Commit Graph

84 Commits

Author SHA1 Message Date
Ryan C. Gordon 32da097a55 Cleanup BAIL_* and GOTO_* macros to be less cluttered.
This also fixes really aggressive compiler warnings about the "if (!ERRPASS)"
construct producing unreachable code.
2017-07-06 11:51:41 -04:00
Ryan C. Gordon 062bdc2c5b Patched to compile with OpenWatcom. 2017-07-05 23:49:10 -04:00
Francesco Bertolaccini e4206408e5 Add support for VDF 2017-06-20 13:22:41 -04:00
Ryan C. Gordon 66b7767cec Added support for password-protected .zip files.
To use: mount a .zip file as usual, open a file as usual, but append '$' plus
the password to the end of the path, like so:

    PHYSFS_File *f = PHYSFS_openRead("/path/to/my/file.txt$MyPassword");

Note that this is the "traditional" PKWARE crypto, like you might get from
PkZip 2.04g or Info-ZIP. They have more advanced cryptography in the modern
.zip spec, but that's not implemented here. If you've ever tried to unzip
an archive and were prompted for a password, you probably saw the traditional
crypto at work.

Note that this is NOT a secure thing: if you ship an app that contains a
password, someone with a debugger can retrieve it. Note also that this
password could be available to any code registered as a PHYSFS_Archiver, and
that due to how archives combine into a single file tree, the wrong archiver
can get the password, at a minimum causing it to fail because passwords don't
match.

In short: make sure you know what you're doing before you use this!
2016-09-12 15:38:30 -04:00
Ryan C. Gordon ab7eb81a4a Platform deinit should happen last, since other deinit bits might depend on it. 2016-09-07 21:21:59 -04:00
Ryan C. Gordon 32d4f4541e HTTPS all the things. 2016-02-25 02:51:28 -05:00
Ryan C. Gordon b479c57bcb zip: Reworked how we organize ZIP entries.
We now use a hashtable for lookups of specific paths, and organize the
entries into a directory tree. The end result is fast lookup and fast
enumeration without having to search a sorted array or tapdance with
substrings...which means the rare, mysterious bug where we failed
to find an existing file should be gone now, too.
2016-02-25 02:20:52 -05:00
Ryan C. Gordon 57bbcc3eb3 Fixed compiler warning about unused variable because assert() is awful. 2015-07-02 23:36:33 -04:00
Ryan C. Gordon f6a62ecad1 Make sure read buffers get reinitialized before refilling (thanks, Michael!). 2014-09-11 00:07:23 -04:00
Ryan C. Gordon 53ef674270 Fixed resource leak when failing to mount a file that isn't an archive. 2014-08-19 02:28:13 -04:00
Willi Schinmeyer 9df22f4bf5 Fix NULL flush() in PHYSFS_Io crash 2014-07-20 13:36:56 +02:00
Ryan C. Gordon 016394067c PHYSFS_stat() returns zero on error, not -1. 2014-02-07 16:22:43 -05:00
Ryan C. Gordon 5aa0dae5e2 Added a FIXME. 2014-01-16 07:03:42 -08:00
Ryan C. Gordon 5dd2cfce35 Fixed a single-line comment. 2014-01-16 07:03:34 -08:00
Ryan C. Gordon 4046759f61 Better basedir detection for various Unix platforms. 2013-08-23 23:41:35 -04:00
Ryan C. Gordon 4eafb44817 Get rid of array of static archivers. 2012-11-30 14:37:37 -05:00
Ryan C. Gordon 36b542ee7f Cleaned out "exists" nonsense in openRead() API. 2012-11-30 13:00:08 -05:00
Ryan C. Gordon 697c31e81b Remove __PHYSFS_setError(), use the new public API instead. 2012-11-30 12:49:25 -05:00
Ryan C. Gordon 7824a093fb Cleaned out the "exists" nonsense in the stat() API. 2012-11-30 12:43:56 -05:00
Ryan C. Gordon ada24e9c5d Moved supportsSymlinks into PHYSFS_ArchiveInfo. 2012-11-30 00:04:52 -05:00
Ryan C. Gordon 16ee8355db Centralize responsibility for filtering symlinks during enumeration. 2012-11-29 11:47:16 -05:00
Ryan C. Gordon e40d80b00f Allow application-supplied archivers.
This lets an application supply its own archivers, where they will work like
 any built-in archiver. This allows abstract directory interfaces the same
 way that PHYSFS_Io allows stream implementations.

