ISFDB Bug List/Archive/Archive02

Jump to navigation Jump to search
  • 10060 FIXED New interviews don't properly integrate. Adding a new interview to a publication will cause the interview to show up on both the interviewer's and interviewee's bibliographies, but a pub_content record is not created, so it doesn't show up in the publication listing. Alvonruff 05:40, 21 Dec 2006 (CST)
  • 20030 FIXED When searching for 'Zelazny' or 'Roger Zelazny' only a blank page is displayed even though there is an extensive entry for him. PortForlorn 1:08 30 Aug 2006 (EST)
Searching on either 'Zelazny' or 'Roger Zelazny' pulls up his page for me. Have you tried deleting cached files in your browser? Ahasuerus 08:27, 30 Aug 2006 (CDT)
I finally had time to research this further and discovered that when there is only one author matching a search request, the code embeds a javascript redirection directly to the author's listing rather than show a one entry list of candidates. But, for security reasons, I never browse with javascript enabled. Thus my page display is blank. So this isn't a bug - but it would be "nice" to see a warning (to those of us who are more paranoid than most) that javascript is disabled. PortForlorn 13:50 15 Sep 2006 (EST)
Oh, I see! Yes, that makes a lot of sense. Perhaps a general warning to the effect that some functionality within the ISFDB requires JavaScript, Al? Ahasuerus 13:42, 15 Sep 2006 (CDT)
  • 20037 FIXED I clicked on the first link in the section on "Editing omnibuses" in the Community Portal and got the following error:

IndexError Python 2.4.2: /usr/local/bin/python Tue Oct 17 23:48:17 2006

A problem occurred in a Python script. Here is the sequence of function calls leading up to the error, in the order they occurred.

 254         PrintHeader(title)
 256         publication = SQLGetPubByTag(tag)
 257         if publication == 0:
 258                 sys.exit(1)

publication undefined, SQLGetPubByTag = <function SQLGetPubByTag>, tag = 'THSHDWSFCH2004'

/www/ in SQLGetPubByTag(tag='THSHDWSFCH2004')
 363         result = db.store_result()
 364         pub = result.fetch_row()
 365         return pub[0]
 367 def SQLGetPubById(tag):

pub = ()

IndexError: tuple index out of range

     args = ('tuple index out of range',)

Even if that is a bad link now, pl.cgi really shouldn't bomb out. (Perhaps every externally callable cgi script should check its arguments.) PortForlorn 23:53, 17 Oct 2006 (CDT)

A problem occurred in a Python script. Here is the sequence of function calls leading up to the error, in the order they occurred.
  300                 elif len(publication[PUB_ISBN]) == 10:
  301                         print '  <li><b>ISBN-10:</b>', publication[PUB_ISBN]
  302                         print '  <li><b>ISBN-13:</b>', isbn13(publication[PUB_ISBN])
  303                 elif len(publication[PUB_ISBN]) == 13:
  304                         print '  <li><b>ISBN-13:</b>', publication[PUB_ISBN]
isbn13 = <function isbn13>, publication = (82853L, 'Das Kuckucksei', 'DSKCKCKSPZ1988', '1988-00-00', 4433L, None, 'tp', 'NOVEL', 'B000DCO2Y0', '', None, None), PUB_ISBN = 8
 /www/ in isbn13(isbn='B000DCO2Y0')
  229         newISBN = '978' + isbn[0:9]
  231         sum1 = int(newISBN[0]) + int(newISBN[2]) + int(newISBN[4]) + int(newISBN[6]) + int(newISBN[8]) + int(newISBN[10])
  232         sum2 = int(newISBN[1]) + int(newISBN[3]) + int(newISBN[5]) + int(newISBN[7]) + int(newISBN[9]) + int(newISBN[11])
  233         checksum = sum1 + (sum2 * 3)
sum1 undefined, builtin int = <type 'int'>, newISBN = '978B000DCO2Y'

ValueError: invalid literal for int(): C
      args = ('invalid literal for int(): C',)

Ahasuerus 17:54, 18 Oct 2006 (CDT)

    • I just found another of these and, sure 'nuf, it had an ISBN10 with some alpha characters embedded within the first 10 positions. PortForlorn 21:10, 18 Oct 2006 (CDT)
Title: The Playboy Book of Science Fiction and Fantasy 
Authors: Anonymous 
Year: 1966-00-00 
Publisher: Chicago, Illinois: Playboy Press 
Price: $0.95 
Pages: 402 
Type: hc --> -->  
IndexError Python 2.4.2: /usr/local/bin/python
Wed Nov 1 01:22:58 2006 

A problem occurred in a Python script. Here is the sequence of function calls leading up to the error, in the order they occurred.

  329         if publication[PUB_CTYPE] != 'MAGAZINE':
  330                 title_id = SQLgetTitleReferral(publication[PUB_PUBID])
  331                 if title_id:
  332                         print '<li>'
title_id undefined, SQLgetTitleReferral = <function SQLgetTitleReferral>, publication = (45399L, 'The Playboy Book of Science Fiction and Fantasy', 'PBBSFF1966', '1966-00-00', 2942L, '402', 'hc', 'NOVEL', None, None, '$0.95', None), PUB_PUBID = 0 
 /www/ in SQLgetTitleReferral(pub_id=45399L) 
  723         result = db.store_result()
  724         title_id = result.fetch_row()
  725         while title_id[0][0]:
  726                 query = "select title_ttype from titles where title_id=%d;" % title_id[0][0]
  727                 db.query(query)
title_id = () 

IndexError: tuple index out of range 
      args = ('tuple index out of range',)

--Marc Kupper 02:27, 1 Nov 2006 (CST)

