Back in the day I did try to remove all posix tools from the build and
packaging process. I failed, because it was impossible to replace sed.
I also tried that and also failed but I still think that it is possible, just some
trade-offs must be made.
For example usage of sed during parser
16.01.2020 18:10, marius adrian popa wrote:
This might help with some documentation that is in html format and to generate
the md version
What the point to change the format that can be displayed out-of-box on any OS to the
format that requires preprocessing with some third-party tools to
10.01.2020 23:27, Evelyne Girard wrote:
In FB4, when adding a timestamp field it creates me timestamp_tz (32752).
normal?
Yes, unless you force FB4 to use previous versions default datatypes either
with (DataTypeCompatibility = 3.0 in Firebird.conf) or with set bind of
timestamp with time
09.01.2020 20:09, Evelyne Girard wrote:
Am I missing something ?
Yes. New binding operator described in README.set_bind.md.
--
WBR, SD.
Firebird-Devel mailing list, web interface at
https://lists.sourceforge.net/lists/listinfo/firebird-devel
09.01.2020 16:26, Paul Reeves wrote:
CMakeLists.txt is not the definitive package list. It is just for CMake,
isn't it?
Yes. But it (and also script in build_win32) are shipped in the packages and supposed
to be used to build the examples after installation AFAIU. Both of them refer dyn and
09.01.2020 13:53, Paul Reeves wrote:
So the question should whether to include dynamic gpre examples in the
packages or just remove them from the repo. As far as I know they have been
deprecated for a long while and have not been shipped in the product since
maybe even InterBase days.
On the
09.01.2020 14:37, Adriano dos Santos Fernandes wrote:
- If they want to use not-updated tools, why they need to get these
tools from fresh versions?
For this question there is a simple answer: because these utilities are
version-dependent.
--
WBR, SD.
Firebird-Devel mailing list, web
06.12.2019 20:08, Paul Reeves wrote:
But I guess we should upgrade anyway. Perhaps a more pertinent question
would be too ask whether we need to stick with v5. We haven't hacked
the code. We just use InnoSetup from a standard install so there is no
reason to stay with an older version unless v6
03.01.2020 08:22, Paul Reeves wrote:
It is useful from a visual perspective though as it is easy to spot files
that have been edited or binaries installed from different packages. I'll
take a look at the code and see if I can fix it.
No need. I already fixed it.
--
WBR, SD.
30.12.2019 17:45, Dimitry Sibiryakov wrote:
BTW, currently this piece of code isn't called at all because of mistype in
variable name.
Do we want to use this trick at all? It was handy at DOS time when files had no version
resource inside, but nowadays it is rather obsolete
Hello, All.
What version of touch is supposed to be used during Windows package building? POSIX one
doesn't support switches that I see in BuildExecutableInstall.bat.
BTW, currently this piece of code isn't called at all because of mistype in
variable name.
--
WBR, SD.
20.12.2019 21:19, Norbert Saint Georges wrote:
I try all the possibilities in numeric (38), decimal (38) it returns me, via pascal API,
always a 32762 and not a 32752
You must use the latest snapshots, not old beta release.
--
WBR, SD.
Firebird-Devel mailing list, web interface at
20.12.2019 21:08, Norbert Saint Georges wrote:
it is declared SQL_INT128 with 32752 but cannot find a way to declare it in an
FB4 table.
An idea or a link?
doc/sql.extensions/README.data_types.txt
--
WBR, SD.
Firebird-Devel mailing list, web interface at
17.12.2019 12:11, Adriano dos Santos Fernandes wrote:
So you suggest that put respect seek position (overwriting data already
written) or that it always append data?
I have no opinion about put. I guess overwriting of data is troublesome.
--
WBR, SD.
Firebird-Devel mailing list, web
16.12.2019 18:03, Adriano dos Santos Fernandes wrote:
Using internal (BLB_*) or external API, looks like garbage is read when
one creates a blob and tries to read from it (using the same handle,
before it's closed).
