The GoSub flag (0x8000) was cleared before being tested, so PushGosub
never returned to the following line when returning using a Return
statement.
Change-Id: Ic1e6977d48a34c52ed9c24c5be0d9acfa3eb7552
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/210395
Reviewed-by: Andreas Heinisch <andreas.heinisch@yahoo.de>
Tested-by: Jenkins
Pass over an omitted argument as a missing parameter instead of a
numeric zero, so that optional parameters, their default values and the
basic function IsMissing() behave accordingly.
Change-Id: I7c4a65dfb1791f8aa900cc9eddd3007422dc41a0
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/209457
Tested-by: Jenkins
Reviewed-by: Andreas Heinisch <andreas.heinisch@yahoo.de>
When calling a Basic procedure using a named argument (for example,
foo(y:=10)), optional skipped parameters were not correctly marked as
missing, causing the function IsMissing() to return incorrect results.
Change-Id: I90a9910165ec00a24deba86f7b3397dbc0720c93
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/205034
Tested-by: Jenkins
Reviewed-by: Andreas Heinisch <andreas.heinisch@yahoo.de>
I added the code that I remove here in commit
22d7827c66 (tdf#168569: support date
values in string pool, 2025-09-26), but this is section to handle
old images, after commit 3bcfb1aac1
(tdf#143707 - Change strategy to support suffix type characters,
2021-08-10) the suffixes are written in a way that is handled in
SbiImage::GetString.
Change-Id: Ia2ad05f1e146dc40686f2baf07f178331f1ff4a7
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/191561
Reviewed-by: Mike Kaganski <mike.kaganski@collabora.com>
Tested-by: Jenkins
Commit f45463d8e2 (tdf#93727 Support date
literals in basic, 2017-05-24) introduced correct parsing of literals
like #2025-09-26#. However, it didn't retain the value type; there was
a discussion about that in gerrit, but no solution was found, and type
was set to double.
Later, a similar problem (storing type of value in compiled image) was
fixed in commit 5eedb3beea (tdf#142460:
properly handle boolean values in string pool, 2021-06-25).
This change reuses the same method to store date type using 'd' char in
the string pool.
Change-Id: I32e8497ece1f30980ba6d4fca248687b817348f2
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/191555
Tested-by: Jenkins
Reviewed-by: Mike Kaganski <mike.kaganski@collabora.com>
SbiRuntime::FindElement ends with a call to CheckArray(). The latter
takes a variable, checks if that is an array and if it has parameters;
and if both are true, tries to obtain an element of the array defined
by the parameters as array indexes (using SbxDimArray::Get).
The problem was, that the variable may actually be array; it may have
parameters; but this array can be the result of a function call; the
variable may be SbxMethod; and the parameters may be for the function,
not array indexes.
Specifically, the call like
Iif(True, Array("A","B"), Array("B","A"))
searched for an entity named Iif; found that it was a RTL method;
prepared the arguments for it (consisting of four elements: 0 - the
method itself; 1 - boolean True; 2 - array ["A", "B"]; 3 - array
["B", "A"]); the function executed, putting the correct value to the
params[0], which was the method itself (so its value became an array);
and finally, the SbxMethod variable with the array value got passed
to CheckArray. The latter found out that it's an array, and that it
has parameters (those four) - and tried to get the element, which
index was defined by the parameters.
Here comes the specifics of LibreOffice Basic's argument handling.
SbxDimArray::Get only cares about first arguments expected by the
array dimensions; it ignores the rest. Since the array was single-
dimensional, it only considered parameter 1 (parameter 0 is, by
convention, reserved for return value). That was a boolean. When it
was False, its GetInteger returned 0; and finally, CheckArray took
element 0 from the array, returning a scalar value, as if it was
Iif(False, Array("A","B"), Array("B","A"))(0)
But when the boolean was True, another Basic specific took place:
its numerical value is -1; and trying to get element -1 from the
array resulted in "Index out of defined range".
This change disambiguates between the cases of array index, and
function parameters, checking if the variable itself, and its 0th
parameter are both SbxMethods. Then it doesn't try to treat the
parameters as indexes into the array.
Change-Id: I597f94a7e688c8bba12d87a1bd5871491c25cac0
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/187145
Tested-by: Jenkins
Reviewed-by: Mike Kaganski <mike.kaganski@collabora.com>
As required by [MS-VBAL], and actual VBA implementation.
It is a question if we also need to change it likewise in other
modes (normal, compatible) - see tdf#150460.
Change-Id: I1e83d57fe4be3bf264a64fe977909cf7684bc798
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/172005
Tested-by: Jenkins
Reviewed-by: Mike Kaganski <mike.kaganski@collabora.com>
Makes them more comprehensible; uses semantic names instead of magic
numbers. Added comments to the constants that are unclear.
Change-Id: Ifcb8a0d93ce057b8af08859b87ad3af8c67479e5
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/171443
Reviewed-by: Mike Kaganski <mike.kaganski@collabora.com>
Tested-by: Jenkins
Don't execute On expression GoSub Statement; On expression GoTo
Statement if the expression lies out of range.
