Messages in this thread | | | Subject | Re: ext2 inode o-o-bound prob aft crash xemacs aft susp during gnus mode | Date | Wed, 19 Jun 1996 18:12:30 -0400 | From | ulmo@q ... |
| |
whoops! hate being sick ... forgot to tell versions of stuff:
2.0
(assorted accompanying package versions according to Changes file)
xemacs 19.13 compiled on my system May 7th, probably under the current libc, 5.3.12.
Was running with gpm in console mode (latest gnus doesn't seem to work with GNU emacs very well).
I'm a bit surprised we went from a ``stable'' libc 5.2.*, to ``experimental'' 5.3.*, to ``experimental'' 5.4.* when really we should be stabalizing everything right about now. Is the 5.4.* series going to end up being the one we should all use with the latest Linux 2.0 kernel?
We really do need better version dependency trees, I agree with that request very largely. This is everyone's responsibility: you must put it in the release notes of your package a whole list of "this requires at least x and at most y of package z" for every package dependency, and another "package c version d requires at least version a and at most version b of this package, but version e of package c requires at least version f of this package, which also requires version w of package n to be compiled before installing package o but deleted after installing package p, according to the instructions within the release notes to package m version g" or whatever is the truth for every package that depends on your package. Making a machine-readable file in some standard format that some parsing program that admin's can use would be superb.
Because, quite frankly, I want the higher security of 5.3.12, but also want to follow the stabilization track that will eventually go into setups centered around 2.0 .... (to think, most of you will know what I mean by those version #s without even mentioning what package I'm talking about. tsk.)
Alright, so these last three paragraphs are just about distribution issues; perhaps I'll wait for a solid Linux 2.0 distribution. But, doing everything Documentation/Changes says is so reliable, that I thought that that was the way it was supposed to be done ... :) and get annoyed by the occasional bugs that do seem to squiggle their way in anyway.
| |