Transient network problems should result in network-Exceptions being
signaled, not buffer pointers errors. I sent a message a week or two
ago, "What is the contract of SocketStream>>#next:?" which talked
about a similar-sounding problem with a different method. Since then
I saw the problem happen again (#next: 4 giving me back an empty Array
with no error signaled at all). I traced through the code for over an
hour trying to figure what could posisbly have happened, but the code
looks correct..
On Sun, Oct 19, 2014 at 6:26 PM, David T. Lewis <
[hidden email]> wrote:
> On Sat, Oct 18, 2014 at 04:44:15PM -0700, Sean P. DeNigris wrote:
>> Squeak4.5
>> latest update: #13965
>>
>> 1. In MC Browser, select
http://source.squeak.org/trunk>> 2. Click open
>>
>> I get "Error: subscript is out of bounds: -2" when
>> SocketStream>>upToAll:limit: computes a negative index via "lastRead - sz +
>> 2", which is (0 - 4 + 2).
>>
>
> I am not seeing this in my image. Anyone else? Possible a transient network problem?
>
> Dave
>