From mboxrd@z Thu Jan 1 00:00:00 1970 From: Jim Wilson To: David Edelsohn Cc: gcc@gcc.gnu.org Subject: Re: Loop unroll fixes Date: Fri, 14 Sep 2001 22:23:00 -0000 Message-id: <200109150523.WAA23869@cygnus.com> References: <200109150311.XAA24844@makai.watson.ibm.com> X-SW-Source: 2001-09/msg00589.html > Please keep in mind that your messages state your opinion of the >events, not some objective facts. I have made many statements of fact, and I have tried to provide documentation for all of them. You have not provided any verifiable documentation to contest any of them. There are two facts of particular notice 1) You have made accusations of malfeasance against other gcc developers, but you have not provided any verifiable evidence to back these claims. 2) You stated that you would deliberately violate gcc process, but you have not provided any verifiable evidence to prove that this was necessary. I have also provided opinions, you are welcome to disagree with them, but that does not change the facts. > I only can express my view of the events as well. I have not seen you provide any verifiable documentation for any of your claims. > I respect your opinion, although >I think you developed your conclusions before investigating all of the >information. This is wrong. I researched every shred of evidence that I could find on gcc.gnu.org. I looked at the PR in question, 3384. I searched every mailing list for the PR number, for Zoltan Hidvegi's name, and I followed every thread I could find. The only information I did not investigate was personal interviews of the involved parties, but I did not have time for that, because you had specified too short of a deadline to allow for that. I included all of this evidence in my original message on this thread. The wealth of references in my message make it clear that I invested a considerable amount of time investigating the issues. For reference, my message is here http://gcc.gnu.org/ml/gcc/2001-09/msg00533.html I see no basis for your opinion that my investigation was insufficient. If you can provide verifiable evidence that I am wrong, then please do so. > The intent of my original message was to notify everyone that I >was not going to allow the bottleneck on this patch to continue. I did >not intend for anyone to immediately review the patch or generate >alternate fixes for bugs highlighted by the patch. I contest this claim. Your original message is here: http://gcc.gnu.org/ml/gcc/2001-09/msg00529.html In this message, you say: If there is no specific, documented, technical objection to this patch within one week, it will be approved. You could have achieved your goal by reminding everyone that the patch was still unreviewed, and that you wanted the patch reviewed in time to be included in gcc 3.0.2. Instead, you issued an ultimatum. You specified a deadline, and you stated that you would take unilateral action in contravention of documented gcc policies if there was no objection. This was designed to provoke an immediate review of the patch, as that is the only way to generate a technical objection to the patch. As supporting evidence, I offer a quote from this message: http://gcc.gnu.org/ml/gcc/2001-09/msg00549.html I quote: Your vindictive response of further delaying the patch does not help GCC or anyone. If you were not interested in an immediate review of the patch, then the fact that I said I would look at it in one week's time should not have been a serious problem. I will admit this one is weak, one could argue that you were contesting the style of the response, and not the delay, but still, I think it has merit as a supporting claim. After all, I did agree to review the patch, I just did not agree to review it immediately. > None of us has any sort of special >privilege or insight to decree what is right or what is wrong or what is >sufficient evidence or what policy is appropriate. And this is exactly why I objected to your original message. Because with your ultimatum, you were claiming special priviledges for yourself to set policy. You have stated this much better than I did. And with that, I'd like to end this thread. If David wants a rebuttal, I won't object. I would suggest that anyone who wants to help solve the problem of patch reviewing contribute to Joe Buck's thread instead of this one. I will followup on Zoltan's patch as I promised. I am at the moment waiting to see what comes from Bernd's review. As a final word, I will admit that I am guilty of unnecessary flamage, but I thought the issues raised here were worth the risk. Jim