[prev in list] [next in list] [prev in thread] [next in thread] 

List:       gentoo-dev
Subject:    Re: [gentoo-dev] Re: Is it possible to ignore / override parts of binpkg environment locally?
From:       Alec Warner <antarus () gentoo ! org>
Date:       2014-05-19 3:19:52
Message-ID: CAAr7Pr9A5xW=eHXhDAS8iNoW6c+piGYb8ar4J5n7UysMt4L1Lg () mail ! gmail ! com
[Download RAW message or body]

On Sun, May 18, 2014 at 6:29 PM, Leho Kraav <leho@kraav.com> wrote:

> On 19.05.2014 03:11, Leho Kraav wrote:
>
>>
>> Do I now always have to rebuild all the packages to get the new changes
>> to user.eclass included? I'd really like to just ignore the environment
>> file coming with the binpkg and the let the local system determine
>> pretty much everything. Then I could just build the packages once and
>> continue developing on binclient.
>>
>> Option B is if I could just make a list of functions that cannot be
>> overriden by later binpkg environment import. In this case egetent,
>> enewuser, enewgroup. My preliminary experiments with "declare -r" or
>> "readonly" in /etc/portage/bashrc didn't really succeed, probably
>> because the processes during binpkg emerge are not related or I just
>> don't know the right way to do this. It appears in the beginning, some
>> stuff runs from the binclient environment, then everything gets switched
>> to binpkg environment, then final cleanup happens again in binclient
>> environment.
>>
>>
> ${QA_INTERCEPTORS} looks like something really interesting for this.
> Except it looks like it's a hardcoded internal list in bin/ebuild.sh that
> cannot be added to with an outside environment variable?
>
>
QA_INTERCEPTORS are meant to detect when certain binaries are called in
global scope (which is illegal). I wouldn't rely on them for anything else.

-A

[Attachment #3 (text/html)]

<div dir="ltr"><img class="ajT" \
src="https://mail.google.com/mail/u/0/images/cleardot.gif"><br><div \
class="gmail_extra"><br><br><div class="gmail_quote">On Sun, May 18, 2014 at 6:29 PM, \
Leho Kraav <span dir="ltr">&lt;<a href="mailto:leho@kraav.com" \
target="_blank">leho@kraav.com</a>&gt;</span> wrote:<br> <blockquote \
class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc \
solid;padding-left:1ex"><div class="">On 19.05.2014 03:11, Leho Kraav wrote:<br> \
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc \
solid;padding-left:1ex"> <br>
Do I now always have to rebuild all the packages to get the new changes<br>
to user.eclass included? I&#39;d really like to just ignore the environment<br>
file coming with the binpkg and the let the local system determine<br>
pretty much everything. Then I could just build the packages once and<br>
continue developing on binclient.<br>
<br>
Option B is if I could just make a list of functions that cannot be<br>
overriden by later binpkg environment import. In this case egetent,<br>
enewuser, enewgroup. My preliminary experiments with &quot;declare -r&quot; or<br>
&quot;readonly&quot; in /etc/portage/bashrc didn&#39;t really succeed, probably<br>
because the processes during binpkg emerge are not related or I just<br>
don&#39;t know the right way to do this. It appears in the beginning, some<br>
stuff runs from the binclient environment, then everything gets switched<br>
to binpkg environment, then final cleanup happens again in binclient<br>
environment.<br>
<br>
</blockquote>
<br></div>
${QA_INTERCEPTORS} looks like something really interesting for this. Except it looks \
like it&#39;s a hardcoded internal list in bin/ebuild.sh that cannot be added to with \
an outside environment variable?<br> <br>
</blockquote></div><br></div><div class="gmail_extra">QA_INTERCEPTORS are meant to \
detect when certain binaries are called in global scope (which is illegal). I \
wouldn&#39;t rely on them for anything else.</div><div class="gmail_extra"> \
<br></div><div class="gmail_extra">-A</div><div class="gmail_extra"><br></div></div>



[prev in list] [next in list] [prev in thread] [next in thread] 

Configure | About | News | Add a list | Sponsored by KoreLogic