<div dir="ltr"><div><div>Gary,<br><br></div><div>First of all, congratulations on getting this far! I know from experience that<br></div><div>it's a non-trivial exercise :-(<br></div><br><div class="gmail_quote"><div dir="ltr" class="gmail_attr">On Sun, Nov 17, 2019 at 9:39 PM Gary Mills <<a href="mailto:gary_mills@fastmail.fm" target="_blank">gary_mills@fastmail.fm</a>> wrote:<br></div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">I have a series of questions to ask the members of this mailing list.<br>
I have information to share with you as well. I already have partial<br>
answers to most of the questions, but I'd still like to hear from you.<br>
<br>
First of all, how much interest do you have in OI on SPARC? My<br>
interest is as a developer and tester. What is your interest?<br>
I'd like to determine the size of the audience.<br></blockquote><div><br></div><div>My interest is indirect; it's good to see that the illumos on SPARC ecosystem<br></div><div>has more participants.<br></div><div> </div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">
How should I contribute to OI on SPARC? I have plans to build more<br>
packages, and to do so with fewer changes to the OI source. I also<br>
have plans to update the distribution from the current 2018 to 2019 or<br>
2020. Does this sound reasonable to you?<br></blockquote><div><br></div><div>I think you need to decide what the bounds are (in terms of exactly what<br></div><div>packages you're going to include - I know that there's a reasonable number<br></div><div>of packages I ignore because they are broken or not interesting, a lot of the<br></div><div>desktop stuff seems less than worthwhile for example, and of course there's<br></div><div>a whole pile of stuff like Go and Node that's never going to work).<br><br></div><div>Keeping up to date is a good idea. From experience, if you start to drift, it's a<br></div><div>devil to catch up.<br></div><div> </div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">
How should you contribute to OI on SPARC? I've filed bug reports<br>
for many of the changes I've made. They can be seen at:<br>
<br>
<a href="https://www.illumos.org/projects/openindiana/issues" rel="noreferrer" target="_blank">https://www.illumos.org/projects/openindiana/issues</a><br>
<br>
I've attached patches to each bug report, but in order for these<br>
patches to be integrated into the OI source, the patches need to be<br>
turned into PRs for github. They also need to be tested on x86 to<br>
make sure they don't accidentally change anything there. Can you help<br>
with any of this? Can you build packages for SPARC from OI source?<br>
Can you help in any other way?<br></blockquote><div><br></div><div>Generally I'm building a lot of stuff on both x86 and SPARC. If you come<br></div><div>across problems, then I'm happy to help. (Although my normal response to<br></div><div>anything that doesn't get along is to excise it for the time being and find<br>something more cooperative to work on.)<br></div><div> </div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">
What type of repository do you prefer? Should it be file-based, as it<br>
is now, or should it be remote, as for OI x86? The repository will<br>
only get larger, as people build more packages and publish them. Will<br>
you download such a large file? I don't know of any way to merge<br>
repositories, so there must be only one.<br></blockquote><div><br></div><div>The one-file repo is fine for a single shot release, for people who only<br></div><div>have one system. (Although I imagine that people who have multiple<br></div><div>SPARC systems are thin on the ground, sadly.) But as soon as you<br></div><div>start considering updates, the download a massive blob model is no<br></div><div>longer viable.<br></div><div><br></div><div>And I think you should keep the sparc repository (if any) completely separate<br></div><div>from the x86 repo the project runs. You'll likely be looking at completely<br></div><div>different needs and policies.<br></div><div> </div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">
Finally, who should coordinate OI on SPARC? Should this be done by<br>
the OI project, or should it be done separately? </blockquote><div><br></div><div>You're the one doing the work, so that's you. Having the project involved<br></div><div>is a drag on their time and a blocker on yours; keeping it as loosely<br></div><div>coupled as possible is best.<br><br></div><div>(This is a general theme; generally the x86 and SPARC sides will want<br></div><div>to avoid treading on each other's toes. And I occasionally find myself tripping<br></div><div>over myself on Tribblix, where I'm both!)<br></div><div> <br></div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">Keep in mind that OI<br>
SPARC uses OI source. Most of it builds and publishes with no<br>
changes. When changes are necessary, my plan has been to introduce<br>
them with no harm to x86 packages built from the same source. Indeed,<br>
some of the changes fix bugs in the corresponding x86 packages. Also<br>
keep in mind that IPS is designed to handle multiple architectures,<br>
making it easy to integrate SPARC with x86. In fact, this is already<br>
done for illumos.<br></blockquote></div><br></div><div>-- <br><div dir="ltr">-Peter Tribble<br><a href="http://www.petertribble.co.uk/" target="_blank">http://www.petertribble.co.uk/</a> - <a href="http://ptribble.blogspot.com/" target="_blank">http://ptribble.blogspot.com/</a></div></div></div>