Should this return an error (writing from an opened existing BLOB does)?
16.12.2019 15:42, Norbert Saint Georges wrote:
I am developing an application to import Oracle DBs to Firebird 4.
Do you mean "automatic import"?
part_id is a number that has the value:203974809258041991228294191469794647457
what should i use in firebird 4 to import part_id?
For
1, 3.0.4
Reporter: Dimitry Sibiryakov
Priority: Minor
Reported by Paul Reeves. Valgring show that memory allocated by getaddrinfo()
is not freed:
==00:10:37:54.796 1607== 119 bytes in 1 blocks are definitely lost in loss
record 38 of 95
==00:10:37:54.796 1607
If an attempt to connect to a valid host failed, memory returned from getaddrinfo() is
not freed.
--
WBR, SD.
Firebird-Devel mailing list, web interface at
https://lists.sourceforge.net/lists/listinfo/firebird-devel
21.11.2019 18:09, Dmitry Yemanov wrote:
However, the same level of "quick" validation is now performed using the page number field
(stored inside the every page).
BTW, Oracle use the same check, but it has SCN written twice, in the beginning of a
page and in the end to detect partial
21.11.2019 16:45, Norman Dunbar wrote:
Apparently, checksums have not been used for some years after it was discovered that the
checksumming process consumed too much of the CPU?
It is not so AFAIK. Just probability of "teared writing" of database pages was dropped
to about zero some day.
Hello, All.
#include "./firebird/impl/types_pub.h"
What's the point of this dot in the beginning? MSVC and GCC always looking for included
files starting from location of current file.
--
WBR, SD.
Firebird-Devel mailing list, web interface at
17.11.2019 10:19, Mark Rotteveel wrote:
Don't you mean this commit instead:
Actually yes. I just copied last commit from blame log.
--
WBR, SD.
Firebird-Devel mailing list, web interface at
https://lists.sourceforge.net/lists/listinfo/firebird-devel
16.11.2019 19:18, Alex Peshkoff via Firebird-devel wrote:
What commit did that change which I can not find?
Your commit:
https://github.com/FirebirdSQL/firebird/commit/1ba1af29f0020a4c22b086ae383af22aec76840a
What branch?
master branch:
15.11.2019 17:58, Alex Peshkoff via Firebird-devel wrote:
Not - no ODS change for access check.
But it adds one more member field to blh struct defined in ods.h.
--
WBR, SD.
Firebird-Devel mailing list, web interface at
https://lists.sourceforge.net/lists/listinfo/firebird-devel
Hello, All.
Subj really change ODS of databases without changing its version?..
--
WBR, SD.
Firebird-Devel mailing list, web interface at
https://lists.sourceforge.net/lists/listinfo/firebird-devel
Hello, All.
unsigned long real_handle =
_beginthreadex(NULL, 0, THREAD_ENTRYPOINT, THREAD_ARG, CREATE_SUSPENDED,
_id);
On Windows 64 bits this code can lose top 32 bits of returned handle causing all
following code to fail completely.
IMHO, whole real_handle usage must be cleaned
: Engine
Reporter: Dimitry Sibiryakov
Priority: Minor
Adding connection-related substitutions can make creation of database-specific
configuration simplier.
For example "$database" could be used in INCLUDE directive (in for
$(database).conf) to all
07.10.2019 17:00, Paul Reeves wrote:
Both servers are using Beta1
Upgrade to snapshot. It had been fixed since then.
--
WBR, SD.
Firebird-Devel mailing list, web interface at
https://lists.sourceforge.net/lists/listinfo/firebird-devel
25.09.2019 11:52, Kjell Rilbe wrote:
Most Unicode characaters are not available on most keyboards. For
example, I have a Swedish keyboard. To enter a "Φ" (U+03A6), I will need
to find it in Windows character table app (browse, browse, browse...) or
find it on the Internet, like I did this time
25.09.2019 10:11, Kjell Rilbe wrote:
It would be a
lot more useful with an ability to specify the character codepoint
inside a string literal, and have that codepoint automatically encoded
into the string using that string's character set and encoding.
