I've been noticing this off and on. It is an intermittent problem. I
think I've only seen it on Windows, and not on Ubuntu, but I'm not
absolutely sure of that.
I'm starting to write test cases for fossil json, and this is happening
when my new json.test script uses repo_init from tester.tcl to get a
nice clean and empty repo to abuse.
Ross@Riven ~/Documents/tmp/ftest
$ tclsh ../fossil4/test/tester.tcl ./fossil json -prot -quiet
project-id: 35e5185813329d3ba328171ececec1f2742232ab
server-id: 23942ec8523e03be10007f61dd901d0afa19bc72
admin-user: Ross (initial password is "7f0599")
SQLITE_NOTICE: delayed 25ms for lock/sharing conflict at line 39938
while executing
"exec $::fossilexe new $filename"
(procedure "repo_init" line 17)
<<and a bunch more lines of backtrace omitted here>>
The context in repo_init looks like:
catch {exec $::fossilexe close -f}
file delete $filename
exec $::fossilexe new $filename
exec $::fossilexe open $filename
I thought it might be a quirk of reusing the same file that was just
deleted, so I tried adding an after 100 command between the file delete
and exec fossil new. That had no effect at all.
I'm currently running tests happily on Windows from an MSYS bash prompt.
Incidentally, the newly introduced support for using ./configure works
great under MSYS bash, even for out of tree builds. In the command
above, the ftest folder is an out of tree build, and the fossil checkout
is in ../fossil4.
This is mostly just an annoyance as usually just running the test
framework again works ok. But intermittent problems bug me, and
intermittent problems in a test framework bug me more...
Any thoughts?
--
Ross Berteig r...@cheshireeng.com
Cheshire Engineering Corp. http://www.CheshireEng.com/
_______________________________________________
fossil-dev mailing list
fossil-dev@mailinglists.sqlite.org
http://mailinglists.sqlite.org/cgi-bin/mailman/listinfo/fossil-dev
