Why not just make mpich2-maint same as mpich-discuss ? A.Chan ----- "Darius Buntinas" <[email protected]> wrote:
The idea is that people send their issues (bug, help requests, etc) to mpich2-discuss. Then if we decide that it's a real bug, we create a ticket for it. Users who are sufficiently sophisticated to determine that it's a bug and know how to use trac, can create tickets themselves, but they need not.
Sending email to mpich2-maint no longer creates a trac ticket.
-d
On 11/12/2009 11:14 AM, [email protected] wrote:
I am confused as why some mpich2-maint bug report goes through ? e.g. this guy from france about time command on mpich2 vs intel MPI. I see his mpich2-maint email but can't find his bug report in TRAC. Since his bug report comes through, should we answer his question or ask him to file a bug report ? (Answering his question right away seems to encourage him to do it again.)
Also, it seems some users are refusing to file a bug report on TRAC: e.g. the guy tried to do cross-compile nemesis on an alpha chip. I guess I will file a bug report for him....
Our bug-report system seems confusing to some users (like old MPI users who ain't really up to all the technology or non CS newbies), e.g. our reply message from mpich2-maint assumes user know the problem he/she has is bug or not. Also the reply message does not explicitly contain the link.
http://wiki.mcs.anl.gov/mpich2/index.php/Creating_New_Bug_Tickets
Since filing a bug report with TRAC isn't easy, I can see some user will just gives up.... Are we stuck with this system ?
A.Chan
That is an option, but some people may expect mpich2-maint to a small set of people (us) rather than 400. Rajeev
-----Original Message----- From: [email protected] [mailto:[email protected]] On Behalf Of [email protected] Sent: Thursday, November 12, 2009 1:27 PM To: Darius Buntinas Cc: [email protected] Subject: Re: [mpich2-core] bug reporting
Why not just make mpich2-maint same as mpich-discuss ?
A.Chan
----- "Darius Buntinas" <[email protected]> wrote:
The idea is that people send their issues (bug, help requests, etc) to mpich2-discuss. Then if we decide that it's a real bug, we create a ticket for it. Users who are sufficiently sophisticated to determine that it's a bug and know how to use trac, can create tickets themselves, but they need not.
Sending email to mpich2-maint no longer creates a trac ticket.
-d
On 11/12/2009 11:14 AM, [email protected] wrote:
I am confused as why some mpich2-maint bug report goes through ? e.g. this guy from france about time command on mpich2 vs intel MPI. I see his mpich2-maint email but can't find his bug report in TRAC. Since his bug report comes through, should we answer his question or ask him to file a bug report ? (Answering his question right away seems to encourage him to do it again.)
Also, it seems some users are refusing to file a bug report on TRAC: e.g. the guy tried to do cross-compile nemesis on an alpha chip. I guess I will file a bug report for him....
Our bug-report system seems confusing to some users (like old MPI users who ain't really up to all the technology or non CS newbies), e.g. our reply message from mpich2-maint assumes user know the problem he/she has is bug or not. Also the reply message does not explicitly contain the link.
http://wiki.mcs.anl.gov/mpich2/index.php/Creating_New_Bug_Tickets
Since filing a bug report with TRAC isn't easy, I can see some user will just gives up.... Are we stuck with this system ?
A.Chan
participants (2)
-
chan＠mcs.anl.gov -
Rajeev Thakur