8
0
mirror of https://github.com/FirebirdSQL/firebird.git synced 2025-01-27 05:23:02 +01:00
Commit Graph

179 Commits

Author SHA1 Message Date
dimitr
7cc12e8489 1) Reworked memory management in SORT. Now we're pools aware there.
2) Reworked error handling in SORT and its callers.
3) Replaced DLS+SortMem with the generic TempSpace class.
4) Replaced GDS temp file management with the TempFile class.
5) Various cleanup.
2006-05-31 08:53:00 +00:00
alexpeshkoff
b056a118b5 1. Avoid use of STL
2. Reworked exceptions to make status_exception always hold correct status vector
3. Avoid use of RTTI
2006-05-19 15:17:02 +00:00
robocop
ebe0a75026 Simplification. 2006-05-04 00:02:58 +00:00
aafemt
46c3ee5090 Fix for my previous commit. Thanks to Adriano 2006-04-28 04:12:09 +00:00
aafemt
f5838a4262 Not elegant fix for treating too big statements as an error 2006-04-27 10:57:11 +00:00
robocop
a90a72232c Adjust isql to the full words stated in the IB5 docs for the command line options. 2006-04-08 00:03:50 +00:00
alexpeshkoff
543932292f fixed regression in switch 2006-04-06 14:43:01 +00:00
robocop
c5f2284b15 Style. 2006-04-06 08:18:53 +00:00
mkubecek
4a94466a5a Fix of isql command line parser: check for repeated switches or database name. 2006-03-28 04:31:35 +00:00
robocop
2484232f23 - Misc.
- Style
- Avoiding one indentation level.
2006-03-26 05:58:29 +00:00
asfernandes
3329d9fc07 Correction 2006-03-26 03:34:14 +00:00
mkubecek
332dce9203 Fix of isql command line parser: allow only abbreviations of standard switch names, check integer arguments to be integers. 2006-03-25 23:41:36 +00:00
robocop
3ed50872b2 Give isql a more useful help when receiving incorrect command line arguments, in sync with other utilities. Courtesy of Michal Kubecek.
Example: isql -?
2006-03-16 09:07:37 +00:00
hvlad
73dcebc962 Replace "../jrd/jrd_time.h" with "../common/classes/timestamp.h" 2006-02-03 13:13:27 +00:00
alexpeshkoff
32ef4f5b67 fixed compilation after changes in exception 2006-01-07 16:58:23 +00:00
robocop
f902dd1311 Parse the server version output (that can be large with redirection) correctly when isql -z is used. 2005-12-30 06:57:06 +00:00
alexpeshkoff
9fa49f5077 make server-version output a bit more readable 2005-12-29 09:32:41 +00:00
alexpeshkoff
4447d705c8 fixed bugs in server version reporting 2005-12-29 08:06:11 +00:00
alexpeshkoff
2d259e3978 Fixed buffer overrun in case of bad data returned by isc_database_info() 2005-12-26 16:10:59 +00:00
robocop
39b9a7fb22 Avoid the intermediate buffer. 2005-12-03 07:01:16 +00:00
alexpeshkoff
62a24d46e4 1. Moved data structures, written to database file, into ods.h.
2. Added new types of clumplets to ClumpletReader and ClupmletWriter.
3. Applied modified ClumpletReader and ClupmletWriter to /remote and partially /jrd.
4. Added active protection from remote brute-force attack to services API.
2005-11-27 20:53:09 +00:00
robocop
e7fa075880 Let isql yell always, irregardless of the switches. 2005-11-08 23:57:23 +00:00
robocop
c6e7a41878 Fix display alignment problem with CHAR type and small binary fields. 2005-10-08 07:02:54 +00:00
robocop
4337f38101 Update isql accordingly to my previous Fix SF Bug #1292007
Move more hardcoded error strings to the msg database.
2005-10-06 07:02:04 +00:00
robocop
d7f1979164 Fix premature EOF while reading a script file when using the SHELL command inside it, found by Paul Reeves. Trying to find a compromise between MS requirements for system() call and its bad side effects on our isql. 2005-09-03 05:54:27 +00:00
robocop
0285583c9f Fix display bug found by Adriano, caused by missing %s and try to make isql output more consistent. QA Tools that check for strict result may complain. 2005-07-08 04:10:27 +00:00
asfernandes
8fe0408bb3 Make SQLDA_DISPLAY display input SQLDA parameters of INSERTs, UPDATEs and DELETEs 2005-06-23 15:12:53 +00:00
robocop
2870acfc87 Fix blob dumping routines in jrd and report errors in isql. 2005-06-16 05:13:35 +00:00
robocop
545cb77032 If an error happens when we are trying to retrieve the first record, we need to print a newline or the error msg will appear padding the column headers. 2005-06-12 23:47:04 +00:00
robocop
f3593f8977 Cater for the longest possible identifiers, quoted. 2005-06-12 01:36:03 +00:00
robocop
c547304496 - Renamed function input to newinput to be in sync with newoutput.
- Implemented PASSTHROUGH command only for the debug build. Use with care.
2005-06-11 00:32:06 +00:00
robocop
35dc1e343c - Cleanup.
- A bit better handling of SHOW EXCEPTION.
2005-05-26 06:45:25 +00:00
robocop
f6d654cf13 Lower the clutter. 2005-05-24 04:42:01 +00:00
robocop
1ac3a5d9b7 - Implement the SHOW COMMENT(S) command.
- Integrate it into metadata extraction, too.
- A couple of minor improvements to isql.
See explanation in show_comments() function.
2005-05-19 08:03:10 +00:00
kkuznetsov
