1 |
On 20-11-2017 11:53:32 -0600, R0b0t1 wrote: |
2 |
> Hello friends! |
3 |
> |
4 |
> On Monday, November 20, 2017, Sergei Trofimovich <[1]slyfox@g.o> wrote: |
5 |
> > On Sun, 19 Nov 2017 22:47:35 -0600 |
6 |
> > R0b0t1 <[2]r030t1@×××××.com> wrote: |
7 |
> > |
8 |
> >> Understanding an existing codebase should not be a technical |
9 |
> >> challenge. I had to resort to reimplementing all of the steps myself, |
10 |
> >> in part to understand if they were done properly in the first place. |
11 |
> > |
12 |
> > Looks like you are an expert in this area now and are perfectly capable |
13 |
> > of submitting the patches. I can review them at least for crossdev. |
14 |
> > |
15 |
> |
16 |
> In my goal to understand bootstrap-rap I am still in the process of creating |
17 |
> something crossdev-like that can be used to generate a toolchain. |
18 |
> |
19 |
> A recurring problem I have had is that this set of related tasks - generating |
20 |
> cross compilers and packages, generating an initramfs, or generating a prefixed |
21 |
> pseudoinstallation - all start by reimplementing some subset of portage. |
22 |
> |
23 |
> For prefix/RAP it makes sense, for the others possibly not. |
24 |
|
25 |
You may also want to understand that cross-compiling (or compilers) in |
26 |
itself is a very difficult topic to get right. Mixing that with Prefix |
27 |
FAICT never got out of the lab-setting in which it was attempted. |
28 |
|
29 |
> >> Unfortunately these are things that the original authors should |
30 |
> >> produce. Experience has shown me that documentation written by |
31 |
> >> ancillary contributors that do not have deep experience with the code |
32 |
> >> base is poor. |
33 |
|
34 |
Like Benda said, documentation can always be improved. |
35 |
|
36 |
In the case for bootstrap-prefix, it used to be documented in terms of |
37 |
steps and why one had to do them that way. Somewhere at the start of |
38 |
2006, when there was like 150 packages, and one arch (ppc-macos), said |
39 |
script didn't exist. As it stands today, the key decisions and |
40 |
workarounds are actually documented, but as RAP actually shows, if you |
41 |
only focus on a specific use-case, you can get rid of a lot of (what |
42 |
appears to be) nonsense. It's the context in which you look at the |
43 |
projects you refer to. |
44 |
|
45 |
> Yes, that is what I am doing with my own code as I have the time to |
46 |
> write it. I mostly still have no idea what is going on in the already |
47 |
> written code. |
48 |
|
49 |
Perhaps open up the dicussion on the related project's mailing lists. |
50 |
At least I haven't come across any request to explain certain |
51 |
bits/decisions yet. |
52 |
|
53 |
Thanks, |
54 |
Fabian |
55 |
|
56 |
-- |
57 |
Fabian Groffen |
58 |
Gentoo on a different level |