For example, the capital letter Ö with
24.09.2019 16:27, Kjell Rilbe wrote:
The built-in function ASCII_CHAR(n) seems to only accept integers 0..255
and not have any character set support whatsoever.
ASCII (which this function has in name) define only 127 symbols.
As a workaround, do string literals include some escape syntax
18.09.2019 12:11, Alex Peshkoff via Firebird-devel wrote:
For these API function to be useful, they must return localized region names according
to current OS settings. I don't think that it is doable in Firebird client API.
May be they should use current connection charset?
18.09.2019 3:13, Adriano dos Santos Fernandes wrote:
On 16/09/2019 08:21, Alex Peshkoff via Firebird-devel wrote:
On 16.09.2019 14:02, Adriano dos Santos Fernandes wrote:
There should be a way to client request additional data (tz offset, tz
region as string, string length) to be returned
16.09.2019 13:21, Alex Peshkoff via Firebird-devel wrote:
On 16.09.2019 14:02, Adriano dos Santos Fernandes wrote:
There should be a way to client request additional data (tz offset, tz
region as string, string length) to be returned together with the data.
A general solution, not a wrong
16.09.2019 12:50, Alex Peshkoff via Firebird-devel wrote:
It was very good (but left unnoticed suggestion) from Dimitry Sibiryakov: add single SQL
statement to tune automatic datatypes coercion. I agree with this suggestion and suggest
to drop all 'set bind' statements, all related DPB items
Hello, All.
What would you recommend for editing Firebird sources on Linux?
--
WBR, SD.
Firebird-Devel mailing list, web interface at
https://lists.sourceforge.net/lists/listinfo/firebird-devel
12.09.2019 14:29, Mark Rotteveel wrote:
So maybe appveyor has it installed. Then I suppose you should install it.
Looks like it. Or I will just always use VS2017 for my builds from now on.
If core developers decided that VS 2017 is the only usable compiler for Firebird 4 they
must remove
12.09.2019 13:12, Dimitry Sibiryakov wrote:
You forgot to include %FB_ROOT_DIR%/extern to include search path.
Or remove "re2/" directory from #include directive.
--
WBR, SD.
Firebird-Devel mailing list, web interface at
https://lists.sourceforge.net/lists/listinfo/firebird-devel
12.09.2019 12:33, Adriano dos Santos Fernandes wrote:
Are you using the latest master?
There were an error, fixed in the latest commit.
Yes. You forgot to include %FB_ROOT_DIR%/extern to include search path.
--
WBR, SD.
Firebird-Devel mailing list, web interface at
I also see that setenvvar.bat doesn't call vcvarsall.bat anymore which result in path
to MSBuild is not set and build failure. Was this change intended?
--
WBR, SD.
Firebird-Devel mailing list, web interface at
https://lists.sourceforge.net/lists/listinfo/firebird-devel
11.09.2019 15:48, Alex Peshkoff via Firebird-devel wrote:
if you mean getParam(n) together with getNumberOfParams() - I tend to agree.
I see no need in getNumberOfParams, but may be someone does.
I'm afraid that getting data by attribute like 'name of failed FK' is close to AI task. If we
11.09.2019 15:05, Alex Peshkoff via Firebird-devel wrote:
One more related idea - return text in firebird-allocated buffer which is later released
by explicit cliens call.
The main visible disadvantage is possible deadlock when error is OOM.
If you return not plain pointer but a smart one
11.09.2019 9:13, Alex Peshkoff via Firebird-devel wrote:
No other method - but if you have reasonable suggestion we will be happy to implement it.
Just an idea - language details are not needed.
How about an iterator on error lines? It could have methods like getGdsCode(),
getLine(),
10.09.2019 14:11, liviuslivius wrote:
It is not repleacement of current plan only another plan format.
