1
0
Fork 0
mirror of https://github.com/KingDuckZ/dindexer.git synced 2024-11-26 01:03:45 +00:00
Commit graph

11 commits

Author SHA1 Message Date
42acc42d8f Fix wrong assertions. 2016-05-18 22:25:41 +02:00
6b5fe85571 Make mime_type and mime_charset on the fly
For the same motivation explained in the parent commit, make string_refs
on the fly instead of storing them.
2016-05-17 21:12:44 +02:00
8c03ba15bc Crash fix - make path on the fly
Newer versions of gcc have a small string optimizations that makes
moving FileRecordData objects behave incorrectly. When a small string is
moved, string_refs into it become invalid. Making a path on the fly
using the path() method also has the side effect of making
FileRecordData smaller in size.
2016-05-17 20:21:44 +02:00
d2588d3c7e Implement hashing task.
Also get rid of the ShorFileRecordData and put the new LeanBase
class to use.
2016-03-08 08:48:12 +01:00
d579b311f2 Add ShortFileRecordData struct to hold the file list in DirTree. 2016-03-04 21:57:09 +01:00
be9fc3eb0b Fix various problems introduced in scan with the last commit.
Fixes a crash due to an assertion trying to access a
past-the-end iterator.
Gets content type detection to work, although only after
hashing is done (see comment in main.cpp for details).
Fixes a problem with the array passed to guess_content_type
being not sorted as the function expects.
Adds more assertions.
2016-02-23 18:56:07 +01:00
22614432a9 Detect content type after scanning and save to DB. 2016-02-22 19:44:48 +01:00
f20d73bdab Update year on copyright notice. 2016-02-05 21:02:40 +01:00
fa12bb2d74 Bugfix - don't store absolute paths to files in the DB.
This was introduced in a previous commit.
2016-01-11 12:53:37 +00:00
f5356e0c29 Change namespace in machinery to mchlib. 2016-01-05 12:49:27 +00:00
4e29200b47 Move scan's core code into new machinery lib. 2016-01-05 12:35:46 +00:00
Renamed from src/scan/recorddata.hpp (Browse further)