This is a work in progress still. The API is still changing, and will remain
 at version 0 until it is finalized (a theoretical future version 1 will be
 for when the final public interface changes, not when we evolve the initial
 API design).
2012-11-28 01:36:13 -05:00
Aleksi Nurmi fd137cba42 SLB archiver 2012-11-12 23:40:29 +02:00
Ryan C. Gordon 5c18db7fe9 Don't sort if there's nothing to do (prevents array underflow, too). 2012-06-07 10:21:43 -04:00
Ryan C. Gordon a865b14bec Added Zip64 support to the .zip archiver.
Now we can handle .zip files > 4 gigabytes, etc.
2012-06-01 05:44:50 -04:00
Ryan C. Gordon 758e5beedb Moved PHYSFS_ArchiveInfo into PHYSFS_Archiver. 2012-03-25 17:17:56 -04:00
Ryan C. Gordon 6222a4a92d Add binary compatibility to PHYSFS_Io. 2012-03-25 17:18:24 -04:00
Ryan C. Gordon a7383c295f Moved some functions around.
We want the #undef malloc, etc, lines as close to the end of the file as
 possible, and __PHYSFS_readAll() not in the middle of the allocator code.
2012-03-25 16:15:10 -04:00
Ryan C. Gordon 107f07da97 Better cleanup if PHYSFS_init() fails halfway through. 2012-03-25 16:13:23 -04:00
Ryan C. Gordon 4cfd7277f3 Made archiver list dynamic, in preparation for external archivers. 2012-03-25 15:59:29 -04:00
Ryan C. Gordon 0de337eb75 Renamed PHYSFS_Archiver::dirClose() to PHYSFS_Archiver::closeArchive(). 2012-03-24 10:36:11 -04:00
Ryan C. Gordon b2e6882f52 Another FIXME. 2012-03-24 00:32:22 -04:00
Ryan C. Gordon dda627412c More FIXME cleanup. 2012-03-24 00:26:04 -04:00
Ryan C. Gordon 9d01a645ed Make sure base dir always has a dirsep at the end of it.
Now the higher level doesn't have to check for this and realloc the string.
2012-03-23 04:14:01 -04:00
Ryan C. Gordon 0584117f85 Changed so that this builds a reasonable default with no command line #defines.
That is, it now builds .zip and native dir support, and that's it.
2012-03-23 00:52:32 -04:00
Ryan C. Gordon 76fdd9cab6 Removed a FIXME: actually, the code DOES set an error, as far as I can tell. 2012-03-22 02:58:45 -04:00
Ryan C. Gordon 46125c95b3 Removed a FIXME. 2012-03-22 02:57:29 -04:00
Ryan C. Gordon a0af6bbb71 Removed fallback for systems that have no userdir.
This would try to build something under the basedir for Windows 95, OS/2, etc,
 but those targets are gone now. Modern systems provide this. If a given
 system can't in the future, they can pull this code out of revision control
 and use it in their implementation of __PHYSFS_platformCalcUserDir().

This change let me remove __PHYSFS_platformGetUserName(), too.
2012-03-22 00:27:46 -04:00
Ryan C. Gordon b1d4afcbd3 Cleaned up some FIXMEs. 2012-03-22 00:21:28 -04:00
Ryan C. Gordon 60aa0e460c Make sure the prefdir has a final dirsep on it.
This matches the behaviour of PHYSFS_getBaseDir() and PHYSFS_getUserDir().
2012-03-21 23:59:43 -04:00
Ryan C. Gordon 43367c0c29 Fixed some brainfarts in the Windows version of PHYSFS_getPrefDir(). 2012-03-21 23:52:44 -04:00
Ryan C. Gordon 3a5a780988 PHYSFS_setSaneConfig() now uses PHYSFS_getPrefDir(). 2012-03-21 23:31:21 -04:00
Ryan C. Gordon 24d6a925d1 Added PHYSFS_getPrefDir(). 2012-03-21 23:30:50 -04:00
Ryan C. Gordon ba676b23e5 Reworked the error reporting API. Now we use error codes instead of strings.
This is based on work originally done by Christoph Nelles.
2012-03-20 15:38:12 -04:00
Ryan C. Gordon 8d7cf56651 Fixed mutex misuse. 2012-03-20 15:28:19 -04:00
Ryan C. Gordon be4ae4d401 Removed __PHYSFS_platformCvtToDependent().
It's only used by archive_dir.c, and it's trivial to make it into
 platform-independent code for that one module.

Bonus: less malloc() pressure...now every access to the native filesystem
 doesn't require a temporary heap allocation.
2012-03-15 01:58:39 -04:00
Ryan C. Gordon 03dbc3f758 Make __PHYSFS_platformDirSeparator into a single char.
This multichar thing was always stupid. Pull it out of revision control if
 you ever need it.
2012-03-15 01:54:57 -04:00
Ryan C. Gordon aaf8868c8d Removed __PHYSFS_convertToDependent(), as dead code. 2012-03-15 00:40:00 -04:00
Ryan C. Gordon 624d7f3a42 Fixed PHYSFS_read() and PHYSFS_write() in the dev branch.
PHYSFS_read() and PHYSFS_write() should return number of objects written,
 not object size.

Thanks to Evgeny Podjachev for the fix!
2012-03-14 05:47:15 -04:00
Ryan C. Gordon 611aeec61b Removed __PHYSFS_platformRealPath().
It was an ugly piece of code, didn't work on The Hurd (apparently), likely
 dangerous in its incomplete understanding of Windows paths, and only
 used in places that should have given us normalized, absolute
 paths in the first place anyhow.
2012-03-11 04:19:36 -04:00