robocop
3389149564
The files rng.cpp, rng.h, rng_proto.h, bookmark.cpp and bookmark.h belong exclusively to PC_ENGINE, that's obsolete functionality.
...
Tenth step: some code was tagged PC_ENGINE but was for SCROLLABLE_CURSORS instead, hence restore it. Also, clean unused data members, obsolete BLR verbs and the like.
2006-02-07 08:08:49 +00:00
robocop
38b063b3de
I misread the logic: it's true (not false) that tells Timestamp to not call the operating system to get the current time. Thanks to D. Sibiryakov.
2006-02-04 02:40:20 +00:00
robocop
97122ee389
Replace jrd_time.h by our timestamp.h, but I left timestamp.h's #includes commented where it *seemed* not necessary.
2006-02-03 08:39:36 +00:00
robocop
8b99699398
Do not lose time getting the accurate timestamp from the operating system if we are going to override it imediately with the user supplied value in isc_encode_date and isc_encode_timestamp. Also, this avoids an exception escaping from the API functions if the system call fails.
2006-02-02 00:03:29 +00:00
robocop
b67b0219ab
Change from getenv to readenv. This is still getenv in Linux/UNIX but GetEnvironmentVariable in Windows.
2006-01-14 04:48:59 +00:00
dimitr
cf3ed06d9c
Fixed POSIX builds.
2005-12-28 10:07:54 +00:00
robocop
a44464d73c
Do not use the harcoded c:\Program Files path while possible: it doesn't have to live in the C: partition and it's localized in non-English Windows versions.
2005-12-25 07:43:04 +00:00
robocop
9202047a0f
Solve chicken and egg problem: when we can't locate the config file, we throw an exception but if we try to trap and decode the failure in another place, it tries to know the location of the messages file (to interpret the status vector), which in turn requires to locate the config file. Therefore, we crash mysteriously.
2005-12-25 04:42:13 +00:00
robocop
82ebc3ab64
Fix the non-VMS version of gds__validate_lib_path:
...
We shouldn't modify the result from getenv().
Also, fix some null terminator issues.
2005-12-23 23:11:57 +00:00
robocop
f00fc0ff56
1.- Make "vec" type safe and get rid of several code contortions. Fix some small bugs meanwhile.
...
2.- Put inside PC_ENGINE several data members that belong to that feature.
3.- Put inside PC_ENGINE the code that does cleanup on those unused data members.
2005-12-02 07:35:34 +00:00
dimitr
29f2f43359
Fixed the bug I introduced with my GDS changes.
2005-11-29 11:36:06 +00:00
dimitr
905a2e46f8
Fixed builtin lookups for error messages and SQLCODEs.
...
IMPORTANT!!! You have to rebuild the tree from scratch!
2005-11-24 11:53:55 +00:00
dimitr
2382db581e
Reworked the BLR printer to use Firebird::string as its buffer.
2005-11-24 08:34:16 +00:00
dimitr
c236dc4a73
1. Fixed memory corruption in filters.cpp
...
2. Increased internal buffer in gds.cpp to avoid overflows
3. Made this buffer dynamically allocated to avoid stack corruptions
4. Minor cleanup
2005-11-23 20:18:19 +00:00
robocop
3d53c1b75b
Let's hope this fixes the incompatibilities between safe_interpret and the old gds_interprete.
2005-09-24 05:09:05 +00:00
alexpeshkoff
d15f6b11fc
avoid unpredictable behaviour of gds__log() on posix in signal handlers
2005-08-29 14:57:43 +00:00
skidder
9f04361ab1
Oops, committed some debugging code which was not supposed to be exposed to public :-)
2005-07-21 00:44:14 +00:00
skidder
286c14843b
Global handles need to be closed or we have the leak during client/engine DLL unload.
...
The cleanest solution would be to port FileObject from BV3 branch here and use it,
but this hack will do either.
2005-07-21 00:42:51 +00:00
skidder
699c914cbf
little debug logging improvements
2005-07-19 23:27:38 +00:00
asfernandes
354a158736
Misc.
2005-06-06 20:44:10 +00:00
asfernandes
59213efe7e
Merge INTL branch into HEAD
2005-05-27 22:45:31 +00:00
alexpeshkoff
63c5cebdcf
fixed embedded build
2005-05-16 11:08:14 +00:00
alexpeshkoff
74c88a4e97
fixed a couple of buffer overflows in gds__prefix*() functions family
2005-05-14 10:45:40 +00:00
robocop
1caceee875
Use static_cast when converting from void*, not reinterpret_cast.
2005-03-26 05:39:44 +00:00
robocop
014d90ff59
Misc.
