2017-11-26 8:21 GMT+09:00 Ilya Mikhaltsou <[email protected]>: > > That's the case of recent macOS versions. > > > > > > Judging from the code, 10.3 returns libruby.a as $librubyarg while it > doesn't have such a static library, but does have a shared library. As a > result, RUBY_LIBS is set to something like "libruby.a -L/.../" which > causes link error. Therefore, libruby.a must be replaced with -lruby if > $librubydir/libruby.a doesn't exist and if $librubyarg is equal to libruby.a > > > > I couldn’t find any reference neither to this nor to such problem with > LIBRUBYARG in RbConfig on OS X 10.3. Can you please substantiate this by > some links to discussions? >
Read the code of the current version of configure.ac. > > The correct variable is $librubya. > > > > > > No. Our goal is to set RUBY_LIBS to canonical "-L$rubylibdir > $librubyarg", which would look like, e.g., "-L/usr/local/lib -lruby". Only > for 10.3 (or systems having the same RbConfig bug), $librubyarg must be > replaced with -lruby to avoid link error. We don't need to care about > libruby-X.Y..Z-static.a at all from the outset. > > > > Our goal is never that, since vim attempts to use static library on other > unixes. The specific case for macOS is exactly due to inexistence of a > static library there. > No. If vim were to make any attempt to link against a static library just like you claim, the code would be if test -f "$rubylibdir/$librubya"; then librubyarg="$librubya" instead of the current if test -f "$rubylibdir/$librubya"; then librubyarg="$librubyarg" Currently, if $rubylibdir/$librubya exists, then it uses $librubyarg instead of $librubya. Basically, $librubyarg is '-lruby'. However, as the code indicates, there's a case where $librubyarg is set to libruby.a. Only for that case, libruby.a is chosen (This is OK since it turns out to exist). Otherwise, -lruby is chosen even if libruby.a exists. > > What we need to care are two things: (1) the in-existence of the static > library and (2) the sanity of the ruby configuration. When both are met, we > decide that there simply is no static library at all. > Too abstract for me to understand. > > > > Besides that, as I already wrote above, your current patch makes the > code about macOS completely useless. And I’m trying to ask you what is the > point of it. > > > > > > > > You argued on the basis of a wrong assumption (See above). Rather, your > patch breaks 10.3 build. > > > > Unfortunately, I don’t have OS X 10.3. I would wonder if this is truly the > only thing that breaks it and if the current version even builds there. As > well, as I would wonder if there is anyone left wishing to build latest vim > on OS X 10.3. > Note that the question was made last May: https://stackoverflow.com/questions/43956208/does-mac-os-x-have-the-dev-ptmx-file , and that Patch 8.0.1166 was merged last September. You can't drop the support for old Mac OS X's in favor of your patch. > -- > -- > You received this message from the "vim_dev" maillist. > Do not top-post! Type your reply below the text you are replying to. > For more information, visit http://www.vim.org/maillist.php > > --- > You received this message because you are subscribed to the Google Groups > "vim_dev" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to [email protected]. > For more options, visit https://groups.google.com/d/optout. > -- -- You received this message from the "vim_dev" maillist. Do not top-post! Type your reply below the text you are replying to. For more information, visit http://www.vim.org/maillist.php --- You received this message because you are subscribed to the Google Groups "vim_dev" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. For more options, visit https://groups.google.com/d/optout.
