[Gc] There should be a library major bump of gc library due to now having several GC_xyz "hidden" functions

Paulo César Pereira de Andrade paulo.cesar.pereira.de.andrade at gmail.com
Sun May 27 11:44:18 PDT 2012


2012/5/27 Ivan Maidanski <ivmai at mail.ru>:
> Hi Juan and Paulo,
>
> Sun, 27 May 2012 15:18:04 +0200 от Juan Jose Garcia-Ripoll <juanjose.garciaripoll at googlemail.com>:
>> On Sun, May 27, 2012 at 1:43 PM, Ivan Maidanski <ivmai at mail.ru> wrote:
>>
>> > > Probably this should be enough, but I fear that none of these changes
>> > will propagate to the problematic distributions unless the library suffers
>> > a formal release.
>> >
>> > What's normal? gc-7.2b is such or not?
>> >
>>
>> Sorry, but I have missed any announcement that 7.2 had moved beyond alpha
>> stage.
>>
>>
>> > From our side, we can prepare gc-7.2c with the required fixes unhiding
>> > symbols you mentioned (could somebody prepare a patch?) and also move them
>> > to public API in gc-7.3alpha.
>> >
>>
>> I am right now a bit busy to produce the patch, but will push it in my
>> short-term to-do list. Please point me to the right repository to make the
>> patch: is this your git one? Does it has a 7.2 to patch again?
>
> Yes, the git repo: https://github.com/ivmai/bdwgc

  I forked it at https://github.com/pcpa/bdwgc

> Ideally, it would be good to prepare 3 patch sets:
> 1. applicable both to master (7.3alpha) and release (7.2) branches (e.g., regarding unhiding GC_push_x)
> 2. applicable to release branch only (e.g. regarding unhiding GC_mark_bits as it is already public in v7.3)
> 3. applicable to master branch only after patch set 1 (e.g. move concerned GC_push_x to public API).

  What I did was to cherry-pick the mark-bit commit in the branch release,
then write another for the GC_push functions and symbol required by
ecl. I then rewrote the patch in master because it would not apply
cleanly (due to modifiers, e.g. GC_ATTR_NONNULL)

> Paulo -
> Would be good if you could help Juan in patch preparation. Thank you.

  I gone ahead and did the patches as listed above, and confirmed it compiles
and works at least for a "(+ 1 1)" test in ecl, but did not address the comment
Juan said about GC_init_explicit_typing as I do not have enough insight about
what needs to be done.

  Ivan, I think I pressed "pull request" in the github interface, but
should still
need more review about the patch, as more functions may need to be made
GC_API.

> Regards,
> Ivan
>
>>
>> Juanjo
>>
>> --
>> Instituto de Física Fundamental, CSIC
>> c/ Serrano, 113b, Madrid 28006 (Spain)
>> http://juanjose.garciaripoll.googlepages.com

Paulo



More information about the Gc mailing list