| Age | Commit message (Collapse) | Author |
|---|
|
While we're at it, avoid an indirection through a pointer, and
use a name that isn't reserved for the compiler (leading _ CAP).
This is a small portion of the associated bug:
BUG=https://code.google.com/p/pdfium/issues/detail?id=112
R=brucedawson@chromium.org
Review URL: https://codereview.chromium.org/880663003
|
|
A suitably corrupted file can cause the parser(s) to repeatedly re-read
sections of the file at increasing parser recursion depth until the
stack is exhausted. There is supposed to be a check for this based upon
the parser "level", but not all call paths pass or update the level as
required.
Much as I hate per-class statics, this introduces one to track the depth
so that the check is enforced no matter how screwy the call path might be
that leads the parser to re-enter itself. This is more palatable than trying
to find all these paths and fix them. We know this is OK since there is
only one thread in here modifying the static.
BUG=451830
R=thestig@chromium.org
Review URL: https://codereview.chromium.org/875263002
|
|
Currently, no callers go through the Interface, which makes having a
separate interface class kind of pointless. After converting callers
away from using the CPDF_DataAvail concrete class, it can be moved
from the header to the .cpp file.
R=bo_xu@foxitsoftware.com
Review URL: https://codereview.chromium.org/873523002
|
|
This is done by explicitly adding a virtual dtor to interface classes,
since the cost is small given that there are already virtual functions.
The exceptions are for classes that have a Release() or Delete() method,
in which case it is non-virtual and protected to indicate that the virtual
class is never the deletion point.
BUG=
R=brucedawson@chromium.org, thestig@chromium.org
Review URL: https://codereview.chromium.org/810883005
|
|
BUG=406806
R=tsepez@chromium.org
Review URL: https://codereview.chromium.org/503883002
|
|
(patchset #1 of https://codereview.chromium.org/493163003/)
Reason for revert:
Needs to address comments before landing
Original issue's description:
> Use number of components from ICC profile and alternate color space
>
> BUG=406806
>
> Committed: https://pdfium.googlesource.com/pdfium/+/be83103
TBR=tsepez@chromium.org,jun_fang@foxitsoftware.com
NOTREECHECKS=true
NOTRY=true
BUG=406806
Review URL: https://codereview.chromium.org/504883003
|
|
BUG=406806
Review URL: https://codereview.chromium.org/493163003
|
|
BUG=387983
R=tsepez@chromium.org
Review URL: https://codereview.chromium.org/454283002
|
|
Should there be cases where this fails to compile, it indicates a mistake,
either an incorrectly declared overrriden virtual method, or a method that
should be declared non-virtual.
The only issues were with CPDF_CustomAccess::GetBlock(), CPDF_CustomAccess::GetByte(),
and CPDF_CustomAccess::GetFullPath(). These don't appear to be used anywhere,
and are removed. Two members are removed that are no longer needed once those
methods are removed.
R=jam@chromium.org, jun_fang@foxitsoftware.com
Review URL: https://codereview.chromium.org/454983003
|
|
Calling `delete` on an object of a type that has virtual functions but
not a virtual destructor is questionable: Since the object has virtual functions,
it likely has subclasses, so if it's deleted through the base pointer and the
destructor isn't virtual, the subclass destructor won't be called.
In most cases, the classes getting deleted can just be marked final to tell
the compiler that it can't possibly have subclasses (this also enables the
compiler to generate better code).
Two classes didn't have any sub- or superclasses but virtual functions -
this doesn't make sense, so make all methods of these classes non-virtual.
(Also delete an unused function on one of the two classes.)
In one case, a class actually did have a subclass that needs to be deleted
virtually, so mark one destructor as virtual.
BUG=none
R=bo_xu@foxitsoftware.com
Review URL: https://codereview.chromium.org/370853002
|
|
|
|
|