  • 20042 (also 10043, merged with this one) FIXED There's a typo in the verification display. Consider BKTG05892. It says Verfied by Grendelkhan on 2006-11-21 10:12:28 when it should say Verified by Grendelkhan on 2006-11-21 10:12:28. grendel|khan 10:14, 21 Nov 2006 (CST). Similarly, when you display a verified pub, e.g. this one, the reference status text at the bottom says it is "Verfied", rather than "Verified". Mike Christie 06:22, 5 Nov 2006 (CST)
  • 20043 FIXED If you are in the alphabetical or chronological bibliography displays, the "Awards" link causes a python error. Mike Christie 07:34, 24 Nov 2006 (CST)
In fact the editing tools generally seem to behave oddly; the titles screens lists nothing; "Find Stray Publications" finds nothing, and the Author Data link causes a python error. Mike Christie 15:48, 24 Nov 2006 (CST)
A problem occurred in a Python script. Here is the sequence of function calls leading up to the error, in the order they occurred.
   57         if found:
   58                 bib.printShortSeries()
   59                 bib.displayWorks('SHORTFICTION')
   60                 bib.displayWorks('POEM')
bib = <biblio.Bibliography instance>, bib.printShortSeries = <bound method Bibliography.printShortSeries of <biblio.Bibliography instance>>
 /www/ in printShortSeries(self=<biblio.Bibliography instance>)
  304                 if self.au_seriesnum:
  305                         print "<b>Short Fiction Series</b>"
  306                 newlist = self.tree.Print(self.au_titles, '', self.au_id, self.au_collabs, self.au_serials, self.au_variants)
  307                 self.au_titles = newlist
newlist undefined, self = <biblio.Bibliography instance>, self.tree undefined, self.au_titles = [], self.au_id = 7154L, self.au_collabs = [], self.au_serials = [], self.au_variants = []

AttributeError: Bibliography instance has no attribute 'tree'
      args = ("Bibliography instance has no attribute 'tree'",)

Reported by Ahasuerus 19:31, 20 Nov 2006 (CST)

AttributeError	Python 2.4.2: /usr/local/bin/python
Mon Dec 18 17:34:54 2006

A problem occurred in a Python script. Here is the sequence of function calls leading up to the error, in the order they occurred.
   57         if found:
   58                 bib.printShortSeries()
   59                 bib.displayWorks('SHORTFICTION')
   60                 bib.displayWorks('POEM')
bib = <biblio.Bibliography instance>, bib.printShortSeries = <bound method Bibliography.printShortSeries of <biblio.Bibliography instance>>
 /www/ in printShortSeries(self=<biblio.Bibliography instance>)
  304                 if self.au_seriesnum:
  305                         print "<b>Short Fiction Series</b>"
  306                 newlist = self.tree.Print(self.au_titles, '', self.au_id, self.au_collabs, self.au_serials, self.au_variants)
  307                 self.au_titles = newlist
newlist undefined, self = <biblio.Bibliography instance>, self.tree undefined, self.au_titles = [], self.au_id = 22958L, self.au_collabs = [], self.au_serials = [], self.au_variants = []

AttributeError: Bibliography instance has no attribute 'tree'
      args = ("Bibliography instance has no attribute 'tree'",)

Ahasuerus 17:36, 18 Dec 2006 (CST)

