Comments on: Call for critical A98 fixes
As you know, the main focus of R3 A98 is to add hosted (resident) extensions. This makes it possible to include extended modules (with your own native functions) bundled within a single executable file, thereby eliminating the need for external libraries to accompany the distribution and installation of the program.
A98 will be released in a few days, but there may be time to make a few other critical fixes, as long as it does not delay the release. Hopefully, after A98, we can return to more frequent updates to address various other bugs and wishes.
Once A98 is tested, the graphics library will be back ported into the hosted extension mechanism (along with a few changes to how DELECT processes dialects into command calls.)
congrats Carl, this arrives at a very good time.
May I suggest CureCode tickets #609, #351, #746 and #1148. Also, the problems related to "unwind" handling (#1535 and all that precede).|
#1148 would help, but there are other protect bugs that need fixing as well, so we might as well do them all at once: #1014, #1015, #1141 and #1141 (which affect #1143). Without all of these fixed we can't work on the security model. Don't know if that is a priority for a98, but perhaps another release soon.
Unwind handling bugs: The core one is #1509, and maybe #1515 for unwinds if that isn't the same bug, and #1519 for good measure. The rest are all side effects of #1509, though those side effects often cause crashes or worse. If these fixes are really difficult they can be put off until a99, but they should be done soon.
Please check all pending bugs in CureCode; all of them refer to changes that have already been submitted, and thus should be included (or dismissed).
Good catch on #609, #351 and #746. We need to do a release that focuses on all of the significant semantic or syntactic changes that need doing, and we need to do it sooner rather than later, before too much code is written. #609 is particularly critical to do as soon as we can.
All bugs introduced or incompletely fixed in a97 should be fixed in a98, including #1476, #1477, #1478 and #1486.
Bug #1452 is blocking a lot of people who are writing network protocols and other low-level code that works with binaries. If it's an easy fix, it would be a good idea to do sooner.|
Agree with Brian. #1452 is a bug I've hit in first couple of minutes when trying to convert some R2 code so I guess other people will do as well. It will be good to fix in the next release.
Ok, good notes. I'll take a look at those items listed, but I'm not going to spend much time on them today, because we can make A99 release as soon as necessary with additional fixes.
Also, I'm interested in any issues related to the host-kit builds, some of which were made earlier in discussion groups. We need to collect these, although we can deal with them over the next few releases (as we move graphics commands to the host.)
Passing by...I see discussion on Bug #1452. By the way, writing a network protocol is very different in R3 compared to R2 ?
Is there already a kind of tutorial to do it in R3 ?
On your mention of backporting the GUI stuff into the release: Will this be ready for A99? If not, I would appreciate a rough estimate.
Time to resume our A98 release goal.
Ok, so we had a huge discussion about self as used in objects and functions, and a number of revisions have been made to contexts, objects, and binding. This should help us close out a number of those related CureCode tickets.
So, it's time to get back to the rest of A98 and make at least a few more critical fixes before release.
Also, A98 does include the embedded extensions mechanism; however, there are still a few rough edges we will need to iron out.
Here's the plan: I'm going to look over the tickets noted above, and fix the ones that can be done quickly. Any deep issues will need to wait for the next go-round, because we don't want further delays in release.
Post a Comment:
You can post a comment here. Keep it on-topic.