short and sweet; those vars expand incrementals parsing, including how to mix them. Suggest you update eselect..
bleh, i suggest you go and read the source... it does. However, it might be that i'm missing some vars in some classes. Now, it would have been nice if you told me which vars are missing...
Now, this is more like it. Some variables moved from space to colon serarated since I wrote env.eselect. Included are: - KDEDIRS - CLASSPATH - INFODIR - ROOTPATH Will fix them once I get back home today.
(In reply to comment #1) > bleh, i suggest you go and read the source... it does. Wouldn't have filed the bug if I hadn't read the source; you lack support for this both in 1.0.6 and in svn. > However, it might be that i'm missing some vars in some classes. Yes, you are. > Now, it would have been nice if you told me which vars are missing... You're missing SPACE_SEPARATED and COLON_SEPARATED. These vars are additionals specifying *other* vars to expand, as I said ("those vars expand incrementals parsing, including how to mix them"). Read the src of portage.py:env_update please; pretty simple (whitespace expanded lists of incrementals and colon parsed)....
(In reply to comment #3) > (In reply to comment #1) > > bleh, i suggest you go and read the source... it does. > Wouldn't have filed the bug if I hadn't read the source; you lack support for > this both in 1.0.6 and in svn. So you probably should start to read it properly.... It _IS_ in. for var in ${vars} ; do # Colon separated?... if has ${var} ${PATH_CLASS} ; then ... continue fi # ...everything else is space separated! if has ${var} ${SPECIAL_CLASS} ; then ... continue fi [[ ${var} == LDPATH ]] && continue # Ok, just a non-cummultative var. ... What needs to change a) LDPATH should now be part of the PATH_CLASS. b) Comment before space class is wrong, and SPECIAL_CLASS is a fugly name. Do you still think it is not implemented?
(In reply to comment #4) > (In reply to comment #3) > > (In reply to comment #1) > > > bleh, i suggest you go and read the source... it does. > > Wouldn't have filed the bug if I hadn't read the source; you lack support for > > this both in 1.0.6 and in svn. > So you probably should start to read it properly.... It _IS_ in. First off, stop wasting my time and _listen_ to what I'm saying, and read the damn env_update function in portage.py please. Second- > for var in ${vars} ; do > # Colon separated?... > if has ${var} ${PATH_CLASS} ; then > ... > continue > fi > # ...everything else is space separated! > if has ${var} ${SPECIAL_CLASS} ; then > ... > continue > fi > [[ ${var} == LDPATH ]] && continue > # Ok, just a non-cummultative var. > ... Kudos. The problem there is that SPECIAL_CLASS and PATH_CLASS are hard coded lists of incrementals; note I said 'hard coded', then re-read comment #3; SPACE_SEPARATED and COLON_SEPARATED are vars that env.d entries can specify to add var names into incremental parsing (hence the original subject). > Do you still think it is not implemented? Yep. Do you think these snippy responses without understanding what the bug is about is wise? :)
re-adding to the summary.
> Do you think these snippy responses without understanding what the bug is about > is wise? :) Do you really think your initial bug report short and sweet; those vars expand incrementals parsing, including how to mix them. actually reflects what needs to be done? Found zmedico's thread on gentoo-dev, will work on it.
(In reply to comment #7) > > Do you think these snippy responses without understanding what the bug is about > > is wise? :) > Do you really think your initial bug report > short and sweet; those vars expand incrementals parsing, including how to mix > them. > actually reflects what needs to be done? Considering the summary was "app-admin/eselect - env.eselect lacks COLON_SEPARATED and SPACE_SEPARATED awareness". Combine that with comment #3, would say it's pretty explicit. You either know about those vars, or you don't- if you don't, you ask. You assumed I was being a moron, and assumed wrong (in this case); next time ask if you're unsure please, saves us all a pissing match (goes without saying drop the snippy crap also, doesn't help things and it's a great way to shoot yourself in the foot :). </lecture>... > Found zmedico's thread on gentoo-dev, will work on it. Thank you; please make sure it's a 2 stage pass for handling of it; his original thread wasn't explicit about it, but a 07foo adding VARX to COLON_SEPARATED *does* affect the parsing of 05bar's VARX setting, despite it being set after the initial parsing of 05bar.
Fixed in SVN.
eselect-1.0.7 is out.
