FS#35527 - [clusterssh] interface doesn't become drawed properly

Attached to Project: Community Packages
Opened by Silvio Knizek (killermoehre) - Wednesday, 29 May 2013, 07:12 GMT
Last edited by Sven-Hendrik Haase (Svenstaro) - Saturday, 13 July 2013, 09:27 GMT
Task Type Bug Report
Category Packages
Status Closed
Assigned To Jaroslav Lichtblau (Dragonlord)
Architecture All
Severity Low
Priority Normal
Reported Version
Due in Version Undecided
Due Date Undecided
Percent Complete 100%
Votes 7
Private No

Details

Description:
In clusterssh 3.28-3 the interface isn't drawed properly. You can use it, every menu is there, but no elements are shown. It's like clicking in the blind. Rebuilding perl-tk and clusterssh doesn't solves this problem.

The error message on stderr is »TclObj_get <some ID> int 0'«, printed as fast as possible.

Additional info:
I use [testing], [community-testing] and [multilib-testing].
The attachment shows cssh with some submenu open.
This task depends upon

Closed by  Sven-Hendrik Haase (Svenstaro)
Saturday, 13 July 2013, 09:27 GMT
Reason for closing:  Fixed
Comment by Dave Seff (Adelie) - Wednesday, 29 May 2013, 17:47 GMT
I can confirm, as I have this issue as well.
Comment by Jay Fougere (LANLocked) - Wednesday, 29 May 2013, 17:58 GMT
This issue persists with AUR ( perl-app-clusterssh 4.01_01-2 ) package as well, even after rebuild of the package:

https://aur.archlinux.org/packages/perl-app-clusterssh/

EDIT:

I was able to get cssh working again by downgrading perl and perl-tk to pre May 26 versions (when perl was upgraded to version 5.18.0.1 from 5.16.3-3 and perl-tk was upgraded to version 804.030-5 from 804.030-4 ).
Comment by Shaun Gowie (sgowie) - Thursday, 30 May 2013, 19:06 GMT
If these errors are logged you run the risk of filling your /var filesystem (or / if you haven't moved /var) rendering the machine almost unusable.
Comment by Shaun Gowie (sgowie) - Monday, 24 June 2013, 19:19 GMT
System performance is also significantly impacted. While cssh is active, even if stderr is directed to /dev/null, a full CPU core is consumed by the error triggering event.
Comment by Damien Gombault (Desintegr) - Tuesday, 02 July 2013, 07:15 GMT
I have upgraded perl-tk from 804.030-5 to 804.031-1.
It seems to work again :)
Comment by Stefan J. Betz (encbladexp) - Tuesday, 02 July 2013, 15:24 GMT
Works for me to!
Comment by Markus Holtermann (MarkusH) - Tuesday, 02 July 2013, 15:29 GMT
Works for me as well.

Name : perl-tk
Version : 804.031-1

Name : clusterssh
Version : 3.28-3

Loading...