Bug 221827 – layman standard output being filtered / bypassed during error output

• Product: Portage Development
• Status: RESOLVED INVALID

[Description Dylan Wakefield 2008-05-12 13:23:30 UTC]

Layman has problems with error output bypassing stdout, this seems very similar to the effects of bug 217444 where error (missing build tool for overlay - red asterisk indicated in output) output is being diverted from stdout, not very handy if involving layman in a script. The easiest way for me to explain it is with a quick example:

dylan ~ # layman -L | grep java
* haskell                   [Darcs     ] (source: http://www.haskell.org/~ge...)
* java-gcj-overlay          [Subversion] (source: http://overlays.gentoo.org...)
* java-overlay              [Subversion] (source: http://overlays.gentoo.org...)

dylan ~ # layman -L | grep java > layman_output.txt
* haskell                   [Darcs     ] (source: http://www.haskell.org/~ge...)

dylan ~ # cat layman_output.txt 
* java-gcj-overlay          [Subversion] (source: http://overlays.gentoo.org...)
* java-overlay              [Subversion] (source: http://overlays.gentoo.org...)



Reproducible: Always

Steps to Reproduce:
1.layman -L | grep java > layman_output.txt


Actual Results:  
The results of any overlays missing support tools, eg. missing Darcs for the haskell overlay, are not passed to stdout, hench, not being redirected to the output file. They do still appear on screen, although they shouldn't.

Expected Results:  
One would expect all output to be sent to stdout.

This looks similar to bug 217444.

[Comment 1 Gunnar Wrobel (RETIRED) 2008-06-02 16:22:07 UTC]

Errors are never directed to STDOUT but to STDERR. Basic unix stuff.

Granted: one could discuss printing non-supported overlays also to STDOUT but I really don't see a good reason for that. 

Maybe I misread your bug report?