Beta 2 setup smoother; unit test broke

Hi Blair,

Workspaces (at least in the CHB) are doing a lot of redrawing, or something.
Browse to a method with "a lot" of source code and hold down the up or down
arrow to move through it.  Beta 2 does a lot of screen updating that beta 1
and D4 don't do, or if they do, I can't see it.

The unit test that broke appears to be suffering because you defined
#queryPort differently than I did.  First, is #queryPort new?  Maybe I
simply never noticed it before.  Anyway, I think you'll find that the number
you are returning is in network byte order rather than the host
representation.  Maybe that's intended, but, the host rep makes more sense
to me.  See the last line(s) below for my suggested change.

queryPort
 "Lookup the port using getsockname().  Answer the result as a SmallInteger,
 or signal a socket error."

 | name nameLength ret |

 name := SOCKADDR_IN new
  sin_family:AF_INET;
  yourself.
 nameLength := SDWORD new value:name size; yourself .

 ( WSockLibrary default
  getsockname:self asParameter
  name:name
  namelen:nameLength ) < 0 ifTrue:[ self error ].

 "wks change
 ^name sin_port."
 ^WSockLibrary default ntohs:name sin_port


The package browser seems to not open $ as much as before; I liked it better
the other way, but, it's not a big deal.  The package folder name case
sensitivity thing appears to be fixed.  I still don't have about blocks in
my packages; should that be fixed?  If so, don't waste time looking for a
bug yet, because it's possible that I got them from the wrong place.


Have a good one,

Bill

--
Wilhelm K. Schwab, Ph.D.
[hidden email]

Bill

"Bill Schwab" <[hidden email]> wrote in message
news:[hidden email]...
>
> Workspaces (at least in the CHB) are doing a lot of redrawing, or
something.
> Browse to a method with "a lot" of source code and hold down the up or
down
> arrow to move through it.  Beta 2 does a lot of screen updating that beta
1
> and D4 don't do, or if they do, I can't see it.

Yes, something fishy is going on there. #549.

>
> The unit test that broke appears to be suffering because you defined
> #queryPort differently than I did.  First, is #queryPort new?  Maybe I
> simply never noticed it before.  ...

It has been around (according to our source code database) since 20th
January 1999, so I don't quite understand why it would have broken your unit
test now. In fact there have been very few changes to the sockets package in
D5, if any.

I don't know if changing it now would be likely to break other peoples'
stuff.

>...
> The package browser seems to not open $ as much as before; I liked it
better
> the other way, but, it's not a big deal.

I'm surprised that has changed, since I didn't think we'd done anything in
that area yet, unless it is a side effect of another change. Any specific
examples?

>....The package folder name case
> sensitivity thing appears to be fixed.  I still don't have about blocks in
> my packages; should that be fixed?  If so, don't waste time looking for a
> bug yet, because it's possible that I got them from the wrong place.

No, if you check the bugs list you will see that it is still outstanding.

Regards

Blair

Hi Blair,

> > The unit test that broke appears to be suffering because you defined
> > #queryPort differently than I did.  First, is #queryPort new?  Maybe I
> > simply never noticed it before.  ...
>
> It has been around (according to our source code database) since 20th
> January 1999, so I don't quite understand why it would have broken your
unit
> test now. In fact there have been very few changes to the sockets package
in
> D5, if any.
>
> I don't know if changing it now would be likely to break other peoples'
> stuff.

Fair enough.  I've probably just been overwriting it with my code.  I add
several related methods, and it would be easy to simply file them in without
noticing the extant methods.



> > The package browser seems to not open $ as much as before; I liked it
> better
> > the other way, but, it's not a big deal.
>
> I'm surprised that has changed, since I didn't think we'd done anything in
> that area yet, unless it is a side effect of another change. Any specific
> examples?

Oddly, I can't get it to do it now =:0    If I notice anything further, I'll
report back.



> >....The package folder name case
> > sensitivity thing appears to be fixed.  I still don't have about blocks
in
> > my packages; should that be fixed?  If so, don't waste time looking for
a
> > bug yet, because it's possible that I got them from the wrong place.
>
> No, if you check the bugs list you will see that it is still outstanding.

Does that mean we can do that on the fly?  If so, what's the URL?

Have a good one,

Bill

--
Wilhelm K. Schwab, Ph.D.
[hidden email]