[28587] in Source-Commits
Re: build-system commit: Output error message when dabuild fails
daemon@ATHENA.MIT.EDU (Alex Chernyakhovsky)
Fri Mar 11 21:30:23 2016
MIME-Version: 1.0
In-Reply-To: <23E5861B-9E4C-4EEF-B007-C42C104B48E4@mit.edu>
Date: Fri, 11 Mar 2016 21:30:18 -0500
Message-ID: <CAB18ysoqHhQLBh5P0Hw96Hu2djtgsu1xjDMbwZTu5-o4wxV13g@mail.gmail.com>
From: Alex Chernyakhovsky <achernya@mit.edu>
To: Jonathan D Reed <jdreed@mit.edu>
Cc: Victor Vasiliev <vasilvv@mit.edu>, source-commits <source-commits@mit.edu>
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
You never want a bare raise. Python's implicit raise is not thread safe, IIRC.
-Alex
On Fri, Mar 11, 2016 at 9:24 PM, Jonathan D Reed <jdreed@mit.edu> wrote:
> Doesn’t that want to be a bare “raise” instead of “raise err” so that the stack trace is preserved? Not that that last frame will be particularly useful, but I think it’s good form.
>
> -Jon
>
> On Mar 11, 2016, at 5:37 PM, Victor Vasiliev <vasilvv@MIT.EDU> wrote:
>
>> https://github.com/mit-athena/build-system/commit/e54c602fd0ee1d4c092c876abcb8d81b31988993
>> commit e54c602fd0ee1d4c092c876abcb8d81b31988993
>> Author: Victor Vasiliev <vasilvv@mit.edu>
>> Date: Fri Mar 11 17:26:38 2016 -0500
>>
>> Output error message when dabuild fails
>>
>> dabuild | 6 +++++-
>> 1 files changed, 5 insertions(+), 1 deletions(-)
>>
>> diff --git a/dabuild b/dabuild
>> index c72bfb6..35a46d9 100755
>> --- a/dabuild
>> +++ b/dabuild
>> @@ -312,7 +312,11 @@ if __name__ == '__main__':
>> print >>sys.stderr, "The lock is in place; unable to proceed"
>> sys.exit(1)
>> try:
>> - main()
>> + try:
>> + main()
>> + except subprocess.CalledProcessError as err:
>> + print err.output
>> + raise err
>> finally:
>> dabuildsys.release_lock()
>>
>
