[
https://issues.apache.org/jira/browse/SSHD-743?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16268935#comment-16268935
]
masc commented on SSHD-743:
---
Thanks! I verified the fix and it seems to work well, stress tested on both
Windows
[
https://issues.apache.org/jira/browse/SSHD-743?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16255256#comment-16255256
]
Goldstein Lyor commented on SSHD-743:
-
I have decided to wrap the call to {{finishWriting}} in a
[
https://issues.apache.org/jira/browse/SSHD-743?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16254055#comment-16254055
]
masc commented on SSHD-743:
---
Thinking about it, you're right.
The latter issue actually does not relate, as it
[
https://issues.apache.org/jira/browse/SSHD-743?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16254043#comment-16254043
]
masc commented on SSHD-743:
---
I'm not 100% it's originating from the c'tor code, but it may be a possibility.
>
[
https://issues.apache.org/jira/browse/SSHD-743?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16254029#comment-16254029
]
masc commented on SSHD-743:
---
No it's not. This issue is about network io related exceptions leaking from
threads,
[
https://issues.apache.org/jira/browse/SSHD-743?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16251735#comment-16251735
]
Goldstein Lyor commented on SSHD-743:
-
This failure is a *separate* issue - so please open a new one for
[
https://issues.apache.org/jira/browse/SSHD-743?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16251291#comment-16251291
]
masc commented on SSHD-743:
---
btw there's more exception leaks, this time observed in {{Nio2Session}} c'tor,
[
https://issues.apache.org/jira/browse/SSHD-743?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16249423#comment-16249423
]
masc commented on SSHD-743:
---
ok. I can't further comment on this without digging deeper, but from my current
[
https://issues.apache.org/jira/browse/SSHD-743?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16248619#comment-16248619
]
Goldstein Lyor commented on SSHD-743:
-
The way the {{Nio2Session}} is implemented using asynchronous
[
https://issues.apache.org/jira/browse/SSHD-743?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16248560#comment-16248560
]
masc commented on SSHD-743:
---
what is the reason for starting another write while handling an actual write
If you have an internal bug you still want to keep that bug from negatively
affecting other sessions within the app. That’s why browsers are process
per session. That’s also a case for overly generic catches in specific
places.
On Sat, Nov 11, 2017 at 9:07 AM Goldstein Lyor (JIRA)
[
https://issues.apache.org/jira/browse/SSHD-743?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16248536#comment-16248536
]
Goldstein Lyor commented on SSHD-743:
-
I guess what bothers me is that {{catch (Exception e)}} - seems a
[
https://issues.apache.org/jira/browse/SSHD-743?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16246108#comment-16246108
]
masc commented on SSHD-743:
---
so possibly the best way to fix this without swallowing exceptions would be to
[
https://issues.apache.org/jira/browse/SSHD-743?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16246092#comment-16246092
]
masc commented on SSHD-743:
---
actually I don't think it makes sense at all to start writing when the writing
[
https://issues.apache.org/jira/browse/SSHD-743?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16246080#comment-16246080
]
masc commented on SSHD-743:
---
plus, I believe the cases you're trying to avoid would be valid exceptions
during
[
https://issues.apache.org/jira/browse/SSHD-743?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16246078#comment-16246078
]
masc commented on SSHD-743:
---
that's true for most cases, but in this specific case, where a write is
attempted
[
https://issues.apache.org/jira/browse/SSHD-743?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16245986#comment-16245986
]
Goldstein Lyor commented on SSHD-743:
-
I am not convinced that suppressing exceptions is better than
[
https://issues.apache.org/jira/browse/SSHD-743?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16245747#comment-16245747
]
masc commented on SSHD-743:
---
as I see it, {finishWrite} is ok and should actually throw. it's just, that it
[
https://issues.apache.org/jira/browse/SSHD-743?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16239422#comment-16239422
]
Goldstein Lyor commented on SSHD-743:
-
I have made some safer changes - see [this
[
https://issues.apache.org/jira/browse/SSHD-743?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16235795#comment-16235795
]
masc commented on SSHD-743:
---
even though, going with the above implementation, this would merely prevent the
[
https://issues.apache.org/jira/browse/SSHD-743?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16235791#comment-16235791
]
masc commented on SSHD-743:
---
about the {{isOpen}} check, as an exception would be valid in this case for
other
[
https://issues.apache.org/jira/browse/SSHD-743?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16235781#comment-16235781
]
masc commented on SSHD-743:
---
I believe {{finsihWrite}} should be allowed to throw at all times.
t's only fatal
[
https://issues.apache.org/jira/browse/SSHD-743?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16235735#comment-16235735
]
Goldstein Lyor commented on SSHD-743:
-
[~gnt] Can you take at look at [this
[
https://issues.apache.org/jira/browse/SSHD-743?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16235644#comment-16235644
]
Goldstein Lyor commented on SSHD-743:
-
Thanks, looking into it...
> Nio2Session sporadically leaks
[
https://issues.apache.org/jira/browse/SSHD-743?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16235634#comment-16235634
]
masc commented on SSHD-743:
---
the issue seems to be {Nio2Session.handleWriteCycleFailure} calling into
[
https://issues.apache.org/jira/browse/SSHD-743?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16196822#comment-16196822
]
masc commented on SSHD-743:
---
this seems to be fixed in `sshd-1.6.0`, at least I cannot replicate the issue
[
https://issues.apache.org/jira/browse/SSHD-743?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15981475#comment-15981475
]
masc commented on SSHD-743:
---
Correction: we set the idle timeout on the ssh client side, not the host.
>
[
https://issues.apache.org/jira/browse/SSHD-743?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15981468#comment-15981468
]
masc commented on SSHD-743:
---
* The issue occurs since `1.3.0` but `1.4.0` was tested as well and didn't make
a
[
https://issues.apache.org/jira/browse/SSHD-743?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15981411#comment-15981411
]
Goldstein Lyor commented on SSHD-743:
-
* The issue marks also 1.4.0 as having this problem, but I do not
29 matches
Mail list logo
