--- Em seg, 17/12/12, Bruce Dubbs escreveu: > De: Bruce Dubbs > Assunto: Re: [lfs-dev] Host System Requirements > Para: "LFS Developers Mailinglist" > Data: Segunda-feira, 17 de Dezembro de 2012, 13:13 > Chris Staub wrote: > > On 12/17/2012 06:38 AM, Fernando de Oliveira wrote: > >> > >> I see two modifications, one easy to do, the other > is perhaps > >> impossible. > >> > >> 1. Change the position of the gcc, so it is not in > the beginning nor the > >> end of the tests. > >> 2. Have a conclusion statement: "x tests passed, y > tests failed, if > >> y > 0, please fix your host system to fulfill > the requirements". > >> > >> Last time I discussed, I had some feelings, but not > the clarity I have > >> now, after I having succeeded to help some, and > understand and follow > >> most support to others' problems. > >> > >> If I have succeeded to explain myself this time, > but my suggestions seem > >> wrong, I do not mind, main point is that I still > feel that something > >> could be improved and many potential users not > being scared out anymore. > > > Another possibility: leave it mostly unchanged, except > for the line that > > actually runs the script, redirect stderr to a file (in > other words, > > "bash version-check.sh 2>errors.log"), then follow > up the script with > > something like "These are the packages that had issues: > [display > > errors.log]" At the same time, the "gcc compilation > failed" message can > > also have a "1>&2" appended to it to make sure > it also goes in the error > > log. Further, when I added a compile check to CLFS I > made the "fail" > > message somewhat more verbose, indicating that you > might want to go over > > the full package list again and check for any missing > development packages.
I like these ideas, and will change my script accordingly. Thanks, Chris. > There are some conflicting objectives here. One is > that we want to keep > it simple. Another is that we would like to highlight > problems. It has been proved by posts in the support that the latter is the choice, if there is no way of keeping the former. Or a second script would be created just to analyze the result of the simpler one. > By far, the biggest problem is having the wrong symlink for > /bin/sh. I > recently highlighted the symlink issue in Section 5.3 of > SVN. I suggest > we stay with that until after the next LFS release (March) > and then > re-evaluate. This was a good providence. Please, should I open a ticket just for us all to remember it on February/March? []s, Fernando -- http://linuxfromscratch.org/mailman/listinfo/lfs-dev FAQ: http://www.linuxfromscratch.org/faq/ Unsubscribe: See the above information page
