49540 Commits

Author SHA1 Message Date
Stephan Aßmus
3f5014ec86 PackageInstaller: Run install scripts in the correct directory
* From debugging with the Gobe and Moho installer, scripts define which
   folder to run them from. The PackageInstaller is supposed to run the script
   in that working directory. The parser seems to have the correct folder in
   "installPath" when adding the script as PackageItem, but that code is rather
   horrible. I've changed it so PackageScript items also set the path, use
   InitPath() to obtain the final working directory and set it before running
   the script.
 * Both Moho and Gobe create the Deskbar link from that script, the folder
   is rewritten in the script via ReplaceAll().
 * Correctly running the script makes a bug visible: Dynamically added files
   in the install location by these scripts are not removed when uninstalling
   the package. When re-installing a package, it is first uninstalled and this
   currently gives an error for both Moho and Gobe, since they create some
   links in these scripts which never worked before. To install again, the
   install folder needs to be deleted manually.
 * Some cleanup along the way... sorry.
hrev46882
2014-02-16 23:32:01 +01:00
Stephan Aßmus
4407020ba4 PackageInstaller: Install apps to non-packaged "apps" folder.
Previously, the target folder was always unset, which meant one needed
to select a target folder manually. When running from a read-only media,
this would still be a problem (a TODO), so that instead of failing to install
the package, the user is simply prompted for a path.
The change means that together with rewriting old paths to packaged
locations to the non-packaged equivalents, just double clicking a PKG and
hitting "Install" will now work. Tested with Moho installer and
Gobe Productive.
hrev46881
2014-02-16 12:19:16 +01:00
Stephan Aßmus
735727da48 PackageInstaller: Converted to use layout API
* Also auto-hide vertical scroll-bars when not needed.
 * Nicer margins everywhere
 * TODO: Install type description could collapse when empty or display
   a default message.
 * More refactoring and cleanup
hrev46880
2014-02-16 12:12:25 +01:00
Stephan Aßmus
a16a1437a7 PackageView: Small coding-style cleanup 2014-02-16 12:12:25 +01:00
Stephan Aßmus
193a4de178 PackageView: Refactored a lot of repetitive code...
... to create menu items.
2014-02-16 12:12:24 +01:00
Stephan Aßmus
236dd276ac PackageInstaller: Make window title properly localizeable 2014-02-16 12:12:24 +01:00
Stephan Aßmus
a1b7cb6256 PackageInstaller: Another round of cleanup. 2014-02-16 12:12:23 +01:00
Stephan Aßmus
2a7d988dc5 PackageInstaller: Cleanup PackageWindow
* Also center main package installation window on screen.
2014-02-16 12:12:23 +01:00
Axel Dörfler
ef18b77adc trim: added debug output.
* Can you give it another try, Alexander? The outcome won't change, I'm afraid,
  but maybe we'll know some more afterwards.
hrev46879
2014-02-15 20:50:57 +01:00
Ithamar R. Adema
8f4db360bc arm/atomic: round of atomic function changes
* change vint{32,64} to int{32,64} in:
  * kernel implementation
  * syscall definitions
  * libroot prototypes
* update syscall functions with new atomic prototypes.
hrev46878
2014-02-15 11:46:12 +01:00
Ithamar R. Adema
f5e19a38de board/arm: cleanup setting of ARM compiler options
The ones with ARCH extension are used for setting up the KERNEL
ones, so no need to try and set both.
Also, the verdex target was not setting the ARCH one, and therefore
never configured gcc for ARMv5.
2014-02-15 11:46:11 +01:00
Ithamar R. Adema
4b47d96a77 arm: fix compiler ARM architecture detection
For determining if to define ATOMIC{,64}_FUNCS_ARE_SYSCALLS, use
the proper macro.
2014-02-15 11:46:11 +01:00
Ithamar R. Adema
5cef6be21f arm/atomic: fixup arch_atomic.h
* Remove _inline functions, since we're not using inlines
* Use compiler barrier instead of GCC builtin
2014-02-15 11:46:11 +01:00
Ithamar R. Adema
3eb1ffdb6c kernel/arm: implement stubs for new functions
These were mostly introduced after the scheduler merge.
2014-02-15 11:46:11 +01:00
Jonathan Schleifer
2937677524 Downgrade libpcre for x86
The new version breaks git, but only once it's in the repository.
