Windows LongPaths – extended-length paths


Maybe you are one of persons who belived for this moment that maximal length of path in Windows is equal to MAX_PATH ( 260 signs). Nothing further from the truth !!!.

In document which you can download below I have described inter alia:

– what is the maximum path length and from which it follows
– in how achieve possibility to create paths longer than MAX_PATH
– details related with WinApi, where path length and it’s type is tested.

this and more issues.

Of course, if some programmer assumed that maximal path length is 260 signs, his application in the best scenario won’t be fully functional in but in the worst scenario we can expect bugs that kind like buffer overflow.
To convince myself to validity of this theory, I decided to test anti-virus applications “tolerance” on such paths. Results of those tests you can review in a mentioned document, here I can only say that quite a lot of anti-virus showed described above signs.

And now, it’s time for a little gallery which IMO encourage you to treatment existence of long paths like a real problem:

AVAST 5.0.594.0


AVIRA 10.0.0.567


Norton 17.6.0.32


Sophos 9.5.1

At the end the newest version of gmer on day 15.10.2010 GMER 1.0.15.15315 ( the last bug reported by me has been patched, but like we can see there is still more problems in this area 😉 ).

The document can be downloaded from here: Windows LongPaths – extended-length paths

Greetz to Emiliano Martinez Contreras for a translation from IceEng to more readable form ;).

This entry was posted in Analiza, Bez kategorii, RE, Security and tagged , , , , , , , . Bookmark the permalink.

Comments are closed.