Re: [Bug-apl] Cannot find libemacs on older mac
Hi, from the GNU APL printout it looks like the file was found, but something is wrong with the file: file /Users/alexey/Applications/gnu-apl/lib/apl/libemacs.dylib ( flat namespace in /Users/alexey/Applications/gnu-apl/lib/apl/libemacs.dylib ) So GNU APL was able to open the file (using dlopen()) but then gets the "flat namespace" error at some point. The second link below says: amandogra commented on Oct 2, 2016 It got solved by OS upgrade. For someone who comes to this... I was on Mac OSX El Capitan, but I upgraded to macOS Sierra. Then, I executed the following command brew unlink libevent && brew link libevent This fixed my issue. The posting was related to a different library but the (rest of the) error message looks very similar (considering that GNU APL cuts off the start the error message text until the last : in it). The full error message was most likely: Expected in: flat namespace Maybe "nm -D" on the dylib file reveals some more information. Best Regards, /// Jürgen On 12/04/2017 04:06 AM, Elias Mårtenson wrote: Perhaps adding the path to the dylib file in DYLIB_LIBRARY_PATH will help? Don't forget to exprt the variable before starting apl. Regards, Elias On 4 December 2017 at 01:56, Juergen Sauermannwrote: Hi, I haven't changed anything in that area. One poster in one of the links below mentioned that an OS-X upgrade had fixed a similar problem that the poster had, so my impression is that this is a problem than cannot be fixed inside GNU APL. Best Regards, /// Jürgen On 12/03/2017 12:02 PM, Elias Mårtenson wrote: 10.6.8 is indeed very old. I thought my Macbook was old and it's running 10.10. I have to admit that I haven't tried the Emacs mode on OSX in a while since I don't use Macs anymore. That said, last time I did try, it worked fine. Jürgen, could there be an issue with the way the libraries are built? Did anything change in the last couple of years? Regards, Elias On 3 December 2017 at 18:58, Juergen Sauermann wrote: Hi Alexey, I am not really familiar with emacs or OS X, so Elias is probably better informed than me. Unfotunately shared libraries are rather platform dependent and so are the error messages that they produce. From the output below it seems that the file libemacs.dylib exists and was found, but could not be opened with dlopen(). It might be that your platform needs additional linker flags for dynamic libraries to work, but I have no idea which ones. See also: https://github.com/tmux/tmux/issues/547 or https://developer.apple.com/library/content/documentation/Porting/Conceptual/PortingUnix/compiling/compiling.html Best Regards, /// Jürgen
[Bug-apl] [Reproducible builds] add support for SOURCE_DATE_EPOCH
Hi, I'm a member of the Arch Linux Security team working on the reproducible builds initiative. Right now, apl compiles almost reproducibly, with the only caveat of embedding some time-based information into a header file. It would be nice if this could be overriden with an environment variable (or stripped out whatsoever). I'm attaching a patch on how I believe this could be achieved. Thanks! -Santiago. signature.asc Description: PGP signature
Re: [Bug-apl] [Reproducible builds] add support for SOURCE_DATE_EPOCH
I see no patch attached. Xtian. On 2017-12-04 17:53, Santiago Torres-Arias wrote: Hi, I'm a member of the Arch Linux Security team working on the reproducible builds initiative. Right now, apl compiles almost reproducibly, with the only caveat of embedding some time-based information into a header file. It would be nice if this could be overriden with an environment variable (or stripped out whatsoever). I'm attaching a patch on how I believe this could be achieved. Thanks! -Santiago.
Re: [Bug-apl] [Reproducible builds] add support for SOURCE_DATE_EPOCH
Hi, I wonder if it was stripped out. I also opened a bug with the patch. I'm trying to re-attach here. Thanks! -Santiago. On Mon, Dec 04, 2017 at 10:33:02PM -0500, Christian Robert wrote: > I see no patch attached. > > Xtian. > > On 2017-12-04 17:53, Santiago Torres-Arias wrote: > > Hi, > > > > I'm a member of the Arch Linux Security team working on the reproducible > > builds initiative. Right now, apl compiles almost reproducibly, with the > > only caveat of embedding some time-based information into a header file. > > It would be nice if this could be overriden with an environment variable > > (or stripped out whatsoever). > > > > I'm attaching a patch on how I believe this could be achieved. > > > > Thanks! > > -Santiago. > > --- apl-1.7/src/buildtag2017-03-17 10:46:53.0 -0400 +++ apl-1.7.new/src/buildtag2017-11-30 11:53:00.173223737 -0500 @@ -15,8 +15,15 @@ if [ -x ./config.status ] then CONFIGURE_OPTS=`../config.status -V | grep "with options" | sed 's/[^"]*//'` fi -BUILD_DATE=`date -u "+%F %R:%S %Z"` -BUILD_OS=`uname -s -r -m` + +if [ -n "$SOURCE_DATE_EPOCH" ] +then +BUILD_DATE=`date -u "%F %R:%S %Z" --date=@$SOURDCE_DATE_EPOCH` +BUILD_OS="" +else +BUILD_DATE=`date -u "+%F %R:%S %Z"` +BUILD_OS=`uname -s -r -m` +fi BUILD_TAG="\" / $SVNINFO\", \"$BUILD_DATE\", \"$BUILD_OS\", \"$CONFIGURE_OPTS\"" # write buildtag.hh signature.asc Description: PGP signature
[Bug-apl] I'm curious about how will react APL compiled with ...
I am curious about how will react APL compiled with say the right PCRE and the right FFT library available when ran on a machine who has neither of thoses lib*.so availble? will it refuse to start or just "disable" the corrosponding quad-functions ? Xtian.
Re: [Bug-apl] I'm curious about how will react APL compiled with ...
I don't think anyone would expect to be able to to move the apl binary to a system where the correct libraries aren't installed. You can't typically do that with any programs installed from the package managers. After all, that's why package dependencies exist. However, it does raise the question as to what dependencies the Arch (and other distributions) versions GNU APL use? It's possible that if I install it from the repository I don't get Postgres support for example. Regards, Elias On 5 December 2017 at 13:44, Christian Robert wrote: > I am curious about how will react APL compiled with > say the right PCRE and the right FFT library available > when ran on a machine who has neither of thoses lib*.so availble? > > will it refuse to start or just "disable" the corrosponding quad-functions > ? > > Xtian. > >