Current one is default then tools not affected.
Yes. I meant non-Firebird tools like MS Management Studio etc.
--
WBR, SD.
Firebird-Devel mailing list, web interface at
10.09.2019 13:30, liviuslivius wrote:
But does it mean that you prefere format with attributes then? Or longer one but without
attributes at all?
I'm not a team member but as an user I would prefer format that is understood by
already existing tools.
--
WBR, SD.
Firebird-Devel
06.09.2019 23:00, Carlos H. Cantu wrote:
Sending screenshots of the isql sessions will not help, since what really matters is the
time of starting/commit the transactions.
You don't need to send screenshots. Send script. Second instance of isql can be run in
the right moment using command
02.09.2019 14:35, Adriano dos Santos Fernandes wrote:
Pools are also used for memory accounting.
valgrind and DrMemory are better on this.
It's the same. As soon someone does not chain everything being allocated
or has non-memory resources, finalizers are needed when delete by pool
is
02.09.2019 13:27, Alex Peshkoff via Firebird-devel wrote:
May be OK for future versions.
..."but now we are going to make a quick hack". Ok, ok.
Don't forget to make this hack for boolean datatype as well: those applications that
not aware of this new datatype should get it as a string.
02.09.2019 13:22, Alex Peshkoff via Firebird-devel wrote:
Delete by pool is very important performance optimization.
Yes, it is nice but has a high price demanding careful design of classes and their
interactions which is not currently the case.
--
WBR, SD.
Firebird-Devel mailing
02.09.2019 13:16, Alex Peshkoff via Firebird-devel wrote:
How to make server always use offset (not timezone code) in selected TS with TZ using
coersion?
Usual way for data coercion: application just set desired data type in metadata during
fetch.
PS: I understand that it may be
02.09.2019 13:06, Alex Peshkoff via Firebird-devel wrote:
Suggested approach makes it possible to tune server to return data in a form required by
partcular client.
And (just in case) I repeat: suggested approach duplicate data coercion functionality
and is superfluous.
--
WBR, SD.
02.09.2019 12:26, Alex Peshkoff via Firebird-devel wrote:
When first pool-enabled classes were added to firebird ~15 years ago allocators
were absolutely unable to peresent firebird pools.
Pools were implemented that time because standard memory manager was slower. Had
someone compared
02.09.2019 12:06, Alex Peshkoff via Firebird-devel wrote:
Certainly it's much, much better to define rules of object deletion at construction time
than to delete something explicitly (here and there) when object is going to be destroyed.
That's the job for smart pointers and containers.
--
31.08.2019 12:43, Mark Rotteveel wrote:
let me know where I should look so I can try to fix it myself?
I would start at places where the messages are used/printed.
--
WBR, SD.
Firebird-Devel mailing list, web interface at
https://lists.sourceforge.net/lists/listinfo/firebird-devel
30.08.2019 19:25, Mark Rotteveel wrote:
Have you never used tools or libraries that didn't support newer features of
Firebird?
Yes. I was able to workaround most of incompatibilities but at the end it was easier to
drop these libraries and use API directly.
About tools: isql supports
30.08.2019 15:57, Vlad Khorsun wrote:
You have never worked in environment where database evolving constantly and
there is
a lot of old a new apps that should work together for a some (probably long)
time.
One must be a real genius to add new tables and new fields preserving backward
30.08.2019 15:27, Alex Peshkoff via Firebird-devel wrote:
For some applications that makes sense.
Look: you have an old database without new data types and an old application that works
with it. The only way to put this old application into troubles is to _change_ type of an
old field
30.08.2019 15:10, Alex Peshkoff via Firebird-devel wrote:
why use new datatype at all?
That's a very good question.
--
WBR, SD.
Firebird-Devel mailing list, web interface at
https://lists.sourceforge.net/lists/listinfo/firebird-devel
30.08.2019 14:29, Dmitry Yemanov wrote:
Which sqltype value is provided by isc_dsql_prepare/describe_bind is irrelevant as
long as application can override it.