  • 20054 FIXED I ran an "Advanced Search" on Author contains "unknown". The first page came up fine, but when I clicked on the URL at the bottom to get the next page, I got an error. Ahasuerus 17:39, 17 May 2006 (CDT)
  • 20053 FIXED When you select Title Merge for an author who has a lot (hundreds) of Works, the list displays the first page with 100 records correctly, but then it gets weird. The next page displays records 50-150 or thereabouts, the following one shows records 140-240 (?), etc. Jack WIlliamson and especially Robert Silverberg are good examples of this behavior. Ahasuerus 08:57, 2 May 2006 (CDT)
  • 20055 FIXED It looks like some Author records have no Works or Publications associated with them, e.g. Harrison Ford. I wonder if the bug is in one of the old conversions, e.g. ISFDB1-to-ISFDB2, or whether it's currently possible to delete the last Work/Publication for an Author and leave the Author record behind? Ahasuerus 16:18, 12 Jun 2006 (CDT)
The database is fine. Harrison's name remains because he is the target of an interview (see James Verniere. Since we don't display interviewees on the interviewee's summary page, it appears that Harrison is a loose end. Alvonruff 12:36, 16 Dec 2006 (CST)
To prevent further confusion, I've added a feature that we lost from ISFDB1->ISFDB2: output of interviews about the author. This allows us to see why Harrison is still present in the database. Alvonruff 13:21, 19 Dec 2006
Has this feature gone live yet? I don't seem to be able to find it at Harrison_Ford. Ahasuerus 13:43, 19 Dec 2006 (CST)
  • 20050 FIXED When a submission has been rejected, the displayed message is misspelled, e.g. "Record 79691 has been move to the Rejected state". 19:22, 28 Nov 2006 (CST)
  • 10036 FIXED When you try to make an existing Title into a pseudonymous Title, the default Title Type is always "Anthology" regardless of what the Title's actual Type currently is. Similarly, if you pull up an existing Collection or Omnibus Publication (with no pre-existing contents) for editing, the default value for Contents is "Anthology". It sould be "Shortfiction" for Collections and "Novel" for Omnibuses instead. Ahasuerus 19:41, 13 Sep 2006 (CDT)
  • 10032 FIXED The following problems appear to be related to the fact that merging Titles will result in all but one Title IDs being dropped, so any other records pointing to them (Variant Titles, Awards, etc) will now be pointing to a non-existing Title record. Should we upgrade the Merge script to relink all records to the resulting Title ID?
    • Suppose that A and B are both Long Works and C is a variant title of B. If you make B a variant title of A, then C will remain a variant title of B even though it should presumably become a variant title of A. Moreover, since it is now a variant title of a variant title, it no longer appears in the Author's Long Works Biblio at all, e.g. Jules Verne's The English at the North Pole (UK 1874). Ahasuerus 18:32, 29 Jun 2006 (CDT)
  • 10027 FIXED Editing a review or interview will result in loss of the subject author. Alvonruff 21:10, 5 Jun 2006 (CDT)
  • 10034 CLOSED Marion Zimmer Bradley's Logn Works Bibliography displays The Ages of Chaos as "[ONone]". If you pull it up in Title Edit, the "Storylen:" field is set to "None". I have no idea how it got there while I was editing MZB's biblio, but there you have it. IIRC, it was listed as a standalone novel and I changed it to a Darkover Omnibus, but there may have been more to it since I have modified well over a dozen MZB titles recently. Ahasuerus 16:04, 18 Aug 2006 (CDT)
This was probably fixed in the cleanup I did a month or so ago, and wasn't the only field that could generate 'None' (which Python generates when a null string is dereferenced). I can't reproduce this any longer, so I'm going to close it. Alvonruff 13:02, 16 Dec 2006 (CST)
  • 10035 WONTFIX - If you are not signed in and try to merge two or more titles, you are prompted to sign in. Once you sign in, you see a dead end page instead of being able to continuing with the merge process. Ahasuerus 13:09, 25 Aug 2006 (CDT)
I'm generally open to fixing these redirect pages after a login, but some of them are intractible. In this case, the submit does an HTML POST (not a GET), and the receiving application must obtain the data passed to it from the environment (it is not passed as an argument, like an HTML GET would be). To handle it, the login page would have to have the merge code placed in it to obtain the data, and then it would have to perform a followup POST, which would probably have to be done from a separate Javascript script. There are a few screens that Wikipedia doesn't handle for similar reasons. So I'm going to let this one slip by unfixed. Alvonruff 12:48, 16 Dec 2006 (CST)
  • 10030 CLOSED Entering an entry in a magazine with a blank page number causes the page not to appear in the display for that magazine, although the XML looks like it's there. This happened with the February 1957 issue of Super-Science Fiction. Mike Christie 19:46, 13 Jun 2006 (CDT)
No. Saw your note which said you entered that title as a type 'EDITOR' - the type should be 'ESSAY'. I went to W. W. Scott's page, found the record, changed it to an ESSAY, and it displays fine. Publications with a type of NOVEL generate a title record that is of type NOVEL, an ANTHOLOGY pub generates an ANTHOLOGY title, and a MAGAZINE pub generates.... well, there is no MAGAZINE title type - it generates an 'EDITOR' title. The EDITOR titles allow works to show on the editor's bibliography (if you look at W._W._Scott you'll see the two issues of Super-Science Fiction under the Editor section). As such, it would be goofy for a magazine listing to display EDITOR records, so they don't, and since the title was of that type, it didn't show up.
Although not a display bug, it is a UI bug as the current setup is clearly confusing for people new to the ISFDB data structures. Alvonruff 21:06, 13 Jun 2006 (CDT)
That makes sense and it means the magazine was processed flawlessly. I don't have a full run of S-S F, but I'll enter the others I have this week -- not till Thursday at the earliest, it looks like. And I don't think it's strictly a display bug; I'd suggest we make sure we have good help descriptions in place when we go live. Thanks. Mike Christie 22:13, 13 Jun 2006 (CDT)
This situation was handled by documenting the help screen for new publications. Alvonruff 12:39, 16 Dec 2006 (CST)
  • 10016 FIXED When modifying the Year field (and only the Year field) in the following Work record, the Moderator screen showed two changes. The first one was the Year change, which was correct. The second change, however, was bogus. The Work had a pre-existing Wikipedia URL, which I didn't touch, yet the screen reported that there was no URL in the old version and there was a URL in the new submitted version. When I approved the submission, everything was filed correctly, though. Ahasuerus 13:09, 17 May 2006 (CDT)
  • 10013 FIXED I added Kris Neville and Lin Carter to as Authors. Then I realized that they were already listed as the Authors of the two individual stories (novels, to be charitable) that comprise this Belmont Double. And, since this Publication is an Anthology, they were displayed as Editors, which is not what I wanted. I tried removing them, but when I blanked out both Author fields, my submission was rejected since the Author field was blank. I can get around it by entering "Anonymous" as the Publication Author/Editor, but shouldn't the error message tell me to enter "Anonymous"? Or, better yet, insert "Anonymous" when there is no Author specified in the form? Ahasuerus 21:17, 13 May 2006 (CDT)
As you know, I'm generally against the app doing stuff I don't tell it to do. The tool now notifies the user to use "Anonymous" if the blank author line was an intentional act. Alvonruff 11:11, 16 Dec 2006 (CST)
  • Sorta FIXED. When editing a Publication, the "changes" counter displayed at the bottom of the XML display page doesn't seem to take changes to Artists, Notes (?) and possibly other fields into account. Ahasuerus 18:16, 14 May 2006 (CDT)
Well the "changes" stuff is only present in 4 of the numerous submission apps, presents a lot of maintenance whenever changes are made, can be argumentative ("I made 2 changes to the note, but it said only 1 change was made"), and it doesn't provide any more feedback than what's in the XML record. So I took the "Changes" line out of those 4 apps. If that feedback is useful, then we can file a new feature to get it properly implemented across all of the apps. Alvonruff 11:02, 16 Dec 2006 (CST)
  • 10038 FIXED When converting a Title into a Variant Title, the software lets you point the title record to itself. Not unexpectedly, it causes all kinds of problems with displaying and editng. You can't delete the title since "there is a title linked to it", but the workaround is to merge it with some other title. I propose that we disallow linking titles to themselves. Ahasuerus 13:38, 1 Oct 2006 (CDT)
  • 10022 FIXED I tried to update the Galaxy Science Fiction Novels edition of Hal Clement's "Mission of Gravity". I changed the catalog number to 33, from blank; and I changed the artist to "Wallace A. Wood", from "Wallace Wood". The first change took but the second did not. I tried again, and discovered in moderating that the first submission is still in the approval queue. Approving either one now does nothing; no SQL statement is generated for the artist record. Mike Christie 13:00, 24 May 2006 (CDT)
  • 10028 FIXED Deleting the last work of Hope_Gangloff didn't delete the author record, because the COVERART record didn't get deleted. The publication looked like "* Electric Storm (2003 , Scholastic, 0439314348, $3.99, 96pp, pb) Cover: Hope Gangloff" in the title-view. If you want to look at a database backup, the publication was LCTRCSTRM2003 before I deleted it. I deleted the COVERART record separately, and that deleted the author record--but it should have been deleted along with the publication in the first place. (Also, the COVERART record was title 139349.) grendel|khan 20:42, 10 Jun 2006 (CDT)
  • 10025 FIXED Deleting a publication and title that has a cover artist attached to it does not delete the title record for Cover:whatever. In the case of title 186209, it should have been deleted along with the author record, as it's the last work for that author. grendel|khan 16:43, 31 May 2006 (CDT)
  • 30002 FIXED If a series is not numbered, then its entries will be displayed sorted by publication date in the Long Works list, which is, one assumes, as it should be. However, in the Series screen, the entries are displayed in the order in which they were entered into ISFDB, i.e. following their internal database record numbers. E.g., see , which displays one way in Silverberg's Long Works list and the other way on the Series screen. Ahasuerus 18:37, 4 May 2006 (CDT)
  • 30010 CLOSED The four bogus "series" in Homer Eon Flint's Short Works Bibliography suggest that there may be a bug in one of our series conversion/manipulation tools. Ahasuerus 12:04, 13 Jun 2006 (CDT)
    The bug was no longer recreatable. Either the software or the data had been fixed. Ahasuerus 20:48, 15 Dec 2006 (CST)
  • 30011 FIXED There were three entries in the database for Plague Daemons, which was written by Brian Stableford and published as by Brian Craig. One title was as by Stableford, one as by Brian Craig and one as by Craig and Stableford together. One of the three records was a variant title, although I don't recall which one. I merged these three records (which, in retrospect, was probably unwise) and the resulting record doesn't appear on Stableford's Long Works page at all, although it appears on the Series Bibliography page for Orfeo. Ahasuerus 16:22, 3 Jul 2006 (CDT)
    The problem was that the resulting Title was pointing to itself. The root cause was presumably fixed as part of 10023 -- see below. Ahasuerus 20:44, 15 Dec 2006 (CST)
  • 10023 FIXED Damon Knight's novella "Dio" is a variant title of "The Dying Man". The ISFDB listed it twice as "Dio" (once as a separate short and once as a variant title of "The Dying Man") and twice as "Dying Man". After mergig these two pairs, the association between "Dio" and the "Dying Man" has been lost and they are now listed separately in Knight's Short Works Bibliography. The problem appears to be that when you merge two Works and one of them has a Variant Title pointing to it, that pointer is not examined to see whether it needs to be repointed to the record that survived the merge. If you are merging two records, then you have a 50-50 chance of ending up with a dangling pointer. Multiple merges can result in self-pointing records, e.g. this Fred Pohl story. Ahasuerus 09:44, 31 May 2006 (CDT)
  • 10055 FIXED While trying to change the name of an Author from "Mary E. Wilkins" to "Mary E. Wilkins Freeman" in this Title, I encountered the following Python error:
IndexError Python 2.4.2: /usr/local/bin/python
Sun Dec 3 02:44:07 2006 

