7
Replies
Hi pnk,
Michael has made an astute observation about your method of retrieving the table. There are likely many tables in the doc (esp. on the reference pages) and it is unpredictable which one will be returned as the first. If you have managed to get the right one so far, I think you may have been lucky. I would consider writing a separate function for this that has contingencies to iterate through all the tables in the doc, either checking the heading text or table format to determine if you've gotten the right one.
Having said that, your text implies that you have managed to get the right table. For getting a string from a F_TextItemsT structure, it is very simple. You don't need to cast it... StringTs are already members of the structure. For example:
i = F_StrSubString(textItems.val[0].u.sdata, paraName);
...should work find. There are some caveats to this example, though:
- For this to work in your code, there must be nothing else in the text of the cell except the P:etc. declaration. Anything else, such as line breaks, markers, whatever, will break up the string into multiple sdata members.
- You will want to keep that check that you currently have to make sure the structure isn't empty. An empty structure will invite a crash. For example:
if(textItems.len > -1)
i = F_StrSubString(textItems.val[0].u.sdata, paraName);
else
i = -1;
There are some other areas of your code that are a bit questionable, some that are more serious than others. One example:
if(!pgfName) return;
If your paragraph ID is valid, this should never return a null pointer. If you are attempting to check if the string is empty, use F_StrEqual() or F_StrCmp() instead.
Also:
props = F_ApiGetProps(docId,bodyFmtId);
if(props.len == 0) return;
There are a number of problems here. First of all, you don't deallocate props anywhere, so this will become a major memory leak. You have to deallocate these structures whenever you are done with them (F_ApiDeallocatePropVals() ). Secondly, the abrupt return in the event of a failure is generally considered suboptimal. Here is an idea... make this call at the very beginning of the function and just reuse props in the while loop. There is no need to populate props every time you want to apply a format to a paragraph... it is the same data every time. Even if you were deallocating properly, you would be forcing a lot of unnecessary processing (read "expensive");
Also, this part is convoluted:
ctr.beg.objId = ctr.end.objId = pgfId;
ctr.beg.offset = 0;
ctr.end.offset = FV_OBJ_END_OFFSET;
F_ApiSetTextRange(FV_SessionId, docId, FP_TextSelection, &ctr);
F_ApiSetProps(docId,ctr.beg.objId,&props);
...all you should need is:
F_ApiSetProps(docId, pgfId, &props);
By forcing a text selection every time, you will really slow down the process.
Also, you haven't deallocated the paragraph name which is another memory leak.
Also, your method of getting the text from the cell paragraph is convoluted.
Also, you could save some more time by breaking out of IsInConvTable() as soon as a match is found. You should do this since you are forcing this process with every paragraph in the main doc.
Also, you haven't made any provisions for cases where one paragraph name is contained in another, possibly returning a false positive. This would take some more string handywork.
So with all that, I've reposted your code with some suggestions. I didn't test any of this so I might have syntax errors, but I think the general ideas will make your code considerably more reliable and robust. I hope this helps some. I don't mean to sound like I'm bashing your code... I think this is good stuff... just want to help out 
Russ
F_PropValsT props;
. . .
props = F_ApiGetProps(docId,bodyFmtId);
if(props.len == 0) return;
. . .
while(pgfId) {
//Traverse through paragraphs in the current document (docId) and get their format names.
pgfName = F_ApiGetString(docId, pgfId, FP_Name);
if(F_StrCmp(pgfName, "") == 0)
{
F_ApiDeallocateString(&pgfName);
pgfId = F_ApiGetId(docId, pgfId, FP_NextPgfInFlow);
continue;
}
//Check if this pgfName is in the conversion table's first column..RICI, R2C1,R3C1...
retVal = IsInConvTable(convTabId, pgfName);
//retVal contains 1 if pgfName is in Conversion table, 0 otherwise..
if(!retVal)
F_ApiSetProps(docId, pgfId, &props);
F_ApiDeallocateString(&pgfName);
pgfId = F_ApiGetId(docId, pgfId, FP_NextPgfInFlow);
}
F_ApiDeallocatePropVals(&props);
}
int IsInConvTable(F_ObjHandleT convTabId, ConStringT paraName)
{
//declare tblId, rowId, cellId, paraId;
F_TextItemsT textItems;
IntT retVal=0, i;
//Get Table and Row Ids
tblId = F_ApiGetId(FV_SessionId, convTabId, FP_FirstTblInDoc);
rowId = F_ApiGetId(convTabId, tblId, FP_FirstRowInTbl);
while(rowId) {
//Get Cell Ids of only the first column in each row of the Conversion table.
cellId = F_ApiGetId(convTabId, rowId,FP_FirstCellInRow);
paraId = F_ApiGetId(convTabId, cellId, FP_FirstPgf);
textItems = F_ApiGetText(convTabId, paraId, FTI_String);
if (textItems.len != 0)
i = F_StrSubString(textItems[0].u.sdata, paraName);
else i = -1;
F_ApiDeallocateTextItems(&textItems);
if (i > -1)
{
retVal=1;
break;
}
rowId = F_ApiGetId(convTabId, rowId, FP_NextRowInTbl);
}
return retVal;
}
AdChoices