View | Annotate | Download (18.4 KB)
Copy 3.0.x branch to trunk
Fix bugs #46615 & #46945 - Fail gracefully if store folder is not writable during update.
change the way we do chmodd'ing. I know, my implementation really sucked... good we have motivated community members who point this out. ;) Thanks to faw for providing a way better function and for discussing and also abiding to our needs. :) LEW21 should maybe credited too... he gave the inspiration without knowing it.
the chmod change i already had within the changelog (by mistake). This should further secure writable directories and created files. Installation need to be tested on different hosts.
check if folder is not empty (may happen for absolute paths). Thanks paul
#14429
Hiding everywhere little buggers, eh?
This commit was manufactured by cvs2svn to create branch 'phpBB-3_0_0'.
dumdidum... sorry. ;)
#13327
safer name for the updater downloads (even if removed after download) - reported by SHS`
#5654 - bzopen_or_bzdopen in bzlib.c appends us our binary mode in a strcat call, we don't need it for bzopen();
- compress is nicer (fixed a bug :P) - UTF-8 code is nicer (fixed a bug :P) - new CAPTCHA. Replaced the old one for size and usability issues. The old CAPTCHA will most likely be released as a separate package
bug fixes schema changes i really hope nothing went wrong
- removed db cache (might re-appear, but for now we do not see the need for it) - all changes to styles/subsilver/template are purely cosmetic (no functional changes) - cosmetics - bugfixes - add index to modules table - use modules ordering code for forums too
make sure custom profile fields are created correctly on registration (#2225)
ok, this one is rather large... the most important change: re-introduce append_sid: old style continues to work, not a performance hog as it was in 2.0.x -> structure is different
apart from this, code cleanage, bug fixing, etc.
- allow searches for multibyte characters in fulltext_native - use preg_quote safely
- Some archivers are worthless in this world
To all people having their bug status set to fixed: SF pserver CVS access is currently down, therefore the snapshots are still out of date.
- fix a bunch of bugs - <!-- $Id$ --> is no longer allowed in template (.html) files - changed layout of private message screens (folders are menu items)...
- Cleaned up some compress stuff - Replaced deprecated functions in the PostrgreSQL DBAL - Added an undefined constant during install - Oh, and we now have backups :D We currently work with all the MySQL flavors, PostgreSQL and SQLite...
- We only write to the archive ONCE for every file and folder when working with tar files - Checksum is faster, uses some precomputation to speed things up - We only needed to pad the end of the archive with 512 nulls, not 1024 :D - Added some comments regarding the other fields ;-)
- fseek -> rewind - fewer read commands, less ugly array stuff :D - much cleaner way to deal with the version flag, stored files are now differently labeled than deflated files ( we follow the zip spec better than 7-Zip :D )
Appending to the string is faster than copying it. Additionally, str_repeat is optimized for strings like "\0". This issue only came up with really, really large files to compress ( 150 MB + !)
- Tar extraction and compression has changed to be more efficent by making fewer read/write commands.
- We don't use the extra info because we can extrapolate the only thing of use (attributes) by using logic. Thus, we can skip large amounts of the file without missing any information.
- More loose definition regarding what gets to be a folder and what does not (More accurate, we lost folders before...) - Changed some fread()s to fseek()s - Much faster, single loop ( This might eat up more mem! Please report findings! )
- Tar extraction is MUCH faster, we parse the archive ONCE and create the directories on the spot instead of reading it twice and playing magic with file pointers. Also removed a potential leak.
- Zip extraction needs a small shot in the arm, files and archive must be handled a little better.
Fixes in both CAPTCHA and Compress
Compress: - The "crc bug" is not a bug, it is actually a feature (the function returns an Adler hash, not a crc hash. This is more usefull for PNG files..) and was "fixed" by using the proper function instead of munging a substring...
- file_get_contents instead of imploding file()s or fread()ing till the maximum filesize - language and style properly use compression - language now prompts user for methods - functions_compress does not need to eval() to get a hex date, instead calls pack()...
- install style and style elements - refresh template and style cache
- I forgot that we can add and remove prefixes :D - A small style fix :D
- Forgot to repair a function ;-)
This code works. All of it works. At least, it should.
Extraction of TAR files: - It now works! - If it could work before (which it couldn't) , it would create nasty issues with spaces. This is resolved - If it could work before (which it couldn't) , it would try to create folders in such a way where some (all) folders could not be created. This is resolved....
- Now creates vaild ZIP archives, "Data Descriptor" is not needed and breaks things.
- added acp_language (language pack management) - minor adjustments to cope with PHP5.x
- Documentation related changes - added resend activation email dialog - fixed issue in session code - log failed/successful admin re-authentication/login - fixed simple forum dropdown box (used in mcp and posting)
- change registration page language on-the-fly - added download function to functions_compress as well as tiny bugfixes - added local_name and author to iso.txt file
Various updates "discovered" while working on website compression function useage
Seems some compression applications don't bother with little things like standards ...
Fix up issues with .tar, .tar.gz and .tar.bz2 archives and directory structure creation
Don't write out when closing unless a file actually exists ...
Brain wanted in Room 101, Brain wanted
It's amazing what a strategically placed = can do for normal operation ...
Further tidy ups for zip compatibility ... still seem to be problems deleting entries via Powerarchiver ... works fine under Linux
Improve compatibility of zip extract method
Save some mem when decompressing .zip
Extract .zip ... damn gzuncompress
Extraction of .tar, .tar.gz and .tar.bz2
argh
concanetate rather than substr_replace
add .tar, .tar.gz and .tar.bz2 support, rejiggle source, seperate compress objects extending the base class
Ignore given list of files
Compression/decompression "libs" and a change to filelist