A problem occurred in a Python script. Here is the sequence of function calls leading up to the error, in the order they occurred.

   47         new = titles(db)
   48         new.cgi2obj()
   50         old = titles(db)
new = <titleClass.titles instance>, new.cgi2obj = <bound method titles.cgi2obj of <titleClass.titles instance>> 
 /www/ in cgi2obj(self=<titleClass.titles instance>) 
  178                 while counter < 100:
  179                         if self.form.has_key('title_author'+str(counter+1)):
  180                                 self.title_authors[self.num_authors] = XMLescape(self.form['title_author'+str(counter+1)].value)
  181                                 self.num_authors += 1
  182                         counter += 1
self = <titleClass.titles instance>, self.title_authors = ['Gertrude Atherton', 'Isabella Banks', 'Ambrose Bierce', 'Robert W. Chambers', 'Amelia B. Edwards', 'Elizabeth Gaskell', 'Charlotte Perkins Gillman', 'W. W. Jacobs', 'E. Nesbit'], self.num_authors = 9, global XMLescape = <function XMLescape>, self.form = FieldStorage(None, None, [MiniFieldStorage('titl...OLOGY'), MiniFieldStorage('title_id', '177270')]), builtin str = <type 'str'>, counter = 9, ].value undefined 

IndexError: list assignment index out of range 
      args = ('list assignment index out of range',)

The same change to the Author list in the related Publication went through successfully. Note that there are only 10 Authors in the Pub record, but 11 in the Title record. Ahasuerus 02:48, 3 Dec 2006 (CST)

  • 10056 FIXED - If you create an ANTHOLOGY with contents that include a NOVEL entry type, the title reference for the ANTHOLOGY title is the title record associated with the NOVEL, not the ANTHOLOGY. I was trying to clean up the "Treasury of Great SF": see 1313, which lists "The Chrysalids" as the title reference.
I assume the bug is that if a NOVEL entry type is spotted, perhaps only as the first entry, then that's used as the title reference. So I'm going to try re-entering the data without any novels and then editpub to add novels. If that works, it's a workaround, but I think this is a fairly high priority bug since it leads to incorrect data, though not actual loss of data. Mike Christie (talk) 09:29, 9 Dec 2006 (CST)
I just tested this with entering story/novel/story, and it still created the title reference to the novel, so it doesn't matter where in the sequence the novel is. Mike Christie (talk) 09:33, 9 Dec 2006 (CST)
I've now tested the following sequence: create anthology with two SHORTFICTIONS; approve that; editpub to add a NOVEL content, approve that. This changes the title reference on the publication to point to the novel. I think this needs to be fixed for the beta and will make a note on that page. Mike Christie (talk) 09:40, 9 Dec 2006 (CST)
Not an editing bug. This was a problem in the heuristic to try to determine what the parent title actually is. I've changed it to match the publication type to the title type. This is more reliable, but has the following side effect: if the publication type is wrong (say, it's set to NOVEL, when the title is an ANTHOLOGY), then NO parent title is displayed. I've also changed the title and pub display code to display the types (for some reason it's never done that before), as well as in the publication history, so the mismatch is a lot more noticable now. Alvonruff 11:02, 15 Dec 2006 (CST)
I was looking up some other stuff and ran across Template:TitleFields:Translator which says “Translator - Do not enter anything in this field; it is reserved for future use.” That explains why there is an “undefined field name” Python blow-up and so the “bug report” can be modified into a “feature request” to remove the field from the edit screens until it does get implemented. Marc Kupper 18:32, 12 Dec 2006 (CST)
• update titles set title_xlate='Ros Schwartz' where title_id=10330  --> -->

