Re: [Zope-Checkins] SVN: Zope/trunk/lib/python/Shared/DC/ZRDB/ Collector #556: sqlvar now returns 'null' rather than 'None'.
This change breaks Gadfly which doesn't seem to like 'null' at all. I poked around in ZGadflyDA/gadfly a bit, but it's not obvious to me how to fix the parser (*.mar files anyone?). Stefan [snip] File "/usr/local/Zope-2_8-branch/lib/python/Products/ZGadflyDA/ gadfly/kjParser.py", line 826, in ParseError raise SyntaxError, 'unexpected token sequence.' + data SyntaxError: unexpected token sequence.near :: "dca437824c3b0',"*" null, '', '" ******************************* current state = 252 expects: '?', 'numeric_literal', 'character_string_literal', '-', '+', ('nomatch1',) current token = ((-8, 'user_defined_name'), 'NULL') On 3. Sep 2005, at 01:20, Tres Seaver wrote:
Log message for revision 38276: Collector #556: sqlvar now returns 'null' rather than 'None'.
Changed: U Zope/trunk/lib/python/Shared/DC/ZRDB/sqlvar.py A Zope/trunk/lib/python/Shared/DC/ZRDB/tests/test_sqlvar.py
-=- Modified: Zope/trunk/lib/python/Shared/DC/ZRDB/sqlvar.py =================================================================== --- Zope/trunk/lib/python/Shared/DC/ZRDB/sqlvar.py 2005-09-02 23:15:46 UTC (rev 38275) +++ Zope/trunk/lib/python/Shared/DC/ZRDB/sqlvar.py 2005-09-02 23:20:49 UTC (rev 38276) @@ -98,6 +98,9 @@ raise raise ValueError, 'Missing input variable, <em>%s</ em>' % name
+ if v is None: + return 'null' + if t=='int': try: if type(v) is StringType:
-- Anything that happens, happens. --Douglas Adams
--On 8. Oktober 2005 15:41:57 +0200 "Stefan H. Holek" <stefan@epy.co.at> wrote:
This change breaks Gadfly which doesn't seem to like 'null' at all. I poked around in ZGadflyDA/gadfly a bit, but it's not obvious to me how to fix the parser (*.mar files anyone?).
No idea about the specific error but I think we should now officially deprecate Gadfly for Zope 2.9 and get rid of it in Zope 2.11 (:-)). It really provides no functionality to the Zope core. We could rip it of as external product...but it should go away from the core. -aj
--On 8. Oktober 2005 16:04:46 +0200 Andreas Jung <lists@andreas-jung.com> wrote:
--On 8. Oktober 2005 15:41:57 +0200 "Stefan H. Holek" <stefan@epy.co.at> wrote:
This change breaks Gadfly which doesn't seem to like 'null' at all. I poked around in ZGadflyDA/gadfly a bit, but it's not obvious to me how to fix the parser (*.mar files anyone?).
No idea about the specific error but I think we should now officially deprecate Gadfly for Zope 2.9 and get rid of it in Zope 2.11 (:-)). It really provides no functionality to the Zope core. We could rip it of as external product...but it should go away from the core.
Argh...there is also a gadfly package coming from Zope 3. So when we remove it from the Zope 2 core we get it back with Zope 3 :-) -aj
Andreas Jung wrote:
Argh...there is also a gadfly package coming from Zope 3. So when we remove it from the Zope 2 core we get it back with Zope 3 :-)
Well, having a lightweight semi-functional rdb engine in the distro has always seemed handy for me, mainly for testing rdb-related components with the certainty that the rdb engine you're testing with will always be there... Maybe we should move it to the testing package? cheers, Chris -- Simplistix - Content Management, Zope & Python Consulting - http://www.simplistix.co.uk
On 10 Oct 2005, at 08:58, Chris Withers wrote:
Andreas Jung wrote:
Argh...there is also a gadfly package coming from Zope 3. So when we remove it from the Zope 2 core we get it back with Zope 3 :-)
Well, having a lightweight semi-functional rdb engine in the distro has always seemed handy for me, mainly for testing rdb-related components with the certainty that the rdb engine you're testing with will always be there...
Maybe we should move it to the testing package?
