Convert code like this:
if (aStr.equalsL(RTL_CONSTASCII_STRINGPARAM("rem")))
to:
if (aStr == "rem")
which compiles down to the same code.
Change-Id: I2dfa0f0a457c465948ecf720daaa45ff29d69de9
Convert code like:
if (aByteStr.equalsL(RTL_CONSTASCII_STRINGPARAM("rem")))
to:
if (aByteStr.startsWith("rem"))
Change-Id: I09e40b3fdc87d59a8176c2a5f39cc6aa5cf5a576
1cbe2313edda8a04f0fe233b4a29ef4e2485f557 "Terminating NUL at end of its buffer
is not considered part of OUString" was a thinko that cut the last character off
the last reported HighlightPortion on a line.
Change-Id: Idbe74676e85749cd93854293c6f49c7581414562
This reverts commit a79cb836b951eb2492e43aadd2ee672b9b67b914, which broke
--with-help builds again, as
strLine.copy(i->nBegin, i->nEnd-i->nBegin)
in BasicCodeTagger::tagParagraph (helpcompiler/source/BasCodeTagger.cxx) depends
on the invariant that a HighlightPortion does not extend past the source string.
Also, I see no reason for that change, as empty comments are handled just fine
already as demonstrated by the following commit...
Change-Id: I384bae9c4cf6a38d0a0a2832fa15bde82126cace
(Had to fix type of some variables holding TextPaM::GetIndex() values from
sal_uLong to sal_uInt16 to avoid -Werror,-Wsign-compare failures when comparing
those against HightlightPortion's nEnd.)
Change-Id: Ia8a0ba682ae28e86e394ee48adff3225eb8de053
Iterating over all entries of a std::map is rather slow, so add a
interval map to manage the free entries.
A first attempt to use boost::icl::interval_set for this was abandoned;
while the releaseId() function would be just 1 line, GCC 4.8.2 at least
is unhappy about boost icl headers for non-obvious reasons:
UnpackedTarball/boost/boost/icl/discrete_interval.hpp:45:225: error:
default argument for template parameter for class enclosing ‘void
boost::icl::boost_concept_check_dummy45(boost_concept_check45*)’
Change-Id: I7b767aefee57df7743dc13a694b6a61abdd536c7
to use cppu::supportsService and other pieces.
Change-Id: I16893b3d31a8055acd214ff23d01e63d38fe0826
Signed-off-by: Stephan Bergmann <sbergman@redhat.com>
...what apparently happened is:
(1) First, 2c6363eb16f50a8cfd81bc8511554af4f9e06478 "Be less pathetic" replaced
MediaDescriptor::impl_normalizeURL with removeFragment.
(2) Then, b378e754ae892a044460cfbe33ccc2e51c01f5ca "CMIS: fix file saving issue"
(plus follow-up b50d3727cf336bb536efe7bf067f484ddf864e9a "WaE: unused
function 'removeFragment'") removed calling removeFragment at all (but re-
introduced a now bogus "Parse URL! ..." comment that (1) had removed), which
(3) 34f4d2574987b272681b5843e8f5edc374f55fde "CMIS file picker: it really does
not like ID Mark" tried to revert again (but kept the bogus comment in), and
(4) odd commit 319b160320a045b1a5b302dafbc2220ee1d4d3c3 "CMIS file picker: it
really does not like ID Mark" (that looks like a merge conflict resolution
gone wrong) erroneously re-introduced MediaDescriptor::impl_normalizeURL
that (1) had removed, which caused a -Wunused-variable warning that
(5) 706c5a54f662ea58e3b3a64f189eb5120191152a "-Werror,-Wunused-variable" tried
to fix, but working under wrong assumptions.
The solution is to go effectively back to the code as it was after (1).
Change-Id: I62dfa9fa3adbea6b2cb10a509416fe3797c577b9
...I assume the whole point of 319b160320a045b1a5b302dafbc2220ee1d4d3c3 "CMIS
file picker: it really does not like ID Mark" was to actually use the result of
impl_normalizeURL.
Change-Id: Id94c7785183d96f2a2c3f08caa72af045a1212ba
