Hi Mitchell, Am 31/05/2014 12:39, schrieb Mitchell Joblin:
- Avoid recreating the database file if it is already present this part of the analysis can be very time consuming
but that effectively invalidates the --create_db parameter, doesn't it? Even when --create_db is specificed, the db file will be re-used, making it harder to generate an analysis from scratch. In particular, this can lead to very hard to track problems: You've specified --create_db and are under the (valid) impression that everything will be set up from scratch, while some probably stale old data will be re-used. Since I've suffered a lot from caching old runs in the mailing list analysis, I like to preserve the opportunity to really start off in a clean state. Or are there any use cases for the change that I'm missing? If you don't specify --create_db and run an analysis multiple times, you will anyway get the effect caused by the patch. Cheers, Wolfgang
Signed-off-by: Mitchell Joblin <mitchell.joblin.ext@xxxxxxxxxxx> --- codeface/cluster/cluster.py | 2 ++ 1 file changed, 2 insertions(+) diff --git a/codeface/cluster/cluster.py b/codeface/cluster/cluster.py index d52d92e..99579dd 100755 --- a/codeface/cluster/cluster.py +++ b/codeface/cluster/cluster.py @@ -1303,6 +1303,8 @@ def doProjectAnalysis(conf, from_rev, to_rev, rc_start, outdir, #Perform appropriate analysis #---------------------------- filename = os.path.join(outdir, "vcs_analysis.db") + ## Check if db file is already present + create_db = not(os.path.isfile(filename)) dbm = DBManager(conf) performAnalysis(conf, dbm, filename, git_repo, [from_rev, to_rev], # kerninfo.subsysDescrLinux,