[Glass] Is there a #upToEnd kind of read for a none blocking GsSocket?

Wed Jul 5 19:39:57 PDT 2017

On Wed, Jul 5, 2017 at 6:22 PM, Richard Sargent via Glass <
glass at lists.gemtalksystems.com> wrote:

> GLASS mailing list wrote
> > On Wed, Jul 5, 2017 at 5:44 PM, Richard Sargent via Glass <
>
> > glass at .gemtalksystems
>
> >> wrote:
> >
> >> GLASS mailing list wrote
> >> > I am trying to use GsHostProcess and provide a polling API for it.
> >> > Something like this:
> >> >
> >> >
> >> > ------
> >> >
> >> > runAndWaitPollingEvery: pollingMsPeriod onExitDo: anExitBlock
> >> >
> >> >   | process delay outStream errStream |
> >> >   delay := Delay forMilliseconds: pollingMsPeriod.
> >> >   outStream := WriteStream new.
> >> >   errStream := WriteStream new.
> >> >
> >> >   process := GsHostProcess fork: aCommandString.
> >> >
> >> >   [ process childHasExited ]
> >> >   whileFalse: [
> >> > delay wait.
> >> >                 "retrieve all we have in the pipes to make room and
> >> avoid
> >> > locks"
> >> > outStream nextPutAll: *process stdout upToEnd.*
> >> > errStream nextPutAll: *process stderr upToEnd.*
> >> > ].
> >> >
> >> >   outStream nextPutAll: *process stdout upToEnd.*
> >> >   errStream nextPutAll: *process stderr upToEnd.*
> >> >
> >> >   process stdout close.
> >> >   process stderr close.
> >> >
> >> >   aBlock cull: outStream contents cull: errStream contents
> >> >
> >> > --------
> >> >
> >> > I tried to implement an GsSocket >> upToEnd like this:
> >>
> >> Mariano, I like what you are trying to do. But, I think the name you
> have
> >> chosen is flawed.
> >>
> >> The fact that you had to write a comment where you used it to explain
> >> what
> >> you meant is a red flag for that. "retrieve all we have in the pipes" is
> >> quite different from the usual Stream implementation for #upToEnd. It
> >> sounds
> >> like you want to use a name more like #allAvailable. (Note that I don't
> >> think that's the best possible name, but I don't have a better one at
> >> this
> >> moment.)
> >>
> >>
> > Hi Richard,
> >
> > Thanks, #allAvailable is indeed better. However, note that such a method
> > may be way too general to be put in GsSocket (I don't know). It might
> only
> > make sense when using pipes underneath? I don't know. Probably, I will
> add
> > it as a extension method with my app prefix (#fa) and call it
> > #faAllAvailableInPipe (or something to make clear it may only work for
> > pipes....if this is the case).
> >
> > Aside from the selector, what do you think about the contents? does that
> > sound correct?
>
> #read:into:startingAt: claims it "Returns the number of bytes read, or nil
> if an error,  or 0 if EOF on the receiver." Structurally, it looks good.
>

Yes, exactly. But as those comments where written for GsSocket itself, and
in this case we are using GsSocket but with underlying OS pipes (rather
than underlying OS sockets) I was not 100% sure it should be ok.

> Correct the return value and add the bytes to the stream instead of the
> number of bytes and it should work.
>
> As a matter of style, I would apply de Morgan's rule to the loop to make it
> more comprehensible.
> [self readWillNotBlock and: [(bytesRead := self read: buffer size into:
> buffer startingAt: 1) > 0]]
> whileTrue: [newStream nextPutAll: (buffer copyFrom: 1 to: bytesRead)].
>
>
Yes, that makes sense. Another thing I am not sure is whether I should use
String or ByteArray as buffers (I am not sure wether the streams are opened
as binary or text).

Anyway..... I am going to sleep now but this is what I will end up testing
tomorrow (with the polling API of the first email )

GsSocket >> faAllAvailableInPipe
| newStream buffer nextBytes bytesRead |
buffer := String new: 1000.
newStream := WriteStream on: (String new: 100).
[self readWillNotBlock and: [(bytesRead := self read: buffer size into:
buffer startingAt: 1) > 0]]
whileTrue: [newStream nextPutAll: (buffer copyFrom: 1 to: bytesRead)].
^ newStream contents

Thanks for your help.

>
>
> > Thanks!
> >
> >
> >
> >>
> >>
> >> > -----
> >> > upToEnd
> >> > | newStream buffer nextBytes |
> >> > buffer := buffer1 species new: 1000.
> >> > newStream := WriteStream on: (buffer1 species new: 100).
> >> > [self readWillNotBlock not or: [(nextBytes := self read: buffer size
> >> into:
> >> > buffer startingAt: 1) isEmpty]]
> >> > whileFalse: [newStream nextPutAll: nextBytes].
> >> > ^ newStream contents
> >> > -----
> >> >
> >> > Does that sound correct?
> >> >
> >> > Thanks a lot in advance,
> >> >
> >> >
> >> > --
> >> > Mariano
> >> > http://marianopeck.wordpress.com
> >> >
> >> > _______________________________________________
> >> > Glass mailing list
> >>
> >> > Glass at .gemtalksystems
> >>
> >> > http://lists.gemtalksystems.com/mailman/listinfo/glass
> >>
> >>
> >>
> >>
> >>
> >> --
> >> View this message in context: http://forum.world.st/Is-
> >> there-a-upToEnd-kind-of-read-for-a-none-blocking-GsSocket-
> >> tp4953564p4953607.html
> >> Sent from the GLASS mailing list archive at Nabble.com.
> >> _______________________________________________
> >> Glass mailing list
> >>
>
> > Glass at .gemtalksystems
>
> >> http://lists.gemtalksystems.com/mailman/listinfo/glass
> >>
> >
> >
> >
> > --
> > Mariano
> > http://marianopeck.wordpress.com
> >
> > _______________________________________________
> > Glass mailing list
>
> > Glass at .gemtalksystems
>
> > http://lists.gemtalksystems.com/mailman/listinfo/glass
>
>
>
>
>
> --
> View this message in context: http://forum.world.st/Is-
> there-a-upToEnd-kind-of-read-for-a-none-blocking-GsSocket-
> tp4953564p4953613.html
> Sent from the GLASS mailing list archive at Nabble.com.
> _______________________________________________
> Glass mailing list
> Glass at lists.gemtalksystems.com
> http://lists.gemtalksystems.com/mailman/listinfo/glass
>

-- 
Mariano
http://marianopeck.wordpress.com
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.gemtalksystems.com/mailman/private/glass/attachments/20170705/b66ed399/attachment-0001.html>