You're just moving the problem around with that. What's the point? jens
On Mon, Oct 10, 2005 at 08:58:53AM +0100, Chris Withers wrote: | Andreas Jung wrote: | >Argh...there is also a gadfly package coming from Zope 3. So when we | >remove it from the Zope 2 core we get it back with Zope 3 :-) | | Well, having a lightweight semi-functional rdb engine in the distro has | always seemed handy for me, mainly for testing rdb-related components | with the certainty that the rdb engine you're testing with will always | be there... | | Maybe we should move it to the testing package? If I'm allowed to say anything, I would like to see gadfly being replaced by sqlite. It's available mostly everywhere except in Windows by default, as many products already use it. sqlite itself is 'public domain' and the python da has the same license as zlib if i recall, which is very permissive as well. -- Sidnei da Silva Enfold Systems, LLC. http://enfoldsystems.com
SQLite implements enough of a subset of SQL to be genuinely useful and it's performance isn't too bad either. There is a Zope DA for it somewhere too. We've never been able to use Gadfly for testing as it's so far from standard SQL that it just causes more problems than it solves - as we saw at the beginning of this thread. A On 10 Oct 2005, at 15:03, Sidnei da Silva wrote:
On Mon, Oct 10, 2005 at 08:58:53AM +0100, Chris Withers wrote: | Andreas Jung wrote: | >Argh...there is also a gadfly package coming from Zope 3. So when we | >remove it from the Zope 2 core we get it back with Zope 3 :-) | | Well, having a lightweight semi-functional rdb engine in the distro has | always seemed handy for me, mainly for testing rdb-related components | with the certainty that the rdb engine you're testing with will always | be there... | | Maybe we should move it to the testing package?
If I'm allowed to say anything, I would like to see gadfly being replaced by sqlite. It's available mostly everywhere except in Windows by default, as many products already use it. sqlite itself is 'public domain' and the python da has the same license as zlib if i recall, which is very permissive as well.
-- Sidnei da Silva Enfold Systems, LLC. http://enfoldsystems.com
-- Logicalware Ltd Stuart House, Eskmills, Musselburgh, EH21 7PQ, UK Tel: +44(0)131 273 5130 http://www.logicalware.com -- This message has been scanned for viruses and dangerous content by MailScanner, and is believed to be clean.
--On 10. Oktober 2005 11:03:54 -0300 Sidnei da Silva <sidnei@enfoldsystems.com> wrote:
If I'm allowed to say anything, I would like to see gadfly being replaced by sqlite. It's available mostly everywhere except in Windows by default, as many products already use it. sqlite itself is 'public domain' and the python da has the same license as zlib if i recall, which is very permissive as well.
IMO there is no need to ship the Zope core with *any* DB package except the ZODB. From the maintenance point of view we should basically ship with packages that we really need and that provide value to the core distribution. In Zope 2.6 we included Docutils with provided reStructuredText functionality. Updating and maintaining the package has been a minor pain in the past. Dealing with exploits in 3rd-party-packages (as it happened yesterday with Docutils) is another issue. The core distribution should only ship with packages where we have a certain amount of expertise to deal with such issues. I assume nobody of us is a sqlite core developer and knows the sources. Otherwise we end up like Plone and ship with more 3rd-party packages from release to release. -aj
Andreas Jung wrote:
IMO there is no need to ship the Zope core with *any* DB package except the ZODB.
I dunno, it's very handy to be able to rely on something RDB-like being installed when you want to tests things like ZSQL methods... cheers, Chris -- Simplistix - Content Management, Zope & Python Consulting - http://www.simplistix.co.uk
--On 11. Oktober 2005 07:50:34 +0100 Chris Withers <chris@simplistix.co.uk> wrote:
Andreas Jung wrote:
IMO there is no need to ship the Zope core with *any* DB package except the ZODB.
I dunno, it's very handy to be able to rely on something RDB-like being installed when you want to tests things like ZSQL methods...
Yes, it would be handy. But one could also install the stuff as standalone product for people that really need it. I am really against this proposal - maintaining yet another nice2have package is a burden just for the sake of make the lifes of some people easier. Other application servers also don't ship with a RDBMS....if you need it, install it. Andreas
Andreas Jung wrote:
Yes, it would be handy. But one could also install the stuff as standalone product for people that really need it. I am really against this proposal - maintaining yet another nice2have package is a burden just for the sake of make the lifes of some people easier. Other application servers also don't ship with a RDBMS....if you need it, install it.
I'm not in favour of adding anything else, I'm just not in favour of killing off Gadfly unless it is replaced with something similar and more functiona ;-) Chris -- Simplistix - Content Management, Zope & Python Consulting - http://www.simplistix.co.uk
Chris Withers wrote at 2005-10-11 07:50 +0100:
... I dunno, it's very handy to be able to rely on something RDB-like being installed when you want to tests things like ZSQL methods...
But, for "ZSQL" method tests you do not need a real DA; a DA fake is sufficient (you test "ZSQL" methods and not SQL command execution). -- Dieter
participants (7)
-
Andreas Jung -
Andrew Veitch -
Chris Withers -
Dieter Maurer -
Jens Vagelpohl -
Sidnei da Silva -
Stefan H. Holek