Hi! I'm new to Debian, this is my first bug report and my english is very bad, so please be patient with me! :-)
I have the same bug, but I think the patch by Brad Jorsch is not the best solution, because the real problem is in the source package. The 'bash-completion_20060301-2.diff.gz' created by the maintainer, Luk Claes, contains: --- bash-completion-20060301.orig/bash_completion +++ bash-completion-20060301/bash_completion @@ -29,12 +29,8 @@ [...] -complete -f -X '!*.@(@(?(e)ps|?(E)PS|pdf|PDF)?(.gz|.GZ|.bz2|.BZ2|.Z))' gv ggv kghostview +complete -f -X '!*.@(@(?(e)ps|?(E)PS|pdf|PDF)?(.gz|.GZ|.bz2|.BZ2|.Z))' gv ggv kghostview kpdf complete -f -X '!*.@(dvi|DVI)?(.@(gz|Z|bz2))' xdvi -complete -f -X '!*.@(dvi|DVI)' dvips dviselect dvitype kdvi dvipdf advi +complete -f -X '!*.@(dvi|DVI)?(.@(gz|bz2))' kdvi +complete -f -X '!*.@(dvi|DVI)' dvips dviselect dvitype dvipdf advi complete -f -X '!*.@(pdf|PDF)' acroread gpdf xpdf kpdf [...] Now, I think the bug is due to the doubled use of 'kpdf': take a look at the first line (modified by Luk) and the last one. BTW, the application kpdf is not capable to open a '.pdf.gz' or similar compressed file (I have never noticed this feature and I made some tests for this), so look at the patch by Brad Jorsch, it modifies the line that Luk Claes left unchanged: --- /etc/bash_completion 2008-02-27 10:53:55.000000000 -0500 +++ /etc/bash_completion 2008-02-27 10:53:57.000000000 -0500 @@ -85,7 +85,7 @@ complete -f -X '!*.@(dvi|DVI)?(.@(gz|Z|bz2))' xdvi complete -f -X '!*.@(dvi|DVI)?(.@(gz|bz2))' kdvi complete -f -X '!*.@(dvi|DVI)' dvips dviselect dvitype dvipdf advi -complete -f -X '!*.@(pdf|PDF)' acroread gpdf xpdf kpdf +complete -f -X '!*.@(pdf|PDF)' acroread gpdf xpdf complete -f -X '!*.@(@(?(e)ps|?(E)PS)?(.gz|.GZ)|cb(r|z)|CB(R|Z)|djv?(u)|DJV?(U)|pdf|PDF|dvi|DVI|gif|jp?(e)g|miff|tif?(f)|pn[gm]|p[bgp]m|bmp|xpm|ico|xwd|tga|pcx|GIF|JP?(E)G|MIFF|TIF?(F)|PN[GM]|P[BGP]M|BMP|XPM|ICO|XWD|TGA|PCX)' evince complete -f -X '!*.@(?(e)ps|?(E)PS)' ps2pdf complete -f -X '!*.texi*' makeinfo texi2html In this way the bug is resolved because 'kpdf' is not ever doubled, but I think that the best solution is not to modify that line, but to remove the one I mentioned in the patch by Luk Claes. Regards. Cristian
