<div dir="ltr"><div>Thanks Julien!</div><div><br></div><div>Sorry I wasn't clear enough. <br></div><div><br></div><div>I wasn't trying to suggest a change to mmc, just find out if I was about to do something that would cause intermittent faults, random link errors or something else irritating.</div><div><br></div><div><br></div><div>Cheers!</div><div><br></div><div>Charles<br></div></div><br><div class="gmail_quote"><div dir="ltr">On Thu, Sep 13, 2018 at 8:38 PM Julien Fischer <<a href="mailto:jfischer@opturion.com">jfischer@opturion.com</a>> wrote:<br></div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><br>
Hi Charles,<br>
<br>
On Thu, 13 Sep 2018, Charles Shuller wrote:<br>
<br>
> It seems like it would be a bit nicer if the standalone interface<br>
> object file were added to the mercury library by doing something<br>
> like: <br>
><br>
> ar rs libmercury_lib.a libmercury_lib_interface.o<br>
> <br>
> Will this break something?? It's not clear to me if the interface is<br>
> kept separate for a good reason or if it just happens to have been<br>
> implemented that way.<br>
<br>
You are assuming that there is a library in to which the object file<br>
could be included being built at the same time as the standalone<br>
interface. While that's true in your case and in the example from<br>
the samples directory, it isn't neccesarily true in general; the set<br>
of libraries for which a standlone interface is being generated may<br>
have alredy been built.<br>
<br>
That said, it would probably be possible to addtionally support<br>
something along the lines of<br>
<br>
$ mmc --include-standalone-interface --make libmylib<br>
<br>
which would cause the compiler to build the library, generate the<br>
standalone interface and then include it in the archive or shared<br>
object.<br>
<br>
Julien.</blockquote></div>