[MacPorts] #70539: base: failure for building some ports (depending on conditions) when Macports is installed at deeper directory hierarchy with "--prefix" option

MacPorts noreply at macports.org
Fri May 23 21:58:55 UTC 2025


#70539: base: failure for building some ports (depending on conditions) when
Macports is installed at deeper directory hierarchy with  "--prefix"
option
----------------------------+--------------------
  Reporter:  nanigashi-uji  |      Owner:  jmroot
      Type:  defect         |     Status:  closed
  Priority:  Normal         |  Milestone:
 Component:  base           |    Version:  2.10.0
Resolution:  fixed          |   Keywords:
      Port:                 |
----------------------------+--------------------

Comment (by RJVB):

 There is at least 1 issue with this: `mkdtemp` on Linux requires at least
 6 (six) random characters.

 I also can't really predict what happens when you keep workdirs around for
 incremental building (as a port maintainer). Is the symlink from the
 "traditional" location to the new location enough to fool `cmake` and
 `meson` into thinking nothing has changed?

 Also, what about name clashes when you build multiple versions of a port
 from multiple port trees?

 I suppose no one has tried to determine if there isn't a simple way to
 determine whether the legacy mechanism could lead to problems, e.g. by
 doing some stats on the length of `worksrcpath`? This new scheme *could*
 be optional, which means crud in the root build directory and a bit easier
 manual cleanup.

-- 
Ticket URL: <https://trac.macports.org/ticket/70539#comment:14>
MacPorts <https://www.macports.org/>
Ports system for macOS


More information about the macports-tickets mailing list