throws InvalidFormatException {
URI partNameURI;
try {
- partNameURI = new URI(partName);
+ partNameURI = new URI(resolvePartName(partName));
} catch (URISyntaxException e) {
throw new InvalidFormatException(e.getMessage());
}
}
return retPartName;
}
+
+ /**
+ * If part name is not a valid URI, it is resolved as follows:
+ * <p>
+ * 1. Percent-encode each open bracket ([) and close bracket (]).</li>
+ * 2. Percent-encode each percent (%) character that is not followed by a hexadecimal notation of an octet value.</li>
+ * 3. Un-percent-encode each percent-encoded unreserved character.
+ * 4. Un-percent-encode each forward slash (/) and back slash (\).
+ * 5. Convert all back slashes to forward slashes.
+ * 6. If present in a segment containing non-dot (?.?) characters, remove trailing dot (?.?) characters from each segment.
+ * 7. Replace each occurrence of multiple consecutive forward slashes (/) with a single forward slash.
+ * 8. If a single trailing forward slash (/) is present, remove that trailing forward slash.
+ * 9. Remove complete segments that consist of three or more dots.
+ * 10. Resolve the relative reference against the base URI of the part holding the Unicode string, as it is defined
+ * in ?5.2 of RFC 3986. The path component of the resulting absolute URI is the part name.
+ *</p>
+ *
+ * @param partName the name to resolve
+ * @return the resolved part name that should be OK to construct a URI
+ *
+ * TODO YK: for now this method does only (5). Finish the rest.
+ */
+ public static String resolvePartName(String partName){
+ return partName.replace('\\', '/');
+ }
}
Enumeration<? extends ZipEntry> entries = this.zipArchive.getEntries();
while (entries.hasMoreElements()) {
ZipEntry entry = entries.nextElement();
- if (entry.getName().equals(
+ if (entry.getName().equalsIgnoreCase(
ContentTypeManager.CONTENT_TYPES_PART_NAME)) {
try {
this.contentTypeManager = new ZipContentTypeManager(
try {
// We get an error when we parse [Content_Types].xml
// because it's not a valid URI.
- if (entry.getName().equals(
+ if (entry.getName().equalsIgnoreCase(
ContentTypeManager.CONTENT_TYPES_PART_NAME)) {
return null;
}
// We assume we can continue, even in degraded mode ...
logger.log(POILogger.WARN,"Entry "
+ entry.getName()
- + " is not valid, so this part won't be add to the package.");
+ + " is not valid, so this part won't be add to the package.", e);
return null;
}
}
} catch(IllegalStateException e) {}
}
}
+
+ /**
+ * A problem file from a non-standard source (a scientific instrument that saves its
+ * output as an .xlsx file) that have two issues:
+ * 1. The Content Type part name is lower-case: [content_types].xml
+ * 2. The file appears to use backslashes as path separators
+ *
+ * The OPC spec tolerates both of these peculiarities, so does POI
+ */
+ public void test49609() throws Exception {
+ XSSFWorkbook wb = XSSFTestDataSamples.openSampleWorkbook("49609.xlsx");
+ assertEquals("FAM", wb.getSheetName(0));
+ assertEquals("Cycle", wb.getSheetAt(0).getRow(0).getCell(1).getStringCellValue());
+
+ }
}