Re: Windows env-Bug?

Gerhard Reithofer wrote:

Nevertheless the TCL behaviour is IMHO a bug. Should I report a bug to
Sourceforge or does someone already feel responsible to work on that?

Forcing a re-ask of the environment for any access to the array would hurt
performance. Maybe a special behavior regarding setting any variable to ""
implies a delete (all platforms?), but the environment handling code
doesn't have platform specific code to it.

If you want to point a finger, point it at libc's setenv() that Tcl uses.

Yes, it's true that Tcl's behavior is not consistent on windows, but that
is the behavior on that platform. What about the next person who wonders
why Tcl can't set an envvar to empty (do we use the no-break space char to
imply empty?), then comes another person who wonders why their DLL changing
an envvar doesn't cause Tcl to fire trace on it or even show it be the same.
^- hint: The real environment is not sync'd to libc's cache of it.

In a way, this is like exit codes. Win apps have a 32-bit exit code. Tcl
only does 7-bit. Do you raise Tcl's size or truncate it to fit? Well, in
our case, both are happening and neither is fully right.

Patch it if you can, but you'll be staring at best a compromise that might
not address each side effectively.

--
"MOM, CAN I SET FIRE TO MY BED MATTRESS?" "No, Calvin." "CAN I RIDE MY
TRICYCLE ON THE ROOF?" "No, Calvin." "Then can I have a cookie?" "No, Calvin."
("She's on to me.")

Attachment: signature.asc
Description: OpenPGP digital signature

Relevant Pages

  • Re: Windows env-Bug?
    ... Forcing a re-ask of the environment for any access to the array would hurt ... doesn't have platform specific code to it. ... If you want to point a finger, point it at libc's setenvthat Tcl uses. ... this is like exit codes. ...
    (comp.lang.tcl)
  • Re: Ambiguities of the Harel/UML FSM model
    ... In the exit sequence case the implementer simply picks *any* orthogonal region, exits all its states, then picks the next region and so forth. ... to a particular computing environment. ... computing environment unless the OOA modeler somehow modifies the OOA model. ... correct behavior on a particular platform. ...
    (comp.object)
  • Re: Closures, was Re: What does Tcl lack?
    ... > some aspects of its environment, ... Tcl already has a notion of a proc being associated with a namespace. ... >>And I never understoud why Tcl needs closures. ... > described in terms of string substitution and eval (in other languages ...
    (comp.lang.tcl)
  • Re: Agile developement -- A Parable
    ... platform was added at a later date. ... incurred the cost to write all code to be portable to all other platforms ... Agile processes allow you to adapt to changes in the environment. ... Even when constructing the project for multiple environments, ...
    (comp.object)
  • Re: Measuring time differences
    ... >> the OP was provided with re-directions by other posters. ... > the OP and your suggestion of gettimeofday I can say quite categorically ... if you actually know the platform you can point the poster in the ... the OP didn't provide specifics as to their environment. ...
    (comp.lang.c)