But 99.9% existing (at the moment) applications cannot override unknown for them data
type. So the BIND option may still be good
30.08.2019 13:56, Dmitry Yemanov wrote:
Which is usually good per se, unless performance is affected (e.g. millions of small
objects inside the pool).
Which is a bad idea per se but if these object has no destructor compiler won't
generate its call. Operator delete[] instead of individual
30.08.2019 13:48, Alex Peshkoff via Firebird-devel wrote:
Suggested solution is use of flexible format letting each client to receive data from the
server in a best fit way. To tune that format three parameters will be used - together
they make it possible to have 6 different formats - 4
27.08.2019 2:59, Adriano dos Santos Fernandes wrote:
If the main object is not explicitly destroyed, i.e., destroyed by pool,
the embedded objects leak.
Because of this embedded objects must use pooled
https://en.cppreference.com/w/cpp/named_req/Allocator
--
WBR, SD.
Firebird-Devel
26.08.2019 12:09, Vlad Khorsun wrote:
Sure, but it is not about the speed of GC. It is about space occupied by
deleted records.
This way or that I think Jiří should try it and find out if there is any
difference.
--
WBR, SD.
Firebird-Devel mailing list, web interface at
26.08.2019 11:43, Vlad Khorsun wrote:
When record is removed from data page ?
No immediately, it still leave the stub. Nevertheless GC of the stub should be faster
than full version.
--
WBR, SD.
Firebird-Devel mailing list, web interface at
26.08.2019 9:17, Vlad Khorsun wrote:
In theory, engine could remove such records before transaction ends. Not
sure
it will be easy to do.
Delete-in-place is implemented in Avalerion and Firebird 4.
--
WBR, SD.
Firebird-Devel mailing list, web interface at
20.08.2019 3:26, Adriano dos Santos Fernandes wrote:
It seems there are three ways to fix the problem:
Actually, four:
4) Don't use std::string or use it with Firebid allocator.
--
WBR, SD.
Firebird-Devel mailing list, web interface at
09.08.2019 13:19, liviuslivius wrote:
But it is as an options listed in "gbak -?"
I'm afraid that meaning of this switch could be a little different from what you
propose and it sets starting transaction for recovery from journal.
--
WBR, SD.
Firebird-Devel mailing list, web
09.08.2019 12:37, liviuslivius wrote:
i have checked this now and i see this was in Interbase:
-START(ING_TRANS) starting transaction ID for restore
I don't see such option in Interbase XE7 documentation. Could you provide a
link?
--
WBR, SD.
Firebird-Devel mailing list, web
06.08.2019 11:54, Alex Peshkoff via Firebird-devel wrote:
you have inentionally used cyrillic letter code to make sure firebird accepts
it?
It is UTF-8 code that accidentaly match code of some letter in code page 1251.
--
WBR, SD.
Firebird-Devel mailing list, web interface at
06.08.2019 11:48, Jiří Činčura wrote:
i.e. CreateFileW can certainly be used.
Use of unicode function for file open is the smallest problem on this way. Connection
string got destructed way before that.
--
WBR, SD.
Firebird-Devel mailing list, web interface at
06.08.2019 11:36, Alex Peshkoff via Firebird-devel wrote:
Why are filenames on windows converted to ANSI, not Unicode? Do not know...
Because Unicode patch was rejected and someone said "we will do it some day in our own
way". Can remember who was it?
--
WBR, SD.
Firebird-Devel
15.07.2019 12:27, Jiří Činčura wrote:
is the compression done before the encryption or after (over the wire data)?
Encrypted data is know to be uncompressable so I don't think developers made a mistake
trying to compress uncompressable.
--
WBR, SD.
Firebird-Devel mailing list, web
27.06.2019 18:38, Adriano dos Santos Fernandes wrote:
It's certainly already possible with UDR or UDF, so seems out of scope
for the system package.
