| Message ID | 1456160567-18072-1-git-send-email-berrange@redhat.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
On 02/22/2016 10:02 AM, Daniel P. Berrange wrote: > Pretty printing of JSON responses is important to be able > to understand large responses from query commands in > particular, eg > > (QEMU) query-chardev > { u'return': [ { u'filename': u'vc', > u'frontend-open': False, > u'label': u'parallel0'}, > { u'filename': u'vc', > u'frontend-open': True, > u'label': u'serial0'}, > { u'filename': u'unix:/tmp/qemp,server', > u'frontend-open': True, > u'label': u'compat_monitor0'}]} > > Unfortunately this was broken during the addition of > the verbose flag in > > commit 1ceca07e48ead0dd2e41576c81d40e6a91cafefd > Author: John Snow <jsnow@redhat.com> > Date: Wed Apr 29 15:14:04 2015 -0400 > > scripts: qmp-shell: Add verbose flag > > This is because that change turned the python data > structure into a formatted JSON string before the > pretty print was given it. So we're just pretty > printing a string, which is a no-op. > > This fixes pretty printing of the command responses > and as an added benefit it will also pretty print > command arguments in verbose mode > > (QEMU) object-add qom-type=secret id=sec0 props={"data":"123456"} > { 'arguments': { 'id': 'sec0', > 'props': { u'data': u'123456'}, > 'qom-type': 'secret'}, > 'execute': 'object-add'} > { u'return': { }} That's an odd mix of regular and Unicode strings. Any ideas why it is happening? But the patch is a strict improvement, so: Reviewed-by: Eric Blake <eblake@redhat.com> > > Signed-off-by: Daniel P. Berrange <berrange@redhat.com> > --- > scripts/qmp/qmp-shell | 4 ++-- > 1 file changed, 2 insertions(+), 2 deletions(-) > > diff --git a/scripts/qmp/qmp-shell b/scripts/qmp/qmp-shell > index 7a402ed..15bbc5e 100755 > --- a/scripts/qmp/qmp-shell > +++ b/scripts/qmp/qmp-shell > @@ -231,10 +231,10 @@ class QMPShell(qmp.QEMUMonitorProtocol): > return qmpcmd > > def _print(self, qmp): > - jsobj = json.dumps(qmp) > if self._pp is not None: > - self._pp.pprint(jsobj) > + self._pp.pprint(qmp) > else: > + jsobj = json.dumps(qmp) > print str(jsobj) > > def _execute_cmd(self, cmdline): >
On Mon, Feb 22, 2016 at 03:02:03PM -0700, Eric Blake wrote: > On 02/22/2016 10:02 AM, Daniel P. Berrange wrote: > > Pretty printing of JSON responses is important to be able > > to understand large responses from query commands in > > particular, eg > > > > (QEMU) query-chardev > > { u'return': [ { u'filename': u'vc', > > u'frontend-open': False, > > u'label': u'parallel0'}, > > { u'filename': u'vc', > > u'frontend-open': True, > > u'label': u'serial0'}, > > { u'filename': u'unix:/tmp/qemp,server', > > u'frontend-open': True, > > u'label': u'compat_monitor0'}]} > > > > Unfortunately this was broken during the addition of > > the verbose flag in > > > > commit 1ceca07e48ead0dd2e41576c81d40e6a91cafefd > > Author: John Snow <jsnow@redhat.com> > > Date: Wed Apr 29 15:14:04 2015 -0400 > > > > scripts: qmp-shell: Add verbose flag > > > > This is because that change turned the python data > > structure into a formatted JSON string before the > > pretty print was given it. So we're just pretty > > printing a string, which is a no-op. > > > > This fixes pretty printing of the command responses > > and as an added benefit it will also pretty print > > command arguments in verbose mode > > > > (QEMU) object-add qom-type=secret id=sec0 props={"data":"123456"} > > { 'arguments': { 'id': 'sec0', > > 'props': { u'data': u'123456'}, > > 'qom-type': 'secret'}, > > 'execute': 'object-add'} > > { u'return': { }} > > That's an odd mix of regular and Unicode strings. Any ideas why it is > happening? But the patch is a strict improvement, so: It is todo with the way qmp-shell parses the command line. The qom-type and id arguments it just parses directly, but the last props argument it gets by using json.loads() which returns unicode strings. I realized we can avoid this issue and improve output style by not pretty printing a python data struture, and instead telling the json formatter to pretty print its output. So I've sent a v2 that does that Regards, Daniel
diff --git a/scripts/qmp/qmp-shell b/scripts/qmp/qmp-shell index 7a402ed..15bbc5e 100755 --- a/scripts/qmp/qmp-shell +++ b/scripts/qmp/qmp-shell @@ -231,10 +231,10 @@ class QMPShell(qmp.QEMUMonitorProtocol): return qmpcmd def _print(self, qmp): - jsobj = json.dumps(qmp) if self._pp is not None: - self._pp.pprint(jsobj) + self._pp.pprint(qmp) else: + jsobj = json.dumps(qmp) print str(jsobj) def _execute_cmd(self, cmdline):
Pretty printing of JSON responses is important to be able to understand large responses from query commands in particular, eg (QEMU) query-chardev { u'return': [ { u'filename': u'vc', u'frontend-open': False, u'label': u'parallel0'}, { u'filename': u'vc', u'frontend-open': True, u'label': u'serial0'}, { u'filename': u'unix:/tmp/qemp,server', u'frontend-open': True, u'label': u'compat_monitor0'}]} Unfortunately this was broken during the addition of the verbose flag in commit 1ceca07e48ead0dd2e41576c81d40e6a91cafefd Author: John Snow <jsnow@redhat.com> Date: Wed Apr 29 15:14:04 2015 -0400 scripts: qmp-shell: Add verbose flag This is because that change turned the python data structure into a formatted JSON string before the pretty print was given it. So we're just pretty printing a string, which is a no-op. This fixes pretty printing of the command responses and as an added benefit it will also pretty print command arguments in verbose mode (QEMU) object-add qom-type=secret id=sec0 props={"data":"123456"} { 'arguments': { 'id': 'sec0', 'props': { u'data': u'123456'}, 'qom-type': 'secret'}, 'execute': 'object-add'} { u'return': { }} Signed-off-by: Daniel P. Berrange <berrange@redhat.com> --- scripts/qmp/qmp-shell | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)