getnamo
December 5, 2016, 9:09am
23
I’ll be taking a deeper look at the memory management of the new async architecture this week. In the meantime I wanted to ask about 7z library.
I noticed that some archives of mine that were created on Linux contain illegal file names for windows. When I unzip these, 7z seems to automatically replace the invalid characters with underscores. The problem is that I use the archive listing functions to generate the archive contents map, so my map contains the illegal characters and I can’t find the files after extraction.
I didn’t see a 7z call that could convert an invalid file name to a valid one. Do you know anything about where this happens? I may need to write some transfer functions if they’re not accessible.
Great, making sure we gc appropriately and don’t have memory leaks would be helpful.
Not sure about the name conversion. The underlying library is found here: GitHub - getnamo/7zip-cpp: Fork of SevenZip++ for modern builds. which links to the raw 7z sources here: GitHub - keithjjones/7z at c8852a6dcaf99ae0fb29de8d97736ffa1b3a180e
You should have plenty of access to modify and add any missing underlying functionality. Note that we use the COM-port approach to interacting with the 7z dll and that there is a UE4 specific branch that has slightly different includes for the lib to hide windows data types.