[Libreoffice-commits] core.git: sax/source

[Eike Rathke (via logerrit)]

 sax/source/tools/fastserializer.cxx |   16 ++++++++++++++++
 sax/source/tools/fastserializer.hxx |    3 +++
 2 files changed, 19 insertions(+)

New commits:
commit 63c4e82c910422f928c8e0bcbb82b6c43d4d5557
Author:     Eike Rathke <erack at redhat.com>
AuthorDate: Tue May 14 16:32:12 2019 +0200
Commit:     Eike Rathke <erack at redhat.com>
CommitDate: Thu May 30 19:21:33 2019 +0200

    Reintroduce mbXescape that got overaggressively removed
    
    with
    
        commit d4d37662b090cb237585156a47cd8e1f1cbe2656
        CommitDate: Fri Oct 12 12:46:45 2018 +0200
    
            loplugin:constfields in reportdesign,sal,sax
    
    which killed also the bGood checking and later because that wasn't
    used anymore also the SAL_WARN_IF() was eliminated with
    
        commit 41eeaace84b45c803fff3ebd5ab981f0ad09393b
        CommitDate: Mon Oct 29 15:12:26 2018 +0100
    
            loplugin:oncevar
    
    The mbXescape variable will be used, as the TODO said..
    
    Change-Id: I11091379c27cf0222677595eb723b61ad3ebe4e8
    Reviewed-on: https://gerrit.libreoffice.org/72302
    Tested-by: Jenkins
    Reviewed-by: Eike Rathke <erack at redhat.com>

diff --git a/sax/source/tools/fastserializer.cxx b/sax/source/tools/fastserializer.cxx
index 4a4091438a40..1a422ad8239a 100644
--- a/sax/source/tools/fastserializer.cxx
+++ b/sax/source/tools/fastserializer.cxx
@@ -59,6 +59,7 @@ namespace sax_fastparser {
         , mbMarkStackEmpty(true)
         , mpDoubleStr(nullptr)
         , mnDoubleStrCapacity(RTL_STR_MAX_VALUEOFDOUBLE)
+        , mbXescape(true)
     {
         rtl_string_new_WithLength(&mpDoubleStr, mnDoubleStrCapacity);
         mxFastTokenHandler = css::xml::sax::FastTokenHandler::create(
@@ -135,6 +136,7 @@ namespace sax_fastparser {
             return;
         }
 
+        bool bGood = true;
         const sal_Int32 kXescapeLen = 7;
         char bufXescape[kXescapeLen+1];
         sal_Int32 nNextXescape = 0;
@@ -195,6 +197,7 @@ namespace sax_fastparser {
                             }
                 break;
                 default:
+                            if (mbXescape)
                             {
                                 // Escape characters not valid in XML 1.0 as
                                 // _xHHHH_. A literal "_xHHHH_" has to be
@@ -239,10 +242,23 @@ namespace sax_fastparser {
                                  * scanning for both encoded sequences and
                                  * write as _xHHHH_? */
                             }
+#if OSL_DEBUG_LEVEL > 0
+                            else
+                            {
+                                if (bGood && invalidChar(pStr[i]))
+                                {
+                                    bGood = false;
+                                    // The SAL_WARN() for the single character is
+                                    // issued in writeBytes(), just gather for the
+                                    // SAL_WARN_IF() below.
+                                }
+                            }
+#endif
                             writeBytes( &c, 1 );
                 break;
             }
         }
+        SAL_WARN_IF( !bGood && nLen > 1, "sax", "in '" << OString(pStr,std::min<sal_Int32>(nLen,42)) << "'");
     }
 
     void FastSaxSerializer::endDocument()
diff --git a/sax/source/tools/fastserializer.hxx b/sax/source/tools/fastserializer.hxx
index b00ccf0923df..e3aa17a78f91 100644
--- a/sax/source/tools/fastserializer.hxx
+++ b/sax/source/tools/fastserializer.hxx
@@ -228,6 +228,9 @@ private:
     rtl_String *mpDoubleStr;
     sal_Int32 mnDoubleStrCapacity;
     TokenValueList maTokenValues;
+    bool mbXescape;     ///< whether to escape invalid XML characters as _xHHHH_ in write(const char*,sal_Int32,true)
+                        /* TODO: make that configurable from the outside for
+                         * some specific cases? */
 
 #ifdef DBG_UTIL
     std::stack<sal_Int32> m_DebugStartedElements;