[Ecls-list] ffi broken in windows

Dustin Long dlong at stevens.edu
Wed Dec 26 22:45:26 UTC 2007


Dustin Long wrote:
> Juan Jose Garcia-Ripoll wrote:
>   
>> In short: I do not want to remove this feature, and some other uses as
>> well, since I got complaints the very first time I disabled it.
>>
>> But worry not, for there your problem and fasl loading are two
>> different things. You want to be able to reload libraries that have
>> changed. Well, one may add a flag to ecl_load_library and use one
>> version (the one that does not duplicate files) for the FFI and the
>> other version for implementing LOAD 
>>     
Looking back on it, I'm not sure if your comments on my last patch were 
correct. My patch only changed whether or not the immediate API call was 
used to reload the .fasl file (in ecl_library_open), it would still 
rerun initialization code (si_load_binary), whether or not the .fasl has 
already been loaded. At least, I made a simple file that just printed at 
the top-level and even with my patch it would print again everytime 
(load ...) was run.

But in case there was something more subtle I was missing, I put 
together what you suggested. Ffi and load act differently when loading 
libraries. Patch is attached.


I've also run into a strange slime error that is unrelated to this (it 
shows up with or without my patch). Whenever I use 
slime-compile-and-load I get the error "Cannot flush input stream" on 
the name of the current lisp buffer. I'm currently looking into what's 
making this happen.

Dustin
-------------- next part --------------
A non-text attachment was scrubbed...
Name: 06_2007-12-26-load-foreign-library.patch
Type: text/x-patch
Size: 2037 bytes
Desc: not available
URL: <https://mailman.common-lisp.net/pipermail/ecl-devel/attachments/20071226/d6a72b6a/attachment.bin>


More information about the ecl-devel mailing list