Bug 305549 – dev-util/elfkickers tries to install missing documentation

• Product: Gentoo Linux
• Status: RESOLVED FIXED

[Description Diego Elio Pettenò (RETIRED) 2010-02-17 13:59:52 UTC]

This bug has been open because your package reported in the log that dodoc was called for non-existent, files. This is generally a lower-level mistake, but it's annoying as it fills with false positives the log of packages that might forget to die on doins failures.

Please do correct this. Thanks.

Diego, on behalf of the QA team

[Comment 1 Diego Elio Pettenò (RETIRED) 2010-02-17 14:00:04 UTC]

Created attachment 220013 [details]
Build log

[Comment 2 Christophe LEFEBVRE 2010-02-25 19:36:55 UTC]

Created attachment 221183 [details]
elfkickers-2.0a-r4.ebuild (Corrected ebuild)

[Comment 3 Christophe LEFEBVRE 2010-03-01 19:23:10 UTC]

Created attachment 221663 [details]
elfkickers-2.0a-r4.ebuild (Corrected ebuild)

[Comment 4 solar (RETIRED) 2010-03-01 19:49:51 UTC]

(In reply to comment #3)
> Created an attachment (id=221663) [details]
> elfkickers-2.0a-r4.ebuild (Corrected ebuild)

I don't see why you convered this over to an EAPi=2 ebuild. There is no advantage. Also please attach diff's vs full ebuilds.

## not quite right. ##
if [ -f $d/*.1 ]; then


# This above is not a proper expression in bash. 
# If the glob of *.1 expanded to more then one file you would end up with 
# [ -f d/foo.1 d/bar.1 d/baz.1 ] ; # which in bash would yield an error.

# something like for x in $d/*.1 do ; [ -e $x ] && dostuff.
# would be better. Even if the existing glob only matches 1 file.

[Comment 5 Christophe LEFEBVRE 2010-03-02 19:24:32 UTC]

Created attachment 221831 [details, diff]
elfkickers-2.0a-r4.ebuild.diff (Corrected ebuild)

Thanks Solar for your help. I have used EAPI=2 in my first correction because it was a way to delete "cd $S and unpack". Now, I have included a for loop for the man pages and ""EAPI=0"".

[Comment 6 Nathan Phillip Brink (binki) (RETIRED) 2010-03-05 18:42:36 UTC]

(In reply to comment #4)
> I don't see why you convered this over to an EAPi=2 ebuild. There is no
> advantage. Also please attach diff's vs full ebuilds.

There is the advantage of less needlessly reproduced code. Many ebuilds in portage have ``unpack ${A} && cd "${S}" || die'' in their src_unpack()s and EAPI="2" added the src_prepare() function to prevent this.

Getting rid of cd "${S}" also slightly improves the readability of the ebuild. When one sees an overloaded src_prepare() in an ebuild one immediately knows to look for sed and epatch statements. For EAPI="2" ebuilds, src_unpack() is reserved only for the special occasions where a nonstandard archive format is used, the tarballs contain other tarballs, etc. I think such distinctions yield better clarity.

[Comment 7 solar (RETIRED) 2010-03-05 22:11:17 UTC]

(In reply to comment #5)
> Created an attachment (id=221831) [details]
> elfkickers-2.0a-r4.ebuild.diff (Corrected ebuild)
> 
> Thanks Solar for your help. I have used EAPI=2 in my first correction because
> it was a way to delete "cd $S and unpack". Now, I have included a for loop for
> the man pages and ""EAPI=0"".
> 

Thank you. I tested that all still worked and pushed it to the tree.


--
Nathan, I would not convert over an existing ebuild that is stable for arches to another EAPI. That would have to happen in -r5