Hi Joep, hi guys,
I'm quite surprised by replies here, where few bytes would be used to put a
version number in a way or another. I can remember discussions about
optimizing code space, the most we can... 0.5% of code space for this seems
huge, don't you think ?
Putting version number in library file is not useful IMO. If the problem is
to be able to remember which files at which version were used, SVN can help
again: one can checkout jallib repos at a given revision ("svn co -r1298
..."). This revision number would used to retrieve the exact same code,
years later, if any needed (except if we delete it, which is, as William
stated, not the purpose of GC). Using code from the trunk at HEAD revision
(alway most up-to-date), withtout fixing a revision number isn't a good
practice for project in production state or the like. Production project
should also use released package (and their equivalent from SVN using
"branches" or "tags" maps).
Again (from my jaluino post before this topic is moving here), putting a
revision number in files can be done automatically (set by SVN at commit
time), but isn't a good practice, as this is a duplication of information
(you can have it with SVN metadata and svn client program), and as this can
be harmful for merging process (it gives meaningless conflicts).
Finally, please note compiler binaries are part of jallib repository, and
thus can be checked out, just like other jal files. You should be able to
retrieve all the exact files needed to reproduce (compiler + libraries).
Being able to get these files, at their correct version, isn't something
related to jallib but to users using it, and to their process of releasing
files. We can provide best practice though.
Yep, not a good idea IMHO...
Cheers,
Seb
>
> >I don't want to see extra code generated for every library
> Well... one code takes less code space and might be easier to implement. It
> does however provide less detail, especially for people using svn as a
> source and in cases with multiple libraru versions installed.
>
> > that seems wasteful of code space.
> I think it is *use* of code space rather then waste, since you get
> something in return. And to put this into perspective: we have less then 80
> libraries, excluding the device files. Most projects will probably use
> between 5 and 20 libraries, so the signatures will take less then 0,5% of
> the code space. And if you need these last few bytes, you could disable it.
>
> Of course, it would be better if we found a way to get the message in the
> asm file as comment without actual code.
>
> Joep
>
>
> >
>
--
Sébastien Lelong
http://www.sirloon.net
http://sirbot.org
--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups
"jallib" group.
To post to this group, send email to [email protected]
To unsubscribe from this group, send email to
[email protected]
For more options, visit this group at
http://groups.google.com/group/jallib?hl=en
-~----------~----~----~----~------~----~------~--~---