OperationalError	Python 2.4.2: /usr/local/bin/python
Mon Dec 11 19:21:22 2006

A problem occurred in a Python script. Here is the sequence of function calls leading up to the error, in the order they occurred.
   65                 UpdateColumn(merge, 'Title',      'title_title',      Record)
   66                 UpdateColumn(merge, 'Translator', 'title_xlate',      Record)
   67                 UpdateColumn(merge, 'Year',       'title_copyright',  Record)
   68                 UpdateColumn(merge, 'Seriesnum',  'title_seriesnum',  Record)
UpdateColumn = <function UpdateColumn>, merge = [<DOM Element: TitleUpdate at 0x837a40c>], Record = '10330'
 /www/ in UpdateColumn(doc=[<DOM Element: TitleUpdate at 0x837a40c>], tag='Translator', column='title_xlate', id='10330')
   34                 print "<li> ", update
   35                 if debug == 0:
   36                         db.query(update)
global db = <_mysql.connection open to 'localhost' at 832280c>, db.query = <built-in method query of Connection object>, update = "update titles set title_xlate='Ros Schwartz' where title_id=10330"

OperationalError: (1054, "Unknown column 'title_xlate' in 'field list'")
      args = (1054, "Unknown column 'title_xlate' in 'field list'")
Note that I've only removed the translator fields from the forms (which technically fixes this bug). Translator support still remains on the needed feature list. Alvonruff 10:43, 15 Dec 2006 (CST)
  • 10058 FIXED - If you are on mkvariant.cgi (Make Variant Title) and in the upper part of the page where you can enter a Parent # if you click [Submit Data] without entering a number (to delete a variant relationship) you will get a Python NameError.
NameError	Python 2.4.2: /usr/local/bin/python
Thu Dec 14 18:05:21 2006

A problem occurred in a Python script. Here is the sequence of function calls leading up to the error, in the order they occurred.
   44         update_string += "    <Subject>%s</Subject>\n" % (db.escape_string(title[TITLE_TITLE]))
   45         update_string += "    <Record>%s</Record>\n" % (title_id)
   46         update_string += "    <Parent>%s</Parent>\n" % (parent_id)
   47         update_string += "  </MakeVariant>\n"
   48         update_string += "</IsfdbSubmission>\n"
update_string = '<?xml version="1.0" encoding="iso-8859-1" ?>\n<Is...rt 1 of 2)</Subject>\n <Record>218771</Record>\n', parent_id undefined

NameError: name 'parent_id' is not defined
      args = ("name 'parent_id' is not defined",)
  • 10011 FIXED When merging Works across Authors, e.g. Philip Klass' Immodest Proposals: The Complete Science Fiction of William Tenn: Volume 1 into William Tenn's biblio, at least some award information is lost. See , which was present in the Klass entry, but didn't make it across during the merge. Ahasuerus 11:44, 9 May 2006 (CDT)
    • Update: This happens when merging Titles within the same Author as well. There were two occurrences of J. F. Gonzalez's "Maternal Instinct", one a "NOVEL" with an associated Publication and one a "COLLECTION" with no associated Publication, but with an award record attached. After they were merged, the resulting "COLLECTION" Work was pointing to the correct Publication, but the Award information was lost. Ahasuerus 22:55, 20 May 2006 (CDT)
IndexError	Python 2.4.2: /usr/local/bin/python
Wed Nov 22 18:06:00 2006

A problem occurred in a Python script. Here is the sequence of function calls leading up to the error, in the order they occurred.
  104         else:
  105                 record = ((0, '', '', '', '', '', '', '', '', ''),)
  106                 printauthrecord(record)
  108         print "<p>"
printauthrecord = <function printauthrecord>, record = ((0, '', '', '', '', '', '', '', '', ''),)
 /www/ in printauthrecord(record=((0, '', '', '', '', '', '', '', '', ''),))
   69         printfield("Wikipedia Entry", "author_wikipedia",  record[0][AUTHOR_WIKI])
   70         printfield("IMDB Entry",      "author_imdb",       record[0][AUTHOR_IMDB])
   71         printfield("Author Image",    "author_image",      record[0][AUTHOR_IMAGE])
   73         print "</table>"
global printfield = <function printfield>, record = ((0, '', '', '', '', '', '', '', '', ''),), global AUTHOR_IMAGE = 11

IndexError: tuple index out of range
      args = ('tuple index out of range',)
the latter URL leads to , which brings up an empty "Title Merge" page. Ahasuerus 08:48, 24 May 2006 (CDT)
  • 10010 FIXED I am not 100% sure, but this is what I think happened. Merging two Works (_The Omega Point Trilogy_, a novel and an omnibus) which both pointed to the same Publication resulted in a Work that pointed to the same Publication (the same URL) twice. An attempt to Unmerge the Work resulted in the contents of the Omnibus listing regular omnibus contents once, but the name of the Omnibus Work was listed twice. I am not sure it wasn't that way originally, though, since I dind't check the contents before I unmerged it. In the end, I deleted the publication and merged the titles back, but now I need to recreate the Publication's contents, which will have to wait until contents editing goes live. Ahasuerus 20:29, 5 May 2006 (CDT)
  • 10044 FIXED Not sure if this is an editing or display bug, but I think it's the former. See this page and click on the Farmer link. I used Alt-130 to enter the e acute accent. I gather from Ahasuerus that there is some support for Unicode, so I'll call this a bug; if I should change it so it's "Jose" rather than "José" everywhere, just let me know. Mike Christie 16:06, 5 Nov 2006 (CST)
I'm probably going to need more data on this one - I tend to use the Windows Character Map to copy and paste non-ascii characters (which is how José got into the database in the first place). How does the Alt-130 thing work? (I get nothing). Is it the data entry that doesn't work, or is the subsequent link busted? Could you enter a bogus magazine as an example to work with? Alvonruff 17:11, 22 Nov 2006 (CST)
The ALT thing is a long standing part of windows (and MS-DOS too). You use the numeric keypad. Try ALT-65 and the letter ... A ... will appear as "A" is ASCII code 65 decimal. The accent é is character code 130. I use these a lot £ (156) ¢ (162) • (149). If you use the Windows Character Map then you can see the code (in hex) as you hover the mouse over a character. Note that these are Windows character codes and not Unicode values. Use ALT with a four digit value to get Unicode characters. For example, the Windows character set predates the European Union and for the Euro € you use unicode 0128 and I use “0147 and 0148” to get the opening/closing double quotes or ‘0145 and 0146’ for the single quotes. Marc Kupper 22:55, 22 Nov 2006 (CST)
Marc's description is exactly what I did. If you haven't used that data entry method, I'll just add that you hold the Alt key down while entered all the digits. You asked the sequence I went through; I was entering the content title for a Farmer story and simply used "é" in entering his name in the author field. It worked and saved correctly, but had the results you see. It should be very easy to duplicate; let me know if you have trouble for some reason. Mike Christie 07:34, 23 Nov 2006 (CST)
  • 30012 FIXED A search on a series name where the series contains one or more subseries but no Works (e.g. Legend of the Five Rings) finds no data. Ahasuerus 19:29, 4 Jul 2006 (CDT)
Orson Scott Card's "Ender's Shadow" Three-Book Set (Ender's Shadow, Shadow of the Hegemon, and Shadow Puppets) 

