|
From: | Richard Frith-Macdonald |
Subject: | Re: problems building - faiiling due to autogsdoc |
Date: | Fri, 20 Jun 2003 07:56:29 +0100 |
On Friday, June 20, 2003, at 03:35 am, Adam Fedor wrote:
On Thursday, June 19, 2003, at 05:31 PM, Alex Perez wrote:On Thu, 19 Jun 2003 22:55:12 +0100 Richard Frith-Macdonald <richard@brainstorm.co.uk> wrote:Please don't do that ... instead, track down the cause of the problemand supply a fix, so other people don't have the problem in future.However, if --disable-procfs gets rid of the symptom. it's probably agood idea to look at what information the operating system is providing to the procfs code in NSProcessInfo.1. there could be a bug in GNUsteps handling of procfs, which could befixed. 2. there could be a bug in BSDs procfs, for which we could get autoconf to enable a workaround (or disable procfs support if it is horribly broken).IIRC, I reported that /procfs problem a good 1 year a go (could be even more). GNUstep tries to access a couple of entries which simply do notexist, like /procfs/self/exe (or sth like that) on BSD.*cough* This kind of crap is why GNUstep doesn't get the attention it deserves. *puts on fireproof suit*Well, you, of course, asked for it. So I went back and searched all the gnustep mailing list archives and I didn't find a mention of a bug report like this or anyone suggesting that this was the solution. I also searched the entire web and didn't find anything like this. Perhaps it was done privately - I don't know about that.
I have no recollection of this bug report either, wheras I *do* remember being mildly annoyed that my requests/suggestions to run under gdb to find out what was going wrong never got a response, so I was never able to do anything about the problem (I wasn't willing to go so
far as buying a new machine and installing BSD to track down the bug).So ... how can we decide when to automatically disable procfs? How can configure decide
which BSD variants have which bits of procfs missing?Also, can Pete confirm that the problem is really with procfs, and not simply that +load is
not being called.
[Prev in Thread] | Current Thread | [Next in Thread] |