2004-12-08 04:12:26 +00:00
robocop
af6a16107a
Misc.
2004-11-30 06:18:39 +00:00
robocop
bb170bd3b6
isc_interpret is fb_interpret again.
...
Please do a full rebuild.
2004-11-28 01:03:05 +00:00
robocop
d6535351a2
Misc: style.
2004-11-24 09:22:07 +00:00
dimitr
1859b19b9a
Added missing BLR handling to the internal blob filter
2004-11-17 18:18:44 +00:00
robocop
eb99bbcccf
Poor man's version of the old gds_alloc_report.
2004-11-17 08:56:07 +00:00
alexpeshkoff
ea9e900ab3
cleanup of methods, storing data in plain char*
2004-11-14 18:05:13 +00:00
robocop
e5be73b048
Misc: extra semicolons do not help. :-)
2004-11-07 10:47:20 +00:00
robocop
e1573bf6ce
Fix wrong comment due to copy/paste.
2004-11-06 07:17:18 +00:00
skidder
933063c7ce
Slightly adjust Dmitry's Firebird::TimeStamp class to make it more useful, change is coordiated with Dmitry. Use thread-safe localtime_r routine when it is available
2004-11-04 19:14:20 +00:00
skidder
3e26513b38
fb_interpret->isc_interpet, deprecate isc_interpete, clean interface a little bit
2004-10-25 03:53:03 +00:00
robocop
8fd795893f
Fixed some possible buffer overruns and marked others.
2004-10-09 01:47:22 +00:00
robocop
5833e93321
VMS-related changes. Just in case.
2004-10-07 09:46:03 +00:00
robocop
d6f9a7874a
Close vulnerability #6 in gds__interprete's new version.
...
In the loop, if several parameter exist, they are copied one after another in the same dynamic buffer. Therefore, for the 2nd param and more, we don't have the full buffer size to copy arguments, but only the remainder. Big buffer overflow could happen here.
If there's no buffer at all (from fix #5 ), the empty literal string is passed.
Now, fb_interpret seems ready to be used.
2004-10-04 08:37:49 +00:00
robocop
4964a1c7a1
Close vulnerability #5 in gds__interprete's new version, reported by Dmitry Sibiryakov.
...
When copying paramters to a dynamic buffer, the null terminator was written beyond the buffer's end.
Additionally, if there's no buffer at all (to be evident when fix #6 is posted), the empty literal string is passed.
2004-10-04 08:07:28 +00:00
robocop
43c46a29cd
Close vulnerability #4 in gds__interprete's new version.
...
Static array with 10 places to hold pointers to parameters (isc_arg_string, isc_arg_number & isc_arg_cstring) may overflow in the loop that fills it and doesn't check bounds.
2004-10-04 07:38:28 +00:00
robocop
e5ad0f1a7f
Misc cleanup.
2004-10-04 04:59:43 +00:00
robocop
bdf5dea379
Close vulnerability #3 in gds__interprete's new version.
...
Traditional sprintf could write past the end of buffer.
2004-10-04 04:44:52 +00:00
robocop
1da60e8b11
Fix mistake in operator until I'm ready to commit our utility snprintf.
2004-10-03 09:31:08 +00:00
robocop
16dfc501c1
Close vulnerability #2 in gds__interprete's new version.
...
Work in progress (regarding SNPRINTF wrapper, too).
2004-09-25 10:23:41 +00:00
robocop
1563cd7761
First attempt to close vulnerability reported by Daniel Urban
2004-09-22 08:56:04 +00:00
robocop
e6cb17cac6
Vulnerability reported by Daniel Urban.
...
First step: reorganize code.
Please do a full rebuild after retrieving those changes.
2004-09-22 01:59:39 +00:00
robocop
337614c51c
1) static_cast is enough from void* to another type
...
2) Style
2004-09-20 08:39:03 +00:00
robocop
9252c03bcb
Fix possible minor problem in debug version
2004-07-07 03:42:54 +00:00
skidder
f4e7ebdd3b
Implement rudimentary framework to track lifetimes of active BLOBs.
...
This solves the following issues:
1) Outrageous memory consumption when blobs are converted from strings during request processing
2) Materialization doesn't cause invalidation of BLOB IDs too early.
Original behaviour caused errors with updatable views or procedures receiving BLOB arguments.
2004-06-22 20:13:10 +00:00
alexpeshkoff
1591a54e5e
Thread cleanup:
...
1. Added macros to declare thread entrypoints
2. THD_mutex_* functions use Firebird::Mutex
3. Thread local storage use fb_tls.h
2004-06-08 13:41:08 +00:00
