| 1 |
> on interix, a missing or wrong interpreter is not handled correctly, and |
| 2 |
> thus leads to crashes without appropriate error messages. i'm really |
| 3 |
> glad portage now checks for those ;) |
| 4 |
> |
| 5 |
|
| 6 |
If bootstrapping dies with dozends of packages for no other reason |
| 7 |
than invalid shebangs, and if nobody fixes them even in the |
| 8 |
fundamental packages, doesn't this mean, that nobody tries to bootsrap |
| 9 |
at all any more? With the exeption of mine. |
| 10 |
|
| 11 |
Nobody could bootstrap without either fixing all this packages or |
| 12 |
commenting out the call to die. What I want to say is, that I see some |
| 13 |
danger that the Prefix projekt kills itself by this blocker. People |
| 14 |
don't get started at all. |
| 15 |
|
| 16 |
> interix is more like a real unix than cygwin; it uses symlinks, etc. |
| 17 |
> also, it does _not_ use PATH to search for libraries, but has propper |
| 18 |
> support for rpath, etc. |
| 19 |
> |
| 20 |
> additionally the x86-winnt profiles, which i started, use the parity |
| 21 |
> compiler for windows [1], which adds support for rpath, lazy loading, |
| 22 |
> unix like shared library building, libtool support, etc, etc. so even on |
| 23 |
> windows, there is not need for PATH hacks, and such. |
| 24 |
> |
| 25 |
|
| 26 |
The parity compiler sounds very ambitious and I am curious. On the |
| 27 |
other hands I also see the advantages in Microsofts PATH approach. |
| 28 |
Dynamic libraries doen't depend on a special path any more and can be |
| 29 |
moved around. All you have to do, is to adapt the PATH variable to the |
| 30 |
new location. That makes your programs more portable. |
| 31 |
|
| 32 |
Fortunatly you can do the some on Unix by use of LD_LIBRARY_PATH. However: |
| 33 |
|
| 34 |
<quote http://en.wikipedia.org/wiki/Rpath_%28linking%29> |
| 35 |
The primary disadvantage of using RPATH is that it overrides the |
| 36 |
LD_LIBRARY_PATH settings which makes things like running a precompiled |
| 37 |
binary out of a user's home directory or some other non-default |
| 38 |
location difficult or impossible. Use of RPATH also makes it |
| 39 |
difficult, if not impossible, to upgrade libraries without forcing a |
| 40 |
reinstallation of all the software dependent on (even the older |
| 41 |
versions of) the libraries. |
| 42 |
</quote> |
| 43 |
|
| 44 |
What I am pondering on, is a relative RPATH, relative to the prefix |
| 45 |
path. By this the Prefix installation as a whole could still be moved |
| 46 |
around without breaking. You could run a precompiled PREFIX |
| 47 |
installation out of different user's home directories. |
| 48 |
|
| 49 |
> |
| 50 |
> P.S.: i built cygwin support into parity - would be curious if (after i |
| 51 |
> fixed the portage prefix-chaining patch, which is a prerequisite) |
| 52 |
> x86-winnt profiles would work unmodified on cygwin. |
| 53 |
> |
| 54 |
|
| 55 |
If you want to reach that goal you will reach it. Cygwin isn't |
| 56 |
difficult. There are a few cases when the .exe magic gets into your |
| 57 |
way, but they are usually easy to fix. |
| 58 |
|
| 59 |
Specially you need the Cygwin patch for coreutils to reflect the .exe |
| 60 |
magic and that patch conflicts with the interix-warn-mount.patch. |
| 61 |
|
| 62 |
Al |
Archives