[
https://issues.apache.org/jira/browse/ETCH-278?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13923850#comment-13923850
]
Aaron Frueh commented on ETCH-278:
----------------------------------
Thanks, you're completely right.
The problem I see is that after this happens the Buffer is corrupted. And we
need to search the remaining buffer for the SIG. If not, we lose all Packages
until suddenly there is a package with the Header at the beginning of the
Package.
e.g.
| Hea | der | Data ............ | Header | Data |
We got half a Header in the savedBuffer. The complete Header is defect because
of a wrong SIG, so we have to discard that package but we don't know if another
Header is still existing inside of the remaining buffer. Even worse the next
Header could be split up again.
Damaged Package:
| Hea | der | Data ............ | He |
Next Package:
| ader | Data .... | .......
I will extend my suggestion to solve that issue. But i'm really not sure if
this is handled correctly in the other etch-Bindings.
> EtchPacketizer isn't able to receive multiple Packages in one Buffer
> --------------------------------------------------------------------
>
> Key: ETCH-278
> URL: https://issues.apache.org/jira/browse/ETCH-278
> Project: Etch
> Issue Type: Bug
> Components: binding-cpp
> Affects Versions: 1.3.0
> Reporter: Aaron Frueh
> Priority: Critical
> Attachments:
> 0001-HotFix-for-EtchPacketizer-receive-of-multiple-packag.patch
>
>
> The packetizer should be able to receive multiple Etch-Packages within one
> buffer. (see also Etch-C-Binding)
> This can happen if the socket-receive is somehow blocked or delayed and the
> internal socket buffer contains multiple packages.
--
This message was sent by Atlassian JIRA
(v6.2#6252)