Change-Id: I5c1de25918b5e812d7ec82034f8d56351374d56a
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/165960
Reviewed-by: Andreas Heinisch <andreas.heinisch@yahoo.de>
Tested-by: Jenkins
1. The regex must match the whole input. Thus, the ^ and $ metacharacters
must be replaced with \A and \z, which only match beginning and end of
input, not any line start / end.
2. The * and ? metacharacters of LIKE must match newline character; thus,
search flags must include UREGEX_DOTALL. To avoid changing TextSearch
implementation, I use icu::RegexMatcher directly.
The direct use of icu::RegexMatcher also allowed to simplify the code by
calling icu::RegexMatcher::matches method. This may perform better than
general-purpose utl::TextSearch::SearchForward.
Change-Id: I75776498b36f236da294462362ed5b36ed8cdf68
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/165700
Tested-by: Jenkins
Reviewed-by: Mike Kaganski <mike.kaganski@collabora.com>
since:
commit 9cc8457abc
Date: Fri Dec 15 23:09:19 2023 -0700
tdf#128122 Updated BASIC CCur to reuse SvNumberFormatter
LibreOfficeOssFuzz.conf has --enable-ld=gold and that will just
crash on an unresolved symbol without any explanation, but building
without will give something useful, e.g.
sbxcurr.cxx:(.text._Z14ImpGetCurrencyPK9SbxValues[_Z14ImpGetCurrencyPK9SbxValues]+0x89f): undefined reference to `SbiInstance::GetNumberFormatter()'
/usr/bin/ld: sbxcurr.cxx:(.text._Z14ImpGetCurrencyPK9SbxValues[_Z14ImpGetCurrencyPK9SbxValues]+0xbf7): undefined reference to `SbiInstance::PrepareNumberFormatter(unsigned int&, unsigned int&, unsigned int&, o3tl::strong_int<unsigned short, LanguageTypeTag> const*, DateOrder const*)'
Change-Id: I4b555760a609ec04fd85f7ba9c198a3d2ac712c5
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/161294
Tested-by: Jenkins
Reviewed-by: Caolán McNamara <caolan.mcnamara@collabora.com>
Commit 5760c94b88 tried to avoid all cases
which could set an error in SbiRuntime::PushForEach. To do that, it checked
the type of xObjVar before trying to get an object from it, which otherwise
could set an error.
But the type of the contained value can be not known until it is retrieved
(which can happen inside SbxValue::Get in a call to SbxValue::Broadcast with
SfxHintId::BasicDataWanted). This happens e.g. when the container passed to
'for each' is a call to some special function, like VBA's 'Selection'. Then
SbxValue::GetFullType would return SbxEMPTY prior to SbxValue::Get.
Let's make sure to call SbxValue::Get first (asking for a Variant, to avoid
errors on type mismatch), and only then, check the actual result data type.
Change-Id: Iaa697f38285505e50504ae09f9307fbd29e09a53
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/147273
Tested-by: Jenkins
Reviewed-by: Mike Kaganski <mike.kaganski@collabora.com>
Setup the "last reschedule time" counter at the first pass,
to avoid useless immediate reschedule.
Update the counter when running "when blocked" reschedules.
This seems to avoid the problem with the bugdoc.
Change-Id: Ib5958a1a2b048f5ec654c69ee9e977e8a26de6f5
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/146215
Tested-by: Mike Kaganski <mike.kaganski@collabora.com>
Reviewed-by: Mike Kaganski <mike.kaganski@collabora.com>
A leftover from commit f6e0527033
Date Mon Mar 21 12:55:40 2011 +0100
Remove most of dead code inlibs-core (basic)
which dropped the related commented out code for "not implemented"
stub, being there since c25ec0608a
"initial import".
Change-Id: Ic09a0c63921f1c5daa277fe7bb6dfc194895b673
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/139978
Tested-by: Jenkins
Reviewed-by: Mike Kaganski <mike.kaganski@collabora.com>
...so that its TOOLS_WARN_EXCEPTION can be used in
comphelper/source/misc/logging.cxx in a follow-up commit. (And while at it,
rename from diangose_ex.h to the more appropriate diagnose_ex.hxx. The
comphelper module is sufficiently low-level for this immediate use case, so use
that at least for now; o3tl might be even more suitable but doesn't have a
Library until now. Also, for the immediate use case it would have sufficed to
only break DbgGetCaughtException, exceptionToString, TOOLS_WARN_EXCEPTION,
TOOLS_WARN_EXCEPTION_IF, and TOOLS_INFO_EXCEPTION out of
include/tools/diagnose_ex.h into an additional new
include/comphelper/diagnose_ex.hxx, but its probably easier overall to just move
the complete include file as is.)
Change-Id: I9f3222d4ccf1a9ac29d7eb9ba1530d53e2affaee
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/138451
Tested-by: Jenkins
Reviewed-by: Stephan Bergmann <sbergman@redhat.com>
Andreas Heinisch