On 6/23/05, F.W.A. van Leeuwen <[EMAIL PROTECTED]> wrote:
> While cross-compiling sqlite3.2.1, my Blackfin C compiler gives the
> following warning:
>
> ".\src\pager.c", line 1342: cc0111: {D} warning: statement is unreachable
> assert( rc==SQLITE_OK );
>
> The context is:
>
> <begin code snippet>
> /* This loop terminates either when the readJournalHdr() call returns
> ** SQLITE_DONE or an IO error occurs. */
> while( 1 ){
>
> // CODE OMITTED WITH NO break;
>
> /* rc = sqlite3OsSeek(&pPager->jfd, JOURNAL_HDR_SZ(pPager)); */
> if( rc!=SQLITE_OK ) goto end_playback;
>
> /* Copy original pages out of the journal and back into the database
> file.
> */
> for(i=0; i<nRec; i++){
> rc = pager_playback_one_page(pPager, &pPager->jfd, 1);
> if( rc!=SQLITE_OK ){
> if( rc==SQLITE_DONE ){
> rc = SQLITE_OK;
> pPager->journalOff = szJ;
> break;
> }else{
> goto end_playback;
> }
> }
> }
> }
>
> /* Pages that have been written to the journal but never synced
> ** where not restored by the loop above. We have to restore those
> ** pages by reading them back from the original database.
> */
> assert( rc==SQLITE_OK );
> pager_reload_cache(pPager);
>
> end_playback:
> <end code snippet>
>
> I think the compiler is right. I am not familiar with the sqlite3 source
> code, but the break at line 1330 breaks the for() loop at line 1324, NOT
> the while() loop at line 1281. Is that as it is intended, or is it a
> potential bug?
The for loop can terminate after i>=nRec. I'm not sure that's possible
in practice
or not though.
