17 Nov
2007
17 Nov
'07
11:10 a.m.
> c) Could you please put SOME debugging back. It was there for a > reason, and removing it outright is bad idea. > > Thanks, > Aaron 1. To summarize my opinion about pacman debugging, please give a look at smoke001.py output with --debug... 2. the old debugging was _fortunately_ broken (or at least misleading) in alpm_depcmp: it printed debug message (match / no match) iff pkgname/provision matched 3. alpm_depcmp is quite an atomic function, and can _prove_ easily if it works correctly or not (for example alpm_list_next doesn't print debug information neither) <- the only information in its debug message: its input parameters (but see also 4.) 4. if you "fix" 2., then even a single-package transcation will lead to _thousands_ of depcmp lines in debug (search for satisfier in local and sync db-s); back to smoke001.py: the fix would add at least 1 million (!) depcmp lines (alpm_graph_init) Do you still want it? If yes, with 2-fix or without it? Bye ---------------------------------------------------- SZTE Egyetemi Könyvtár - http://www.bibl.u-szeged.hu This mail sent through IMP: http://horde.org/imp/