which is OK, but when you pull it up in Editing Tools: Publication, the title is "Orson Scott Card's ". Looks like quotes are not being handled properly. I'll leave it as is for now. Ahasuerus 21:00, 4 May 2006 (CDT)

  • 10006 WORKSFORME Fred Saberhagen's biblio had a problem. Somebody had merged his The Third Book of Swords (1984) with The Second Book of Lost Swords: Stonecutter's Story (1988). When I unmerged the record, something strange happened, but I am not sure I captured the sequence of events. In the end, both Work records had the following Awards data:
Title: The Third Book of Swords 
Author: Fred Saberhagen 
Year: 1984 
Series: Book of Swords 
Series Number: 3 
1985 - The Third Book of Swords Locus Poll Award, Best Fantasy Novel (Place: 22) 


Title: Stonecutter’s Story 
Author: Fred Saberhagen 
Year: 1988 
Series: Book of Lost Swords 
Series Number: 3 
1985 - The Third Book of Swords Locus Poll Award, Best Fantasy Novel (Place: 22) 

How does one get rid of this field? Ahasuerus 18:12, 2 May 2006 (CDT

I selected a Saberhagen award-winning title, unmerged it, and found that only 1 title (the original) had an award attached to it. This may have been fixed somewhere in the past. I did make an unrelated fix to the unmerge, such that it generates title records that are identical to the publication record, instead of the original title (which is what would be desired in the above case). Alvonruff 19:03, 28 Nov 2006 (CST)
  • 10001 FIXED When merging titles, the screen that shows all the discrepancies doesn't show the Notes section, so a submitter may accidentally delete Notes data that he didn't even see. You can see the contents that is about to be deleted in the Moderator's screen, but the Moderator may not know why the submitter deleted the Notes. The same problem exists with the Wikipedia link field Ahasuerus 22:41, 30 Apr 2006 (CDT)
  • 10051 FIXED Copying bug report from community portal:
Consider THRLLNGSTP2005. I created it as follows:
  1. I created THRLLNGSTN2005 with the Add Publication to This Title tool.
  2. I cloned it and changed the ISBN, noting that it had 1284 listed as contents at that point.
  3. I accepted the edit in the moderation screen, and saw there that it had contents listed. (Specifically, one copy of the title it was created for.)
  4. The resultant new publication was listed twice in the title list, and appeared to contain four copies of itself when viewed, or two when edited.
Hope this helps track it down. (Originally reported by Grendelkhan.) Mike Christie 09:41, 25 Nov 2006 (CST)
  • 10053 FIXED (Possibly the same bug as 10044) Apostrophe problems seem to exist with content entry. In the December 1956 Fantastic Universe there's a story by "Vithaldas O'Quinn" (almost certainly a pseudonym of Hannes Bok). The data entry seems to go fine, but then when you click on the author name in the contents, the URL says ". . . Vithaldas%20O'Quinn" but the page title says "Vithaldas O\’Quinn". Clicking on "Short Works", the title is "Vithaldas O\\’Quinn" and the URL is ". . . Vithaldas_O\%E2%80%99Quinn". Mike Christie 16:55, 26 Nov 2006 (CST)
  • 10052 FIXED TitleRemove fails for book reviews. I tried deleting Sam Merwin's book review "REVIEW" type record from the June-July 1953 Fantastic Universe, and it approved OK but the record is still there. Mike Christie 09:52, 25 Nov 2006 (CST)
  • 20045 FIXED The "Recent Verifications" display has links to the publications, but they error out: it appears the script is expecting a tag but getting an ID. Mike Christie 17:47, 24 Nov 2006 (CST)
  • 10049 FIXED It looks to me as if the "COVERART" entry type should be removed from the content fields. Experimenting with the way cover art records worked I discovered that if you add a coverart content entry to a publication that has a cover artist already recorded, the result is that two copies of the added artist are added to the list of cover art artists. E.g. if you have magazine "Test Mag" with "Emsh" as the artist, and you add a content record with "Ed Emshwiller" as the artist, the publication will subsequently show three cover artists: "Emsh", "Ed Emshwiller", and "Ed Emshwiller". It seems there's no need for the "COVERART" type in the newpub or editpub screens anyway, since the functionality works fine from the pub entry fields. The only exception I can think of might be dos-a-dos books, where you might want two independent cover art records. Mike Christie 10:01, 23 Nov 2006 (CST)
Removed from the editpub.cgi app only. Alvonruff 12:10, 24 Nov 2006 (CST)
ProgrammingError Python 2.4.2: /usr/local/bin/python
Thu Nov 23 22:17:59 2006 

A problem occurred in a Python script. Here is the sequence of function calls leading up to the error, in the order they occurred.

