On Tue, Nov 5, 2019 at 9:11 AM Elijah Newren <newren@xxxxxxxxx> wrote: > On Mon, Nov 4, 2019 at 8:14 AM Fossies Administrator <Jens.Schleusener@xxxxxxxxxxx> wrote: > > > On Mon, Nov 4, 2019 at 7:07 AM Fossies Administrator > > > <Jens.Schleusener@xxxxxxxxxxx> wrote: > But I thought it might also be worthwhile to you to report what the > false positives found by that program were; I've included them at the > end of this email in the form of a patch. The places where the program > seemed to struggle were: > > * In dealing with translation files. It didn't recognize them as > such and often tried to translate foreign words to a nearby English > one. > * In handling variable names: acronyms might be similar to english > words (cas, for compare and swap, looks like case), abbreviations > might look like alternate words (ans, short for answer, looks like > and). > * Testcases with intentional spelling errors > * Proper names that were similar to English words (Ned -> Need, > Claus -> Clause) > * miscellaneous tech jargon or package names (e.g. 'filetest' module > being replaced with 'file test', 'ith' as in not first or second > but the item at position i being replaced with 'with', 'mmaped' > being replaced with 'mapped', 'CREAT' changing to 'CREATE', > 'UserA' (out of a sequence of UserB, UserC, etc.) changing to > 'users', 'spawnve' function name being replaced with "spawn", > 'CAs' (certificate authorities) being replaced with 'case', etc.) Ooh, one more I remembered that I wanted to point out. It found the spelling error 'achiving', but it wanted to replace it with 'achieving' rather than the correct 'archiving'. Given that the correct is the same edit distance from the spelling error, it made me wonder whether the dictionary in use just needed to be expanded a little.
