Some Common Error Messages
From DLXS Documentation
(Difference between revisions)
| (11 intermediate revisions not shown.) | |||
| Line 11: | Line 11: | ||
<Error>syntax error before: ot within region dsc))</Error></pre></code> | <Error>syntax error before: ot within region dsc))</Error></pre></code> | ||
| - | This is the result of having indexed unnormalized texts with xmlrgn. Normalization puts all the attributes into the order they are presented in the DTD (among other things). This is not a requirement of XML, but it is of nsgmls (invoked in the make xml process, but it's still sgml-based). Normalizing and reindexing will resolve the error. | + | This is the result of having indexed unnormalized texts with xmlrgn. Normalization puts all the attributes into '''the order they are presented in the DTD''' (among other things). This is not a requirement of XML, but it is of nsgmls (invoked in the make xml process, but it's still sgml-based). Normalizing and reindexing will resolve the error. |
| - | *'''pitstge''' -- | + | *'''pitstge''' -- Sometimes seen when searching for a word in an index that is somehow corrupted and/or truncated. For example, <code>pitstge: n = 154526 out of range. Size = 136844.</code> Searches in the index prior to the area of corruption/truncation work, which can mislead you into believing that it is a problem with the query. You will need to reindex to treat this error. |
| + | *'''pitstgr''' -- Sometimes seen when searching for a region in an index that is somehow corrupted and/or truncated. For example, <code>pitstgr: Start (0) out of range (size = 2, length = 6).</code> Searches in the index prior to the area of corruption/truncation work, which can mislead you into believing that it is a problem with the query. You will need to reindex at least the XML and the fabricated regions to treat this error. | ||
| + | *'''sortkey not equal''' -- Seen when the number of "main" regions (id, maintitle, mainauthor, and maindate) are unequal for a given text (or "region main"). For example, <code>Message: ID size="56", sortkey size="28" not equal, key=collid.</code> In this case, there are twice as many id values as region mains. This occurs when you have two or more IDNOs and have not fabricated region id to focus in on only one IDNO type. You might also see this when multiple authors are not honed down to just one mainauthor. More information is available in the section on [[Building_the_Text_Class_Index|Text Class indexing]]. | ||
*'''returning without an answer''' -- Sometimes seen when running xpatbldu, especially when the server is loaded. There may be a bug in xpatbldu that's triggered under certain conditions related to the size of memory allocated for the indexing process. If, for example, your index build command is <code>xpatbldu -m 256m -D $DLXSROOT/idx/c/coll/coll.dd</code>, try changing the memory to 255 or 254. | *'''returning without an answer''' -- Sometimes seen when running xpatbldu, especially when the server is loaded. There may be a bug in xpatbldu that's triggered under certain conditions related to the size of memory allocated for the indexing process. If, for example, your index build command is <code>xpatbldu -m 256m -D $DLXSROOT/idx/c/coll/coll.dd</code>, try changing the memory to 255 or 254. | ||
*'''madiopa''' -- seen when there are errors in the blank.dd. For example, <code>madiopa: Line 47, tag id To is invalid here</code>. In this case, on line 47 of the blank.dd, one of the mappings of punctuation to a space (to enable phrase searching) had gotten corrupted and the tag was no longer valid XML (/From> instead of </From>). | *'''madiopa''' -- seen when there are errors in the blank.dd. For example, <code>madiopa: Line 47, tag id To is invalid here</code>. In this case, on line 47 of the blank.dd, one of the mappings of punctuation to a space (to enable phrase searching) had gotten corrupted and the tag was no longer valid XML (/From> instead of </From>). | ||
| + | *'''Unable to open file: /l1/obj/c/coll/coll.xml. Errors in Data Dictionary.''' -- May be seen when using an index from the command line. If the file exists in the proper directory, check its size. If it is over 2 gigs, xpatu has acted as if it built an index, but it has not. Files must be below 2 gigs to be indexed. | ||
