Joe Orton wrote:
So again, I guess I presumed back then, and would still now, that these problems were understood and were going to be addressed. But that didn't actually happen; the problems are still there. Again, that is the basis of my veto.
Yep, it is still there, and that is because I have been devoting about 10 hours a day every day weekends included (this last weekend excluded thank goodness) to the httpd and apr code over the last three weeks, and that was still not enough to get everything done.
In that time the session stuff got a major refactoring - again. The documentation was written and rewritten, specifically from the point of view of somebody who had never seen this stuff before, and this had the knock on effect of even more rewritten and refactored code.
Then I started slowly and tentatively committing the code, and rpluem did a thorough evaluation, which had to be analysed, acted on and committed. Then Roy presented some issues, so that pushed out the work for a few days more.
At the same time, two external deadlines were looming, and they could not be pushed out any longer, so I have had to devote my attention back to those. The apr and httpd job is not done, and I have no intention of leaving the job undone until it is truly complete. I would have seen no point in even trying to get this stuff introduced if the intention was to not do it properly.
I made the conscious decision to do the httpd stuff before the apr stuff because the apr stuff builds clean, and the httpd stuff does not. Httpd takes priority right now for that reason. I have every intention of addressing the apr concerns. I absolutely cannot do so this week.
I am sorry if I failed to do enough to help explain or discuss the problems last year, and as a result of that the veto seems unduly harsh.
As I said, they are valid concerns, and I have every intention of addressing every one. All I ask is some patience: the apr stuff needs a lot of work, not the least of which is the Microsoft Crypto implementation. There would be little point in going to all this effort if the code only ever worked on one platform.
Regards, Graham --
smime.p7s
Description: S/MIME Cryptographic Signature
