || defined(VMS)
    > # define PATH_SEP ']'
    > # define PATH_SEP2 ':'
    > # define SUFFIX_SEP ';'
    > # define NO_MULTIPLE_DOTS
    > # define Z_SUFFIX "-gz"
    > [...]
    >
    > then you can expect it to be annoying and/or difficult to convert it to
    > use $PARSE instead of its own name-parsing code.
    >
    >
    >> If you don't use F$PARSE,
    >>there is an excellent possibility that someone can construct a legal
    >>file spec that you cannot parse correctly.

    >
    >
    > Someone _can_, but it may be pretty unlikely.
    >
    >
    >> Legal file specs, even if
    >>you limit them to ODS-2 file specs, can get pretty weird. You have to
    >>remember that VMS inherited a lot of cruft from DEC operating systems
    >>that only the old timers remember!

    >
    >
    > Assuming that most folks won't actually be using ";" or ":" or "]" in
    > file names (or "<>" to delimit a directory, or ...), then much of such
    > sub-ideal code can be expected to work (for most people, most of the
    > time).


    Well, if having your code work for most people, most of the time, is
    sufficient. . . .

    It's not that difficult to get it right and have it work for everyone,
    all of the time.



  4. Re: Any filename problems going from VAX/ODS-2 to Alpha/ODS-5/case blind?

    In article <07082009325595_20200296@antinode.org>, sms@antinode.org (Steven M. Schweda) writes:
    >
    > Assuming that most folks won't actually be using ";" or ":" or "]" in
    > file names (or "<>" to delimit a directory, or ...), then much of such
    > sub-ideal code can be expected to work (for most people, most of the
    > time).


    Works for most of the people most of the time? You sound like Bill
    Gates.

    I have gainfull employment fixing that kind of code.

  5. Re: Any filename problems going from VAX/ODS-2 to Alpha/ODS-5/case blind?

    On Aug 20, 1:00 pm, koeh...@eisner.nospam.encompasserve.org (Bob
    Koehler) wrote:
    > In article <07082009325595_20200...@antinode.org>, s...@antinode.org (Steven M. Schweda) writes:
    >
    >
    >
    > > Assuming that most folks won't actually be using ";" or ":" or "]" in
    > > file names (or "<>" to delimit a directory, or ...), then much of such
    > > sub-ideal code can be expected to work (for most people, most of the
    > > time).

    >
    > Works for most of the people most of the time? You sound like Bill
    > Gates.
    >
    > I have gainfull employment fixing that kind of code.


    If only his software were even _that_ good! You give him too much
    credit.

    AEF


  6. Re: Any filename problems going from VAX/ODS-2 to Alpha/ODS-5/case blind?

    On Aug 20, 10:32 am, s...@antinode.org (Steven M. Schweda) wrote:
    > From: "Richard B. Gilbert"
    >
    > > AEF wrote:
    > > > That's why you should always use F$PARSE (or its system services
    > > > equivalent) whenever possible.

    >
    > Well, duh.
    >


    I didn't mean *you* in particular. I meant "you" as in "one" as in the
    general reader or programmer. Perhaps you thought I meant the former.

    [...]
    > Steven M. Schweda sms@antinode-org

    [...]

    AEF


+ Reply to Thread