Re: Fishy assignment in editdoc (editeng) ?
For editdoc.cxx:
== op compares the individual members of the font if the impl pointer is
not equal. If all members are the same, the strange assignment makes
sure that both also point to the same internal instance of the font. If
you want to avoid that strange assignment, you would need to check in
every if statement above whether or not the new value differs from the
old value before making an assignment.
Best regards,
Malte.
On 26.05.2019 21:24, julien2412 wrote:
Hello,
Taking a look to the new reports generated by Cppcheck, I noticed
"duplicateConditionalAssign" and specifically this code:
2028 // If comparing the entire font, or if checking before each
alteration
2029 // whether the value changes, remains relatively the same thing.
2030 // So possible one MakeUniqFont more in the font, but as a
result a quicker
2031 // abortion of the query, or one must each time check bChanged.
2032 if ( rFont == aPrevFont )
2033 rFont = aPrevFont; // => The same ImpPointer for
IsSameInstance
See
https://opengrok.libreoffice.org/xref/core/editeng/source/editeng/editdoc.cxx?r=40d25921#2028
Another similar place:
xmlsecurity/source/xmlsec/nss/securityenvironment_nssimpl.cxx
800 if( xcert == nullptr ) {
801 xcert = nullptr ; <-- whereas in other parts of the
same file, we see a throw instruction
802 } else {
Any thoughts?
Julien
--
Sent from:
http://document-foundation-mail-archive.969070.n3.nabble.com/Dev-f1639786.html
___
LibreOffice mailing list
LibreOffice@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/libreoffice
___
LibreOffice mailing list
LibreOffice@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/libreoffice
Re: Fishy assignment in editdoc (editeng) ?
On 26/05/2019 21:24, julien2412 wrote:
Taking a look to the new reports generated by Cppcheck, I noticed
"duplicateConditionalAssign" and specifically this code:
2028 // If comparing the entire font, or if checking before each
alteration
2029 // whether the value changes, remains relatively the same thing.
2030 // So possible one MakeUniqFont more in the font, but as a
result a quicker
2031 // abortion of the query, or one must each time check bChanged.
2032 if ( rFont == aPrevFont )
2033 rFont = aPrevFont; // => The same ImpPointer for
IsSameInstance
See
https://opengrok.libreoffice.org/xref/core/editeng/source/editeng/editdoc.cxx?r=40d25921#2028
From a quick look, this looks plausible (if the two fonts compare
equal, but have potentially different innards, make them share a single
set of innards).
Another similar place:
xmlsecurity/source/xmlsec/nss/securityenvironment_nssimpl.cxx
800 if( xcert == nullptr ) {
801 xcert = nullptr ; <-- whereas in other parts of the
same file, we see a throw instruction
802 } else {
That was apparently nonsense,
"operator new doesn't return null anyway". Thanks for pointing it out.
___
LibreOffice mailing list
LibreOffice@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/libreoffice
Fishy assignment in editdoc (editeng) ?
Hello,
Taking a look to the new reports generated by Cppcheck, I noticed
"duplicateConditionalAssign" and specifically this code:
2028 // If comparing the entire font, or if checking before each
alteration
2029 // whether the value changes, remains relatively the same thing.
2030 // So possible one MakeUniqFont more in the font, but as a
result a quicker
2031 // abortion of the query, or one must each time check bChanged.
2032 if ( rFont == aPrevFont )
2033 rFont = aPrevFont; // => The same ImpPointer for
IsSameInstance
See
https://opengrok.libreoffice.org/xref/core/editeng/source/editeng/editdoc.cxx?r=40d25921#2028
Another similar place:
xmlsecurity/source/xmlsec/nss/securityenvironment_nssimpl.cxx
800 if( xcert == nullptr ) {
801 xcert = nullptr ; <-- whereas in other parts of the
same file, we see a throw instruction
802 } else {
Any thoughts?
Julien
--
Sent from:
http://document-foundation-mail-archive.969070.n3.nabble.com/Dev-f1639786.html
___
LibreOffice mailing list
LibreOffice@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/libreoffice