   39         # Find pubs that refer to this title
   40         pubs = SQLGetPubsByTitle(int(record))
   41         if len(pubs) > 0:
   42                 NoneFound = 0
pubs undefined, SQLGetPubsByTitle = <function SQLGetPubsByTitle>, builtin int = <type 'int'>, record = '182283' 
 /www/ in SQLGetPubsByTitle(titlerec=182283) 
  438         results = []
  439         db.query(query)
  440         result = db.store_result()
  441         pub = result.fetch_row()
global db = <_mysql.connection open to 'localhost' at 837180c>, db.query = <built-in method query of Connection object>, query = 'select pubs.* from pubs,pub_content where pub_content.pub_id=pubs.pub_id ) order by pubs.pub_year' 

ProgrammingError: (1064, "You have an error in your SQL syntax; check the manual that corresponds to your MySQL server version for the right syntax to use near ') order by pubs.pub_year' at line 1") 
      args = (1064, "You have an error in your SQL syntax; check the ... to use near ') order by pubs.pub_year' at line 1") 

Ahasuerus 22:20, 23 Nov 2006 (CST)

  • 10015 NAB For when the contents-editing tools go live: Publication CWLSXFRVSH2005 includes the title which references it. Perhaps it would be useful if by default, when adding a new publication to a NOVEL, its contents were set to that title, like in this publication. grendel|khan 22:10, 16 May 2006 (CDT)
    Scratch that; I edited the publication (to add pagecount and edit publication date, see update 38513), and it changed the previously unset pub_ctype to 'NOVEL', and in the process made the 'Contents' data vanish. Hopefully I didn't mash anything permanent. And none of the other publications of that title have Contents filled in. Does the request still make sense? grendel|khan 22:20, 16 May 2006 (CDT)
    This is known and expected behavior when there is a mismatch between Publication Type and Title Type, in this case Collection vs. Novel. Ahasuerus 18:23, 22 Nov 2006 (CST)
Reviews are a finicky area - like serials, there is currently no database linkage between the review and the title. Matches are via the title string. In the given case, the old "canonical title" was The Avram Davidson Treasury (which matched the review titles by the same name), and now it's The Avram Davidson Treasury: A Tribute Collection (which no longer matches the review titles). I changed the title of the reviews, and now they show back up on the title bibliography. Currently a limitation in the review implementation. Alvonruff 20:52, 9 May 2006 (CDT)
  • 10018 WORKSFORME I'm unsure as to how, but I managed to create an empty TitleMerge edit, number 38807. I can't apply or delete it, or even view it. I wish I could remember how it got there. grendel|khan 11:56, 18 May 2006 (CDT)
Yeah, the XML data just has a submitter and nothing else, so not much to go on. I forced it into the rejected state. Alvonruff 12:56, 18 May 2006 (CDT)
  • 10007 FIXED When editing Publication data, clicking on the "Edit Content" link (e.g. results in an error. If this happens because content editing hasn't been enabled yet, we may want to disable the link. Ahasuerus 20:33, 3 May 2006 (CDT)
    No longer an issue ever since the new "Edit Contents" functionality was added. Ahasuerus 17:47, 22 Nov 2006 (CST)
  • 10019 FIXED When merging two publications, if you pick the nondefault tag, it won't show on the confirmation screen just before it's actually committed, but it will apply as picked; the issue is just that it's displayed wrong. The tags in this example are, respectively, BKTG14233 (for pub 6802) and BKTG14232 (for pub 6803). The generated XML is:
    <?xml version="1.0" encoding="iso-8859-1" ?>

<IsfdbSubmission> <PubMerge> <Submitter>Grendelkhan</Submitter> <Subject>Child of Thunder</Subject> <KeepId>6802</KeepId> <DropId>6803</DropId> <Tag>6803</Tag> </PubMerge>

</IsfdbSubmission> However, the confirmation screen doesn't show the proper tag selection, even though it will supposedly apply. This probably isn't a big deal, as it looks like publication merges aren't supported yet (despite showing the proper log spew, after posting this same edit previously as 38886, title 15478 looks the same.) You'd have to do some mojo to deal with cover art if it came up, I suppose. I'll just delete extra publications for now, except for this example here. grendel|khan 17:20, 18 May 2006 (CDT)

