<div class="gmail_quote">2009/12/16 Rémi Denis-Courmont <span dir="ltr"><<a href="mailto:remi@remlab.net">remi@remlab.net</a>></span><br><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
<div class="im"><br>
On Wed, 16 Dec 2009 11:29:24 +0100 (CET), <a href="mailto:git@videolan.org">git@videolan.org</a> (git version<br>
control) wrote:<br>
> vlc | branch: master | Jean-Baptiste Kempf <<a href="mailto:jb@videolan.org">jb@videolan.org</a>> | Wed Dec 16<br>
> 11:29:04 2009 +0100| [5fdee8ec9899456231c2bd584614ba41c4191ed3] |<br>
> committer: Jean-Baptiste Kempf<br>
><br>
> lpcm: abort() is a better solution here<br>
<br>
</div>No, it's much worse. assert() prints a diagnostic message with a source<br>
code file name, a line number and the assertion. abort() just terminates<br>
the process. Thus cannot be traced unless core dumping is enabled or a<br>
debugger is running, neither of which are normally true.<br><br></blockquote><div><br>I dislike both solutions although assert() is preferred above abort().<br><br>In case a stream with LPCM is started in a VLM and other streams are already running, then both assert() and abort() will kill the entire process and thereby disrupting service. In this case I prefer gracefull degredation and the startup of that one stream with LPCM should just fail.<br>
<br>My 2cts<br><br>Kind regards,<br>Jean-Paul Saman.<br></div></div>