- Zope-Dev - Zope lists

Problem in ZGadflyDA/__init__.py
by Gregor Hoffleit 11 Apr '00

11 Apr '00

There's a small but sometimes fatal problem in ZGadflyDA/__init__.py: __init__.py currently checks for the presence of a gadfly directory in Zope's var directory, and creates it if it's not yet there: j=os.path.join d=j(j(INSTANCE_HOME,'var'),'gadfly') if not os.path.exists(d): os.mkdir(d) os.mkdir(j(d,'demo')) Now if you start up z2.py as root, since __init__.py is called *before* z2.py tries to setuid/setgid, the gadfly directory will be created with the original uid and gid, i.e. root. Depending on the umask, this leaves the Zope process with insufficient permissions to access the gadfly directory. The suggested change is easy: Similar code is already in ZGadflyDA/db.py, so it should be safe to remove the lines above from ZGadflyDA/__init__.py. Gregor

2 1

incompatible time modules?
by Stefan Kuzminski 11 Apr '00

11 Apr '00

Hi all, I get the following error on starting zope.. Traceback (innermost last): File "/home/langur/Zope-2.1.6-src/z2.py", line 427, in ? import ZServer File "/home/langur/Zope-2.1.6-src/ZServer/__init__.py", line 123, in ? from HTTPServer import zhttp_server, zhttp_handler File "/home/langur/Zope-2.1.6-src/ZServer/HTTPServer.py", line 116, in ? from HTTPResponse import make_response File "/home/langur/Zope-2.1.6-src/ZServer/HTTPResponse.py", line 99, in ? from medusa import http_server, asyncore File "/home/langur/Zope-2.1.6-src/ZServer/medusa/http_server.py", line 746, in ? tz_for_log = compute_timezone_for_log() File "/home/langur/Zope-2.1.6-src/ZServer/medusa/http_server.py", line 729, in compute_timezone_for_log if time.daylight: AttributeError: daylight I built everyting from scratch on a linux box. The zope code seems to expect different things from the time module than the Python distribution includes. Is there a special zope version of the time module? I'm using python 1.5.2. Thanks, Stefan Kuzminski

1 0

Collector Status Update
by Tres Seaver 11 Apr '00

11 Apr '00

Working through the Collector, I have seen a need to separate out the patch for a bug from its description: for instance, the description is formatted using newline_to_br, which breaks any patch to Python code; patches need to be formatted inside <PRE> tags. I therefore added a shiny new "Patch" field, and will work on migrating those bugs/features with patches to move the patch. In the future, please stuff actual patches into this field (use 'diff -c' by preference). Tres. -- ========================================================= Tres Seaver tseaver(a)digicool.com tseaver(a)palladion.com

1 0

Re: [Zope-dev] TreeTag
by Tres Seaver 11 Apr '00

11 Apr '00

Chris Withers <chrisw(a)nipltd.com> > > <RANT> > Now that's the collector has been opened up you can see that certain > things have stagnated ;-) > If you search for TreeTag, you get seven items. > One is a patch fixing a problem I have (which is still there in 2.1.6), > another is a patch fixing a > problem I didn't know about, and one adds some really useful functionality. > They are all between 4 and 8 months old and yet all except one are still > 'pending' > Even with DC's limited resources that's not very good... > </RANT> > > There must be a better way, personally I'd like to know how I can help > and get these patches into the next point release as quickly as possible. > > Any ideas? As the guy who opened the Collector, and who is actively wading through it at the moment, I can sincerely say, "I feel your pain." I am working on Collector issues, prioritizing them for the next several releases -- patches as always get priority (assuming we can understand them, of course). We are actively working on ways to enable more community in the evaluate/integrate/test cycle, but are obviously not there yet. > PS: I was just about to go and fix these problems myself, so I'm very > glad the collector has opened up ;-) I'm glad that opening it up helped you. If you'd like to "vote" for a patch, try it out on a scratch Zope, with a copy of your "real" Data.fs, and verify that it doesn't raise any new problems, and that it fixes the reported problem. Send me mail, with the issue #'s in the subject, and let me know that it worked (or not -- and patches to patches are accepted, too :) Tres. P.S. I have two of the tree-tag patches (#913/1045) checked in already. -- ========================================================= Tres Seaver tseaver(a)digicool.com tseaver(a)palladion.com

2 2

Persistency of database method classes (pluggable brains)
by Dr. Peer Griebel 11 Apr '00