Installing it manually alongside the old one works. Rebuilding git and
uploading a new package does not work, as for the package management,
the new version is exactly equal to the old, as the port revision isn't
newer. We need to come up with a proper way to handle this.

Also removes zsh, as that requires the new libpcre.
hrev46877
2014-02-15 11:39:39 +01:00
Jonathan Schleifer
47343f61a5 Update libpcre to 8.33-1 for x86 hrev46876 2014-02-15 02:54:58 +01:00
Jonathan Schleifer
a781beade7 Add zsh for x86 hrev46875 2014-02-15 02:48:36 +01:00
John Scipione
ceafe5b48a Fix 2 shortcut bugs, one in Web+, one in S&T
It's really 2 instances of the same bug, we only want to consider a
fixed set of modifier keys ignoring the lock keys and left/right keys.
hrev46874
2014-02-14 18:51:13 -05:00
Jérôme Duval
480a333cff Update freetype package for x86, x86_gcc2 and x86_64. hrev46873 2014-02-14 22:35:02 +01:00
Jérôme Duval
ea81f96218 mesa: update x86_gcc2 packages to 7.9.2-6 hrev46872 2014-02-14 17:52:12 +01:00
Jonathan Schleifer
97ab2b9b2f AboutSystem: Add myself to current maintainers hrev46871 2014-02-14 01:26:38 +01:00
Jonathan Schleifer
be3f404ca1 Terminal: Add option to use bright instead of bold text
A saturated add was chosen instead of tint_color, as the results of
tint_color were too extreme.
hrev46870
2014-02-14 01:22:37 +01:00
Stephan Aßmus
3e2e0e63cd BHttpRequest: Improve cookie string building loop...
... to avoid some checks. Does it make the code more readable? Not
that it was hard to follow before.
hrev46869
2014-02-13 12:27:33 +01:00
Stephan Aßmus
1514eb3753 BHttpRequest: More elegant way to build cookieString hrev46868 2014-02-13 10:21:37 +01:00
Stephan Aßmus
6aeaeade6b BHttpRequest: Small fixes
* Allow BString to find a semicolon more efficiently.
 * Replace a dynamic stack allocated array with BStackOrHeapArray
   in one more place.
hrev46867
2014-02-13 10:12:55 +01:00
Adrien Destugues
3e358c1fca HttpRequest: use BStackOrHeapArray
Sometimes we get enough bytes at once from the connection to trigger a
stack overflow. Allocate memory on the heap instead.
hrev46866
2014-02-13 09:36:40 +01:00
Alexander von Gluck IV
abcbb5d617 arm/paging: Convert to new-style CPU management
* Aka, post-scheduler changes
* Luckily ARM paging code is very simular to x86 paging
hrev46865
2014-02-12 23:54:35 -06:00
Alexander von Gluck IV
35171b073d arm: Miscellaneous build fixes
* Use atomic_get_and_set for return value
* Atomics are no longer volatile
* Add missing arch_cpu_pause stub
* Move arch_cpu_idle to arch_cpu header to match
  other architectures
hrev46864
2014-02-12 23:37:15 -06:00
Alexander von Gluck IV
8018e8fa91 arm: Rework hrev46863 to use gcc built-in
* Those calls were indeed v7+ only, and our toolchain
  is v6.
2014-02-12 23:34:08 -06:00
Alexander von Gluck IV
92b2e03d0d arm: Add initial memory barrier functions
* These likely need reviewed by someone better
  at arm assembly. (#10537)
hrev46863
2014-02-12 23:11:11 -06:00
Alexander von Gluck IV
fca8407a2a radeon_hd: Fix GPIOInfo pin shortage on some cards
* Some cards have a lot of pins!
* Should fix #10536
* Add a check to give a friendy warning if we find
  more GPIO pins than what we are prepared for.