| + | *'''Rset = -1 ''' -- This is the sign of an empty result set. Seen in debugging messages as <code>XPatResultSet iterator limit (-1) exceeded</code>. Perhaps the result a bad fabricated region (one that works for another collection but not this one with less encoding), perhaps an ancient bookmark to something that no longer exists in an ever-changing collection. You will need to assess the cause and treat as appropriate. | ||
Current revision
- invalid endpoints -- Sometimes seen when fabricating regions, during the
make postprocess. For example:
sangria:bhlead % make post
cp /l1/prep/b/bhlead/bhlead.prepost.dd /l1/idx/b/bhlead/bhlead.dd
touch /l1/idx/b/bhlead/bhlead.init
/l/local/xpat/bin/xpatu -q /l1/idx/b/bhlead/bhlead.dd < /l1/prep/ b/bhlead/bhlead.extra.srch |
/l1/bin/t/text/output.dd.frag.pl /l1/ idx/b/bhlead/ > /l1/prep/b/bhlead/bhlead.extra.dd
Error found:
<Error>Pair regions 'did' have invalid endpoints.</Error>
Error found:
<Error>syntax error before: ot within region dsc))</Error>
This is the result of having indexed unnormalized texts with xmlrgn. Normalization puts all the attributes into the order they are presented in the DTD (among other things). This is not a requirement of XML, but it is of nsgmls (invoked in the make xml process, but it's still sgml-based). Normalizing and reindexing will resolve the error.
- pitstge -- Sometimes seen when searching for a word in an index that is somehow corrupted and/or truncated. For example,
pitstge: n = 154526 out of range. Size = 136844.Searches in the index prior to the area of corruption/truncation work, which can mislead you into believing that it is a problem with the query. You will need to reindex to treat this error. - pitstgr -- Sometimes seen when searching for a region in an index that is somehow corrupted and/or truncated. For example,
pitstgr: Start (0) out of range (size = 2, length = 6).Searches in the index prior to the area of corruption/truncation work, which can mislead you into believing that it is a problem with the query. You will need to reindex at least the XML and the fabricated regions to treat this error. - sortkey not equal -- Seen when the number of "main" regions (id, maintitle, mainauthor, and maindate) are unequal for a given text (or "region main"). For example,
Message: ID size="56", sortkey size="28" not equal, key=collid.In this case, there are twice as many id values as region mains. This occurs when you have two or more IDNOs and have not fabricated region id to focus in on only one IDNO type. You might also see this when multiple authors are not honed down to just one mainauthor. More information is available in the section on Text Class indexing. - returning without an answer -- Sometimes seen when running xpatbldu, especially when the server is loaded. There may be a bug in xpatbldu that's triggered under certain conditions related to the size of memory allocated for the indexing process. If, for example, your index build command is
xpatbldu -m 256m -D $DLXSROOT/idx/c/coll/coll.dd, try changing the memory to 255 or 254. - madiopa -- seen when there are errors in the blank.dd. For example,
madiopa: Line 47, tag id To is invalid here. In this case, on line 47 of the blank.dd, one of the mappings of punctuation to a space (to enable phrase searching) had gotten corrupted and the tag was no longer valid XML (/From> instead of </From>). - Unable to open file: /l1/obj/c/coll/coll.xml. Errors in Data Dictionary. -- May be seen when using an index from the command line. If the file exists in the proper directory, check its size. If it is over 2 gigs, xpatu has acted as if it built an index, but it has not. Files must be below 2 gigs to be indexed.
- Rset = -1 -- This is the sign of an empty result set. Seen in debugging messages as
XPatResultSet iterator limit (-1) exceeded. Perhaps the result a bad fabricated region (one that works for another collection but not this one with less encoding), perhaps an ancient bookmark to something that no longer exists in an ever-changing collection. You will need to assess the cause and treat as appropriate.
