mailing list of musl libc
 help / color / mirror / code / Atom feed
From: Rich Felker <dalias@aerifal.cx>
To: musl@lists.openwall.com
Subject: Re: [ANN] sabotage 0.9.7 milestone (x86[_64]) release
Date: Thu, 8 Nov 2012 17:40:32 -0500	[thread overview]
Message-ID: <20121108224032.GY20323@brightrain.aerifal.cx> (raw)
In-Reply-To: <20121108142949.828d3bb8.idunham@lavabit.com>

On Thu, Nov 08, 2012 at 02:29:49PM -0800, Isaac Dunham wrote:
> On Wed, 7 Nov 2012 20:31:04 -0500
> Rich Felker <dalias@aerifal.cx> wrote:
> 
> > 
> > > > > Is there a reason Opera would not work with this release?  Its only
> > > > > external dependencies are X11 and freetype/fontconfig.
> > > >
> > > > And glibc.
> 
> And libstdc++.

If it's using a shared libstdc++, then it would either need to be run
with a different library path pointing at the glibc libstdc++.so, or
musl will need to provide a sufficiently compatible ABI, in the C++
sense, for the musl-built libstdc++ to satisfy the needs of a
glibc-linked binary.

> > I'm not clear on whether any of the issues affect Opera, but for such
> > a large program I'd be fairly surprised if they don't (albeit happily
> > surprised, of course).
> Here's what I got (using a somewhat out-of-date libc.so; I know at
> least one function has since been added, but this gives a general
> idea of the scope):
> $ /opt/musl/lib/libc.so lib/opera/opera 
> Error relocating /usr/lib/libstdc++.so.6: __nl_langinfo_l: symbol not found
> Error relocating /usr/lib/libstdc++.so.6: __strftime_l: symbol not found
> Error relocating /usr/lib/libstdc++.so.6: strtold_l: symbol not found
> Error relocating /usr/lib/libstdc++.so.6: __wctype_l: symbol not found
> Error relocating /usr/lib/libstdc++.so.6: __wcsftime_l: symbol not found
> Error relocating /usr/lib/libstdc++.so.6: __towupper_l: symbol not found
> Error relocating /usr/lib/libstdc++.so.6: __strxfrm_l: symbol not found
> Error relocating /usr/lib/libstdc++.so.6: __strtod_l: symbol not found
> Error relocating /usr/lib/libstdc++.so.6: __wcsxfrm_l: symbol not found
> Error relocating /usr/lib/libstdc++.so.6: __uselocale: symbol not found
> Error relocating /usr/lib/libstdc++.so.6: __freelocale: symbol not found
> Error relocating /usr/lib/libstdc++.so.6: __strcoll_l: symbol not found
> Error relocating /usr/lib/libstdc++.so.6: __towlower_l: symbol not found
> Error relocating /usr/lib/libstdc++.so.6: __iswctype_l: symbol not found
> Error relocating /usr/lib/libstdc++.so.6: __strtof_l: symbol not found
> Error relocating /usr/lib/libstdc++.so.6: __duplocale: symbol not found
> Error relocating /usr/lib/libstdc++.so.6: __newlocale: symbol not found
> Error relocating /usr/lib/libstdc++.so.6: __wcscoll_l: symbol not found

All of these would be fixed if we could use a musl-built libstdc++.so
instead of the glibc-built one.

> Error relocating /lib/libgcc_s.so.1: dl_iterate_phdr: symbol not found

Fixed.

> Error relocating lib/opera/opera: __pthread_register_cancel: symbol not found

Interesting. I had not really thought about what could be done to
satisfy the glibc thread-cancellation ABI. For musl-linked programs,
we use the ancient (glibc 2.1 era) cancellation push/pop functions
which are much cleaner and don't need special GCC hacks. But there
should be some way to provide at least an ABI layer for the glibc way.

> Error relocating lib/opera/opera: memmem: symbol not found

Also fixed; we have memmem now.

> Error relocating lib/opera/opera: __res_ninit: symbol not found

This could probably be stubbed out. musl's resolver is completely
stateless.

> Error relocating lib/opera/opera: __pthread_unregister_cancel: symbol not found
> Error relocating lib/opera/opera: qsort_r: symbol not found

Should be easy to add.

> Error relocating lib/opera/opera: __sigsetjmp: symbol not found

This is just another symbol alias.

Rich


  reply	other threads:[~2012-11-08 22:40 UTC|newest]

Thread overview: 19+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2012-11-06 16:34 John Spencer
2012-11-06 18:38 ` Rich Felker
2012-11-07  6:04 ` Isaac Dunham
2012-11-09  1:38   ` lightweight web browsers (was: [ANN] sabotage 0.9.7 milestone (x86[_64]) release) John Spencer
2012-11-09  3:57     ` idunham
2012-11-09  4:02       ` Rich Felker
2012-11-09  6:20         ` Jens Staal
2012-11-07 12:38 ` [ANN] sabotage 0.9.7 milestone (x86[_64]) release Kurt H Maier
2012-11-07 13:13   ` Rich Felker
2012-11-07 13:32     ` Daniel Cegiełka
2012-11-08  1:31       ` Rich Felker
2012-11-08 22:29         ` Isaac Dunham
2012-11-08 22:40           ` Rich Felker [this message]
2012-11-09  1:58           ` Kurt H Maier
2012-11-09  9:08             ` Szabolcs Nagy
2012-11-08 17:24 ` Jens Staal
2012-11-09  1:27   ` John Spencer
2012-11-09  6:04     ` Jens Staal
2012-11-09 14:27       ` John Spencer

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20121108224032.GY20323@brightrain.aerifal.cx \
    --to=dalias@aerifal.cx \
    --cc=musl@lists.openwall.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
Code repositories for project(s) associated with this public inbox

	https://git.vuxu.org/mirror/musl/

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).