1 |
Hi, |
2 |
|
3 |
So I figured I better write the patch for dahdi-tools against musl ... |
4 |
so I proceed to download a stage3 tar from |
5 |
http://distfiles.gentoo.org/experimental/amd64/musl/ (vanilla). |
6 |
|
7 |
I extract it, and mount --rebind /dev, /proc and /sys, and copy in |
8 |
/etc/resolve.conf ... |
9 |
|
10 |
chroot ... and so far so good. |
11 |
|
12 |
emerge --sync |
13 |
emerge -uDNav @world. |
14 |
|
15 |
And this blows up on pam-1.3.1-r2. Looks like |
16 |
https://bugs.gentoo.org/687234. |
17 |
|
18 |
I've seen mention of a musl overlay? |
19 |
|
20 |
What's the best way to proceed? |
21 |
|
22 |
As it stands I can't "make prepare" gentoo-sources (which 4.19.X) is |
23 |
apparently the newest available for musl profile. Reports seems to |
24 |
indicate that this be related to "old linux-headers" (which is at ). |
25 |
|
26 |
Kind Regards, |
27 |
Jaco |
28 |
|
29 |
On 2020/03/27 07:43, Jaco Kroon wrote: |
30 |
> Hi, |
31 |
> |
32 |
> On 2020/03/27 03:25, Joshua Kinard wrote: |
33 |
>> On 3/23/2020 04:21, Jaco Kroon wrote: |
34 |
>>> Hi, |
35 |
>>> |
36 |
>>> https://bugs.gentoo.org/713668 relates. |
37 |
>>> |
38 |
>>> * Searching for /usr/include/execinfo.h ... |
39 |
>>> sys-libs/glibc-2.29-r7 (/usr/include/execinfo.h) |
40 |
>>> |
41 |
>>> As I see I can either add an explicit depend on glibc which I'd prefer |
42 |
>>> not to. Or someone from the musl team could possibly assist on how to |
43 |
>>> get the backtrace() set of calls on musl please? |
44 |
>>> |
45 |
>>> Alternatively I need to add a test and simply path debug.c to only |
46 |
>>> provide stub function for print_backtrace(FILE *fp) that just does |
47 |
>>> fprintf(fp, "No backtrace() available to print a backtrace.\n"); |
48 |
>>> |
49 |
>>> Suggestions? |
50 |
>>> |
51 |
>>> Kind Regards, |
52 |
>>> Jaco |
53 |
>> Some quick searching on google, it looks like the cleanest fix for that bug |
54 |
>> is dahdi-tools needs to be patched to only include execinfo.h or only use |
55 |
>> backtrace() on glibc-based systems, and that patch then sent to the |
56 |
>> dahdi-tools upstream developers for inclusion in a future release. That |
57 |
>> way, we're not dragging that patch around forever in the tree or in the musl |
58 |
>> overlay. |
59 |
> Thanks. I'll see action accordingly. |
60 |
> |
61 |
>> It also doesn't look like musl itself will ever implement execinfo.h or |
62 |
>> backtrace(), per this message in 2015 from the lead musl developer: |
63 |
>> https://www.openwall.com/lists/musl/2015/04/09/3 |
64 |
>> |
65 |
> Implementing libunwind is overkill for my needs, I'll be happy to help |
66 |
> push things upstream if somebody else cares enough to implement. |
67 |
> |
68 |
> Kind Regards, |
69 |
> Jaco |
70 |
> |
71 |
> |