11 Apr '00

It's just me again... The ZSQL document mentions to define a database method class in a file in the Extensions directory. The ZCMG document says "you cannot define classes or objects in your External Methods and expect to be able to assign them to attributes of Zope objects." This is what I see in my application: I'm doing a SQL query. The result is stored in an array of object my own database method class. I save this array in an FSSession object to be used later on. This later usage results in problems. The array does not consist of objects of my own class (or of class r as defined in RDB.py) but of type extension class Record.Record. So how can I make my class definition persistent? The database method class lives in the Extension directory which means this class is not persistent by definition :-(( But perhaps I just overlooked something very simple. Still praying... Thank you! -- connection reset by Peer _______________________________________________________________________ Dr. Peer Griebel Tel. +49 7581 4831 23 Geschäftsführer Fax. +49 7581 4831 11 Knoll Informationssysteme GmbH http://www.knoll-is.de Dreiköniggasse 17 mailto:peer@knoll-is.de 88348 Saulgau privat: mailto:peer.griebel@gmx.de

1 0

Sorting/Reversing dtml-in with URL arguments
by Kevin Smith 11 Apr '00

11 Apr '00

Is it possible to set the sort order and reverse flags on a dtml-in tag using URL arguments or REQUEST variables? For example, sorting "employees" by "name" in index_html would be something like: index_html?sort for <dtml-in employees> ... </dtml-in> Kevin Smith Kevin.Smith(a)theMorgue.org carpe noctum

2 1

Adding data attributes to a pluggable brain?
by Dr. Peer Griebel 11 Apr '00

11 Apr '00

Hi, I just started using pluggable brains. The first thing I observed is that changes to the brain class are only seen in Zope after a restart of Zope. This differs to the behaviour of external methods. For newbies this can be quite irritating. But now to my question: How can I add a data attribute to my brain? Whenever I try to say self.attr=42 in a method of the brain class I get an Attribute Error. I can use a construct like self.__dict__['attr']=42. This seems to work. But is it the right way? Thank you for your help! -- connection reset by Peer _______________________________________________________________________ Dr. Peer Griebel Tel. +49 7581 4831 23 Geschäftsführer Fax. +49 7581 4831 11 Knoll Informationssysteme GmbH http://www.knoll-is.de Dreiköniggasse 17 mailto:peer@knoll-is.de 88348 Saulgau privat: mailto:peer.griebel@gmx.de

1 0

Re: [Zope-dev] Problem in ZGadflyDA/__init__.py
by Tres Seaver 11 Apr '00

11 Apr '00

Gregor Hoffleit <flight(a)hoffleit.de> > To: Jim Fulton <jim(a)digicool.com>, zope-dev(a)zope.org > Subject: > > There's a small but sometimes fatal problem in ZGadflyDA/__init__.py: > > __init__.py currently checks for the presence of a gadfly directory in > Zope's var directory, and creates it if it's not yet there: > > j=os.path.join > d=j(j(INSTANCE_HOME,'var'),'gadfly') > if not os.path.exists(d): > os.mkdir(d) > os.mkdir(j(d,'demo')) > > Now if you start up z2.py as root, since __init__.py is called *before* > z2.py tries to setuid/setgid, the gadfly directory will be created with the > original uid and gid, i.e. root. > > Depending on the umask, this leaves the Zope process with insufficient > permissions to access the gadfly directory. > > The suggested change is easy: Similar code is already in ZGadflyDA/db.py, so > it should be safe to remove the lines above from ZGadflyDA/__init__.py. So that we don't lose it in the list, please submit this change as a "bug-with-patch" to the Collector: http://classic.zope.org:8080/Collector/developer_input Thanks! Tres. -- ========================================================= Tres Seaver tseaver(a)digicool.com tseaver(a)palladion.com

1 0

Actual objects in ZODB not visable to management screen or the Catalog
by Jason Spisak 10 Apr '00

10 Apr '00

This is seriously strange. I have not seen this before...ever. What would prevent and Zclass instance of meta_type 'ZGL_Client' that is viewable by the URL http://localhost:8080/955393378581117/955393587277964 To not be seen anywhere in the Management interface, or indexed by the Catalog using the 'Find Items' tab. Basically it doesn't exist anywhere that I can see, except to put in that URL. "Objects should appear in folder lists" is checked for both this object and it's parent object. The management interface shows nothing, and the Catalog only indexes it if I type in index_object after the URL and force it. And for the final kicker...the old "Find" tab shows that there are no 'ZGL_Client' meta_type in the ZODB. The instance reacts fine to all of the changes I make to the ZClass too. It is just invisible. Any ideas? -- Jason Spisak 444(a)hiretechs.com

1 0

TreeTag
by Chris Withers 10 Apr '00

10 Apr '00

<RANT> Now that's the collector has been opened up you can see that certain things have stagnated ;-) If you search for TreeTag, you get seven items. One is a patch fixing a problem I have (which is still there in 2.1.6), another is a patch fixing a problem I didn't know about, and one adds some really useful functionality. They are all between 4 and 8 months old and yet all except one are still 'pending' Even with DC's limited resources that's not very good... </RANT> There must be a better way, personally I'd like to know how I can help and get these patches into the next point release as quickly as possible. Any ideas? cheers, Chris PS: I was just about to go and fix these problems myself, so I'm very glad the collector has opened up ;-)

1 0