diff options
| author | Tor Lillqvist <tlillqvist@suse.com> | 2012-10-03 16:39:54 +0300 |
|---|---|---|
| committer | Tor Lillqvist <tlillqvist@suse.com> | 2012-10-03 16:39:54 +0300 |
| commit | 325309a7774e10fea79d92f424ae398ea242d142 (patch) | |
| tree | fc7478f955781717c4c01c74fb8765b3597e775e /expat/README | |
| parent | 7bd77ca9974b9467df38921c55ed80b3e5d3dae3 (diff) | |
Mention the disable-dynloading issue
Change-Id: I6d2349b8c4d97d266faebdb630c65e0441e25b18
Diffstat (limited to 'expat/README')
| -rw-r--r-- | expat/README | 9 |
1 files changed, 9 insertions, 0 deletions
diff --git a/expat/README b/expat/README index 1af988937eb0..9289284ec470 100644 --- a/expat/README +++ b/expat/README @@ -36,5 +36,14 @@ libexpat in builds where that is used, why is a separate Unicode one needed when an internal expat is used? Good question. Next question. Patches welcome. +Another interesting issue is the disable-dynloading case: When we link +all needed LO code into one executable (as for iOS apps) or into one +shared library (as in a tentative approach for Android apps) there is +a problem with linking two libraries (ascii_expat_xmlparse and +expat_xmlparse) that contain exactly the same symbols... Obviously +just the one that happens to be mentioned first on the linking command +line gets linked in, and code that then thinks it is calling the other +one will mysteriously crash or silently fail. Working on that... + From: [http://expat.sourceforge.net/] |