I am announcing the End-Of-Life of the http://opensolaris.org/cgi/arc.py
service. This service is used by both the Cadmium extension to
Mercurial and the ON gate infrastructure to obtain ARC case status
information. This EOL is required as part of the migration to the new
site infrastructure.
The replacement for the current service is available from two locations,
one on opensolaris.org, and one on Sun's internal network:
http://arc.opensolaris.org/cgi-bin/arc.cgi
http://candi.sfbay.sun.com/cgi-bin/arc.cgi
The new arc.cgi service has a different interface to the existing arc.py
service, and produces different output. The changes were made to make
it easier to invoke the service, and easier to parse the resulting
output. Specifically, the output is now well-formed CSV, where all
fields are double-quoted (") and any double-quotes are escaped, so '"'
within a field becomes '""'. In particular, this format is then the same
as that accepted as default by the Python csv class
(http://docs.python.org/library/csv.html), so there is no need to write
a custom parser for the output, unlike the current arc.py script.
The easiest way of showing how to use the new interface is to give an
example:
http://arc.opensolaris.org/cgi-bin/arc.cgi?case=PSARC/2009/003&case=PSARC/3000/01&case=oops
"error","oops"
"exists","PSARC","2009","003","FastTrack","closed approved fast-track
01/12/2009","conflict"
"absent","PSARC","3000","01"
In addition, if the script has an internal error, (e.g. can't open a
data file) it will output an error of the following form:
"fatal","descriptive error message"
The data source used by the new script is also new, specifically the
data is provided by the same mechanism that is now used to publish ARC
cases to http://arc.opensolaris.org/caselog. The raw data file can be
found at the following locations (internal/external):
http://arc.opensolaris.org/caselog/case_status.csv
/net/candi.sfbay.sun.com/opt/oso/arc/caselog/case_status.csv
This file has some differences to the old data feed:
1. It only includes data for the ARCs that are published to
arc.opensolaris.org, namely FWARC, LSARC, PSARC and WSARC.
2. The full case name is available, rather than the 40 character
truncated version provided by the existing mechanism.
3. The case type is not available - this is not used by the current
arc.py script in any case.
4. It is a well-formed CSV file (see above).
As an aid to migration, a backward-compatible version of the existing
arc.py script is available from the following locations:
http://arc.opensolaris.org/cgi-bin/arc.py
http://candi.sfbay.sun.com/cgi-bin/arc.py
This takes the same arguments and output the same format as the existing
arc.py script, but uses the new data file. As a consequence, the output
will differ from the current script as described in points 1 and 2 above.
The current http://opensolaris.org/cgi/arc.py script will be removed on
Monday 11th May, and the backwards-compatible versions on arc.oso and
candi.sfbay will be removed on Monday 8th June.
--
Alan Burlison
--
_______________________________________________
tools-discuss mailing list
tools-discuss@opensolaris.org