Oracle has different opinion:
https://docs.oracle.com/database/121/ARPLS/d_lob.htm
It is better to have all routines in one package than
27.06.2019 18:09, Adriano dos Santos Fernandes wrote:
Here a small example of how it could be used for writing data in chunks
Isn't one function to create a blob in temporary space enough instead of all
this?
For example:
execute block returns (b blob)
as
begin
b =
26.06.2019 8:46, Simonov Denis via Firebird-devel wrote:
the DECFLOAT type is defined in the SQL:2016 standard as a separate type, so you don’t
need to remove anything. We are talking only about the implementation of NUMERIC (19+, x)
and DECIMAL (19+, x)
Ok. What's wrong with these types
21.06.2019 15:53, Alex Peshkoff via Firebird-devel wrote:
I.e. I suggest to replace high precision numeric's implementation using decfloat with
native 128-bit integer when possible and ttmath in other cases. That will make it possible
to use 128-bit integers in all cases when 64 is not enough
21.06.2019 15:53, Alex Peshkoff via Firebird-devel wrote:
Native code was compiled w/o opimization - even with -O1 loop was optimized and test
completed at once.
You always can fool optimizer changing you loop into something like this:
for (int i = 0; i < n; ++i)
{
e += (a /
24.06.2019 9:56, Alex Peshkoff via Firebird-devel wrote:
I must say I've used rather old gcc 4.8 to try and it does support int128 for x64 b ut
does not support them for x86.
Some intrinsics were broken in 5.1 but I found 5.4 and later working well.
--
WBR, SD.
Firebird-Devel mailing
21.06.2019 15:53, Alex Peshkoff via Firebird-devel wrote:
Except existing in fb4 (decfloat based) were checked native gcc's __int128 and
ttmath
BTW, did you check native GCC's decfloat as well?
--
WBR, SD.
Firebird-Devel mailing list, web interface at
21.06.2019 15:53, Alex Peshkoff via Firebird-devel wrote:
internal binary layout of 128-bit integer is the same for __int128 on x64 (on
x86 unsupported), ttmath's 128-bit class on x64 and ttmath on x86.
128 bits integers are supported on x86. At least I have no problem operate with them
14.06.2019 17:09, Mark Rotteveel wrote:
And that is not really workable if the client does want to use offset based values and for
example doesn't know the zone Firebird uses, or if Firebird uses an offset that is too large.
It doesn't matter. In Vlad's suggestion server always sends to
14.06.2019 15:17, Adriano dos Santos Fernandes wrote:
Every one stores the UTC time more the region_or_ offset.
Nobody is talking about *store* here.
The displayable value is calculated only when one wants to display the value.
Yes, and this is the moment when _server_ executes a
13.06.2019 16:45, Adriano dos Santos Fernandes wrote:
Considering the reality that we don't have SQL parser in the client for
tools usage, and that generic tools (ISQL and like) will be the clients
for the feature, no, they can't issue a server call to decode each
column its going to display.
13.06.2019 16:02, Adriano dos Santos Fernandes wrote:
So one (client) could request "give-me that timestamp-tz (utc timestamp
+ region/offset) columns and additionally give-me its string
representation or offset or whatever".
System table or package suits that. If user application needs that
12.06.2019 17:27, Adriano dos Santos Fernandes wrote:
What a complete misunderstood of things!
First of all character sets are fixed per table/column. User can
insert/update data with different character set, but in the end it
doesn't matter, data is stored in the table/column character set. TZ
12.06.2019 13:43, Adriano dos Santos Fernandes wrote:
I mean another function with the sole purpose of convert from/to
ID/string, i.e., nothing within the timestamp.
Not needed if server provides a table/package/UDR for it.
The only way would be to remove the functionality in the client,
12.06.2019 13:43, Adriano dos Santos Fernandes wrote:
Then we will not have a single source of truth anymore.