a066a949fb Renamed ERR to ps_ERR in the enum processing_state {...} to fix compile error on Solaris 2005-05-14 14:50:41 +00:00
robocop
df8b6d12e4 Apply Miroslav Penchev fix for ISQL_disconnect_database overwriting the Quiet flag permanently. 2005-05-12 07:00:54 +00:00
robocop
3c361f66ba Simplify array bounds printing for canonic case. 2005-05-12 05:32:25 +00:00
robocop
3fd1859bd1 Workaround for SF Bug #450404. (Isql can't guess what the user intended.) 2005-05-02 07:11:39 +00:00
robocop
3f68b85b89 Another fix for the -b (Bail On Error) option when SQL commands are issued and no db connection exists yet.
Apply Miroslav Penchev's patch for bug with -Q always returning 1 to the operating system, discovered by Ivan Prenosil.
2005-04-28 06:26:24 +00:00
robocop
b2010a42e4 Fix indentation. 2005-04-25 05:23:57 +00:00
robocop
3379bed424 Isql shouldn't try to connect twice when extracting metadata, since the second attempt may produce a confusing message (the db name has been made null at this time). 2005-04-04 08:31:11 +00:00
robocop
4fbf25e341 Fix chaotic output. 2005-04-03 06:58:40 +00:00
robocop
7acd0a6d39 - Temporary fix for Pavel's firebird-Field-test-1173948.
- Some little tweaks.
2005-04-02 07:07:07 +00:00
robocop
f800fd59a7 I find hard to work with functions that span 800 or more lines. Therefore, I modularized process_statement. I took the opportunity to get rid of some local variables and fix performance info printing. 2005-04-02 03:51:43 +00:00
robocop
c5c02ed2e9 1.- Replacing raw call to gds__msg_format by ISQL_msg_get; the routine was there for years and almost unused.
2.- UDFs are still created by sys_flag being NULL, so show_functions() should acknowledge that fact.
3.- Implement naive code to show an approximate script-related line number when a failure happens. (Unlike my private FB1's isql, it doesn't rely on strings subject to localization.)
4.- With SET PLANONLY ON, PLAN was activated, but with SET PLAN OFF, PLANONLY wasn't deactivated.
5.- Include isc_info_sql_stmt_get_segment among statements that are like SELECT.
6.- Solve a bad interaction between PLANONLY and SQLDA_DISPLAY, where the first being activated didn't allow to see the output XSQLDA.
7.- More error detection.
2005-03-31 07:50:32 +00:00
robocop
29690221e1 1.- Some small buffers with predefined lengths don't need to be allocated dynamically.
2.- ISQL_get_index_segments shouldn't overflow in extreme cases. Better print an indication that the extraction wasn't completed (ellipsis).
3.- Fix bug with EXTRACT_list table that was printing ); if it didn't find the required table. This in turn caused bugs in other functions.
4.- STDERROUT is used in all places with second argument set to true. Simplify it.
5.- Delete the code previously commented for WAL.
6.- Fix add_row and copy_table. These features were requested to be made available in release mode at v1 time by Paul Beach. They should now be able to cope with quoted identifiers, too.
7.- Complement usage of readNextInputLine when called directly to avoid reading garbage or causing memory errors.
8.- Fix several strings that should have been declared with MAXPATHLEN since the first moment.
9.- Misc bug fixes and B.O. closed.
2005-03-26 05:33:55 +00:00
aafemt
ad88d1e805 Fix for Field-test item #1168910 2005-03-23 10:10:38 +00:00
skidder
56186e7f4d Calling clear_history like that breaks the builds. But even when build is fixed, call like that causes SIGSEGV due to editline history not being initalized when ISQL is used in non-interactive mode. Since clear_history doesn't free all the editline memory there is no big point in using it here anyways... 2005-03-09 06:53:03 +00:00
robocop
ea866740f1 1.- Stop possible crash with long terminators. Pending other checks for the future.
2.- Complete Damyan Ivanov's patch to allow BAIL detect really most error conditions.
3.- Create two functions to handle part of the bulky processing of isql's built-in commands.
4.- Make SET SQLDA_DISPLAY work in release versions and derive functionality to two new functions.
5.- Change Borland's moronic usage of a FIFO implemented in a linked list to handle the needs of a LIFO, simplying code and avoiding the associated logic pain.
6.- Avoid several SQL> prompts when using the INPUT command interactively.
7.- Free several resources on exit. Stop some additional memory leaks.
8.- Signal we are processing the INPUT command only when the file can be opened.
9.- Put the cursor declaration inside the code protected for scrollable cursors, in case it's needed there, since isql doesn't do positioned updates.

Remember to report random behavior, crashes, segmentation faults, core dumps and BSOD's.
2005-03-04 09:17:34 +00:00
robocop
4be87294fa sqldaid and sqldabc are unused crap borrowed from IBM 15 years ago.
Don't show them as they contain garbage.
2005-02-07 09:00:33 +00:00