* Technically the maximum GPIO pins could be
  ATOM device max * 2 (16 for i2c and other)
  however lets leave some room for expansion.
hrev46862
2014-02-12 21:25:47 -06:00
Jérôme Duval
ed66187183 mesa: update x86 packages to 10.0.2-3 hrev46861 2014-02-12 21:59:50 +01:00
Adrien Destugues
916382a10f Style fixes, no functional changes
* Pointed out by Axel and Ingo.
hrev46860
2014-02-12 14:05:27 +01:00
Adrien Destugues
f6658d2c2c Fix one more status_t/ssize_t mixup. hrev46859 2014-02-12 09:50:34 +01:00
Jérôme Duval
0ca8f08038 Add Web Open Font Format to mime_db. hrev46858 2014-02-11 23:25:43 +01:00
Stephan Aßmus
be94359ce4 PackageItem: Fixed rewriting item paths
* The version of BString::Replace() that was used takes the offset at which
   replacing should start, not the number of replacements. So did this ever
   work? Use ReplaceFirst() instead.
 * Address the performance issues which Pawel commented on, plus some more.
 * I've left some debug output in the code (commented out), since I want to
   work on this some more. For example, I noticed that GoBe Productive puts
   files into /boot/beos/...
hrev46857
2014-02-11 23:19:12 +01:00
Stephan Aßmus
ad1c3c44bf PackageTextViewer: Fix include position 2014-02-11 23:19:11 +01:00
Stephan Aßmus
6568820634 PackageInstaller: Refactoring and cleanup
* Extraced common base class BlockingWindow from PackageTextViewer and
   PackageImagerViewer.
 * Use layouted version of PackageTextViewer.
 * Center ReadMe on screen.
 * Apply margins around the text view.
2014-02-11 23:19:11 +01:00
Stephan Aßmus
3421284044 PackageImageViewer: Re-use B_QUIT_REQUESTED 2014-02-11 23:19:10 +01:00
Stephan Aßmus
79350dde31 PackageInstaller: Yet more cleanup and code simplifications 2014-02-11 23:19:09 +01:00
Stephan Aßmus
4450737441 PackageInstaller: More cleanup and code simplifications 2014-02-11 23:19:09 +01:00
Stephan Aßmus
bdcd6afc18 PackageInstaller: Cleanup and code simplification 2014-02-11 23:19:08 +01:00
John Scipione
5b10d763d0 exfat: use diff structs 4 vol labels and filenames
... and rename the structs to reflect this.

* Volume labels contain up to 11 uint16s (11 2-byte UTF-16 characters).
* Filenames are packed into 1 to 17 structs of 15 uint16s each (for a total
  of 255 2-byte UTF-16 characters).
* Use 2 different packed structs in the exfat_entry union (same bytes, accessed
  with different structs) to access these 2 things.
* Remove a check that assumed the length returned the number of 2-byte UTF-16
  characters, i.e. the number of uint16s the string uses. It doesn't, it
  returns the number of characters contained in the string which might be 2 or
  4-bytes wide. We're doing the conversion wrong for 4-byte UTF-16 characters
  anyway, more on that later.
hrev46856
2014-02-11 13:26:55 -05:00
Jérôme Duval
675878fcfe shared kit: fixed headers dependency. hrev46855 2014-02-11 18:47:24 +01:00
Jérôme Duval
7202cfb455 build fix follow up on status_t/ssize_t changes.
* added a TODO in packagefs.
hrev46854
2014-02-11 18:26:41 +01:00
Jérôme Duval
bcccee9322 Appearance: use FT_CONFIG_OPTIONS_H instead of ftoption.h hrev46853 2014-02-11 17:53:14 +01:00
Adrien Destugues
19f3bae071 Fix more issues detected by gcc4 compiler:
More ssize_t/status_t mixup.
hrev46852
2014-02-11 17:51:22 +01:00
Adrien Destugues
6555120f3b ...and also fix a mismatched prototype
gcc2 doesn't seem to notice.
hrev46851
2014-02-11 13:26:32 +01:00
Adrien Destugues
00dfae0bce Fix build.
Sorry, renaming headers isn't detected by Jam and not everything was
rebuilt.
hrev46850
2014-02-11 12:50:27 +01:00
Adrien Destugues
c3d0dd7a5e HttpRequest: support gzip and deflate compression.
* Use the ZlibDecompressor to decompress the data
* Advertise support in accept-encoding

This should make web browsing feel even faster on wesites that support
these compresion schemes. It also fixes some websites (www.ru,
rainloop.net, ...) that serve gzipped resources even to browser not
supporting it.
hrev46849
2014-02-11 12:06:42 +01:00