When user will create a client timestamp-tz value, he will need to fill
that fields too, manually (with kill the current functionality) or
automatically (making difficult what is
11.06.2019 23:46, Vlad Khorsun wrote:
I already offer to deliver to client both ID and displacement (offset,
bias)...
It costs nothing as current 10 bytes in wire protocol anyway padded up to 12
bytes.
In this case I second your suggestion.
PS: IMHO, there is nothing bad to require a
11.06.2019 21:52, Vlad Khorsun wrote:
I want to eliminate needs to translate something at client side.
The simplest way would be to drop IDs and deliver to client only bias. It is enough for
all usages of timestamp with timezone I can think about.
--
WBR, SD.
Firebird-Devel mailing
11.06.2019 16:41, Paul Reeves wrote:
It seems to me that the implementation of CURRENT_TIME/STAMP is the
wrong way round. It should support the old behaviour.
It does, but you must put "execute statement 'set time zone bind legacy'" into ON
CONNECT trigger in every database.
--
WBR,
11.06.2019 16:22, liviuslivius wrote:
But why is that boom?
Why not show e.g.
11.06.2019 16:08 (GMT+01:00)?
a) Because server sends "timezone ID" instead of bias.
b) Because even if server sends bias, isql calls decodeTimestampTz() in any
case.
--
WBR, SD.
Firebird-Devel mailing list,
11.06.2019 15:41, liviuslivius wrote:
Can i ask what is the problem? When i retrive data by client lib it try to convert it to
local timezone or what?
Copy isql + fbclient.dll somewhere, connect to any database, select current_timestamp
from rdb$database - BOOM!
--
WBR, SD.
11.06.2019 11:44, Alex Peshkoff via Firebird-devel wrote:
If different ICUs have different ID for same TZ than:
- ID should be extended with ICU version,
- suggested by Vlad table should have ICU version in the header. Yes, client will store >1
copies of that table if it connects to different
30.05.2019 18:01, Alex Peshkoff via Firebird-devel wrote:
Not necessary. Provider is changed in databases.conf for particular database.
Yes, but auth module has usage of current version of provider hardcoded.
--
WBR, SD.
Firebird-Devel mailing list, web interface at
30.05.2019 17:09, Emil Totev wrote:
SecurityDatabase = $(dir_secDb)/security4.fdb
You must use security3.fdb with ODS 12 in this case.
--
WBR, SD.
Firebird-Devel mailing list, web interface at
https://lists.sourceforge.net/lists/listinfo/firebird-devel
25.05.2019 15:57, Adriano dos Santos Fernandes wrote:
As I understand, backtrack happens due to reduce/reduce conflicts and we
have 17 of then in the rules.
Accepting of PR #189 could reduce the count to 11. Most likely the technique applied
there can be used for the rest as well.
--
20.05.2019 18:00, Adriano dos Santos Fernandes wrote:
Comments, advices?
http://tracker.firebirdsql.org/browse/CORE-4579
--
WBR, SD.
Firebird-Devel mailing list, web interface at
https://lists.sourceforge.net/lists/listinfo/firebird-devel
06.05.2019 14:59, Massimo Fazzolari wrote:
I use this udfmalloc, on windows is defined as:
function udfmalloc(Size: Integer): Pointer; cdecl; external 'ib_util.dll' name
'ib_util_malloc';
Do you check returned value for nil?
BTW, "Integer" is not always right declaration for the
06.05.2019 12:08, Massimo Fazzolari wrote:
There is no code changes to my UDF source code between the version compiled with FPC 2.6.3
(which works on both linux and windows) and FPC 3.0.4 (which doesn't work on Windows, but
works on Linux).
Some bugs (such as wrong calling convention usage)
28.04.2019 14:48, Mark Rotteveel wrote:
For support of clients that don't support the Firebird 4 types, it would be helpful if the
decfloat bind setting also governs the extended precision numerics, or if a similar
setting (both statement + dpb) is created. Maybe 'decfixed bind', but that might
601 - 700 of 2579 matches
Mail list logo