  • Merging publications - Refrain from use until I look at this. I had completely forgotten that this existed, and cannot vouch for it's veracity. Alvonruff 12:49, 25 May 2006 (CDT)
    Ah, nuts. I forgot why I wasn't running that publication-merge through, and submitted it. It's still visible as submission 38888, and the publications have not been merged. (Also, does publication merging, when turned on, deal with records for the cover artist?) grendel|khan 08:02, 2 Jun 2006 (CDT)
    This has been now "fixed" by removing the ability to merge Publications. Ahasuerus 17:40, 22 Nov 2006 (CST)
  • 10020 WONTFIX Something is causing Publications to be listed multiple time per Work. For example, this page displays the same Publication three times. I don't know how the first two got there, but the third one joined them when I merged two Works which were pointing to the same Publication. Ahasuerus 22:49, 21 May 2006 (CDT)
This is a known issue with some Publications somehow acquiring multiple pointers to the same Title. This will be addressed by adding logic to the data integrity checker as per Al. For now, I fixed this Publication manually by deleting the Title pointer (which deleted all pointers to this Title), re-adding the Title to the Publication, and then merging the newly created Title record with the existing one. Ahasuerus 17:37, 22 Nov 2006 (CST)
  • 10040 FIXED When adding a Publication to an existing Non-Genre Title, the default Pub Type is "Anthology" instead of "Novel". Ahasuerus 11:54, 8 Oct 2006 (CDT)
  • 10045 NAB Content editing: the pulldown for content type does not include Review. If you add extra content, the additional fields do have Review in the pulldown, so there is a workaround. Mike Christie 19:31, 8 Nov 2006 (CST)
The entire concept of this one was throwing me, as reviews a a different beast from regular titles as it needs to track both the author of the review as well as the author of the reviewed work - so reviews have their own separate content section, and by design you can't select REVIEW from the regular content section. Looking at an entry for Fantastic Universe shows what you're talking about: you've entered Merwin's book review column (Universe in Books) as a book review itself, which isn't really the desired effect, as the column shows up as a review, and makes for a confusing display if you click on it in Merwin's bibliography (it shows up with Review of: Universe in Books, with Book Author: NULL). The convention we've been using is that each individual review goes into the REVIEW section, and the overall article is an ESSAY. (In fact there are probably hundreds of magazines where we don't even track the encapsulating article, just the individual reviews). Alvonruff 16:48, 22 Nov 2006 (CST)
Aha. That clears that up. I guess the "REVIEW" type then shouldn't show up in the added pulldowns -- what I thought was a workaround is in fact undesirable behaviour? I'll go back and change the "REVIEW" types to essay in the Fantastic Universes I've entered. I think it's good to have the essay entry and certainly does no harm. Mike Christie 16:51, 22 Nov 2006 (CST)
The interesting thing will be what you have to do to clear it up - you may have to delete it, and re-enter it as an essay. I agree that it is good to have the essay entry, and I've been putting it in myself for the last year or so. Alvonruff 17:06, 22 Nov 2006 (CST)
This now support lengths up to 32 characters. Alvonruff 16:23, 22 Nov 2006 (CST)
  • 10033 NAB - resolved with Contents Deletion going live. This publication displays "A Darkover Retrospective", an essay by Marion Zimmer Bradley, 4 times. When the Publication is pulled up in the Pub editor, the essay appears twice. When I tried deleting one of the essay's instances, the submission tried deleting "The Waterfall" (another Work in the same Pub) instead, but failed. Leaving as is for debugging purposes. Ahasuerus 22:01, 26 Jul 2006 (CDT)
  • Not entirely surprising, as the ability to delete a title is currently on the To-Do list. The real question is: what editing action did you take to attempt to delete the title? (Blank out the the title?). Curious as to what people think is the most-natural way of deleting a title... Alvonruff 19:22, 12 Aug 2006 (CDT)
  • Ah, yes, I remember it now. (I seem to say this a lot these days...) Yes, I believe I blanked out the title/author in the Content Editing section. Perhaps we could have a "Delete" checkbox instead? Ahasuerus 21:48, 12 Aug 2006 (CDT)
  • 10041 FIXED Adding a particularly long and punctuation-heavy Web page address to Author Data resulted in an error. See submission 52726. Ahasuerus 21:48, 17 Oct 2006 (CDT)
  • 10042 FIXED I managed to type info intended for a title's Notes section ( in the Synopsis section and then approved this edit. When I went back to correct this, I discovered that attempting to just delete the text doesn't work. The deletion shows correctly on the title edit page and on the moderator approval page. But the text isn't deleted after approval. Similarly for trying to replace the text with a single blank. The only way I cound "remove" the text was to replace it with a non-blank character (e.g., ".") (which, of course, will still show a Synopsis entry). PortForlorn 00:56, 22 Oct 2006 (CDT)
  • 10046 FIXED An attempt to merge Author records 34775 (Richard J. O'Brien) and 34499 (Richard, J. O'brien) resulted in the following error:
SQL Updates:


update authors set author_canonical='Richard J. O'Brien' where author_id='34499' --> --> 
ProgrammingError Python 2.4.2: /usr/local/bin/python
Sat Nov 11 22:54:13 2006 

A problem occurred in a Python script. Here is the sequence of function calls leading up to the error, in the order they occurred.

  169         merge = doc.getElementsByTagName('AuthorMerge')
  170         if merge:
  171                 AuthorMerge(db, recno, doc)
  173         submitter = GetElementValue(merge, 'Submitter')
AuthorMerge = <function AuthorMerge>, db = <_mysql.connection open to 'localhost' at 82adc0c>, recno = '56232', doc = <xml.dom.minidom.Document instance> 
 /www/ in AuthorMerge(db=<_mysql.connection open to 'localhost' at 82adc0c>, recno='56232', doc=<xml.dom.minidom.Document instance>) 
   94         id = GetElementValue(merge, 'Canonical')
   95         if id and id != KeepId:
   96                 moveAuthorColumn(db, 'author_canonical', KeepId, DropId)
   98         id = GetElementValue(merge, 'Legalname')
global moveAuthorColumn = <function moveAuthorColumn>, db = <_mysql.connection open to 'localhost' at 82adc0c>, KeepId = '34499', DropId = '34775' 
 /www/ in moveAuthorColumn(db=<_mysql.connection open to 'localhost' at 82adc0c>, column='author_canonical', keep='34499', drop='34775') 
   68         update = "update authors set %s='%s' where author_id='%d'" % (column, value, int(keep))
   69         print "<li> ", update
   70         db.query(update)
   72 ########################################################################
db = <_mysql.connection open to 'localhost' at 82adc0c>, db.query = <built-in method query of Connection object>, update = "update authors set author_canonical='Richard J. O'Brien' where author_id='34499'" 

ProgrammingError: (1064, "You have an error in your SQL syntax; check the manual that corresponds to your MySQL server version for the right syntax to use near 'Brien' where author_id='34499'' at line 1") 
      args = (1064, "You have an error in your SQL syntax; check the ...e near 'Brien' where author_id='34499'' at line 1") 

Ahasuerus 23:10, 11 Nov 2006 (CST)

  • 10047 WONTFIX I changed the spelling of Sam Merwin's name on "A World Apart" in the offending issue of Fantastic Universe from "Sam Merwin Jr." to the regularized "Sam Merwin, Jr.". I expected "Sam Merwin Jr." to be deleted automatically the way Author record are when the last associated Title and Publication is deleted ot changed to another Author, but it stuck around. See this URL. Ahasuerus 23:10, 11 Nov 2006 (CST)
There were two book reviews which referenced the "Sam Merwin Jr." version, so it didn't delete the author name. There is currently no direct linkage between books and their reviews, so finding this situation automatically will be difficult. I have added a tool on the author's page that will show these stray titles. I tested it on a copy of yesterday's ISFDB, and was able to fix the Merwin situation without difficulty. Alvonruff 15:03, 22 Nov 2006 (CST)
This only happens when the Publication type is not "Anthology" or "Omnibus". Once the Publication type is changed to match Title type, the problem goes away. Ahasuerus 14:04, 22 Nov 2006 (CST)