Magic-Mirror/gecko-dev
1
0
Fork 0
mirror of https://github.com/mozilla/gecko-dev.git synced 2024-12-12 00:50:40 +00:00
Code Issues Actions Packages Projects Releases Wiki Activity

Bug 1665706: Use length of first DeviceCapabilities array for paper info to allocate space for the others. r=jfkthame

Browse Source 
We rely on the arrays being the same length and now that we allocate double the
amount we need, this should be safe.

Differential Revision: https://phabricator.services.mozilla.com/D90588
This commit is contained in:
Bob Owen 2020-09-17 19:09:50 +00:00
parent 1501ffe928
commit ed17e7db93
1 changed files with 27 additions and 15 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

42
widget/windows/nsPrinterWin.cpp
View File

@ -112,7 +112,8 @@ PrintSettingsInitializer nsPrinterWin::DefaultSettings() const {
template <class T>
static nsTArray<T> GetDeviceCapabilityArray(const LPWSTR aPrinterName,
WORD aCapabilityID) {
WORD aCapabilityID,
size_t aCount = 0) {
nsTArray<T> caps;
// We only want to access printer drivers in the parent process.
@ -120,15 +121,25 @@ static nsTArray<T> GetDeviceCapabilityArray(const LPWSTR aPrinterName,
return caps;
}
// Passing nullptr as the port here seems to work. Given that we would have to
// OpenPrinter with just the name anyway to get the port that makes sense.
// Also, the printer set-up seems to stop you from having two printers with
// the same name.
// Note: this (and the call below) are blocking calls, which could be slow.
int count = ::DeviceCapabilitiesW(aPrinterName, nullptr, aCapabilityID,
nullptr, nullptr);
if (count <= 0) {
return caps;
// Both the call to get the size and the call to actually populate the array
// are relatively expensive, so as sometimes the lengths of the arrays that we
// retrieve depend on each other we allow a count to be passed in to save the
// first call. As we allocate double the count anyway this should allay any
// safety worries.
int count;
if (aCount) {
count = aCount;
} else {
// Passing nullptr as the port here seems to work. Given that we would have
// to OpenPrinter with just the name anyway to get the port that makes
// sense. Also, the printer set-up seems to stop you from having two
// printers with the same name. Note: this (and the call below) are blocking
// calls, which could be slow.
count = ::DeviceCapabilitiesW(aPrinterName, nullptr, aCapabilityID, nullptr,
nullptr);
if (count <= 0) {
return caps;
}
}
// As DeviceCapabilitiesW doesn't take a size, there is a greater risk of the
@ -198,16 +209,17 @@ bool nsPrinterWin::SupportsCollation() const {
}
nsTArray<mozilla::PaperInfo> nsPrinterWin::PaperList() const {
// Paper names are returned in 64 long character buffers.
auto paperNames =
GetDeviceCapabilityArray<Array<wchar_t, 64>>(mName.get(), DC_PAPERNAMES);
// Paper IDs are returned as WORDs.
auto paperIds = GetDeviceCapabilityArray<WORD>(mName.get(), DC_PAPERS);
// Paper names are returned in 64 long character buffers.
auto paperNames = GetDeviceCapabilityArray<Array<wchar_t, 64>>(
mName.get(), DC_PAPERNAMES, paperIds.Length());
// Paper sizes are returned as POINT structs with a tenth of a millimeter as
// the unit.
auto paperSizes = GetDeviceCapabilityArray<POINT>(mName.get(), DC_PAPERSIZE);
auto paperSizes = GetDeviceCapabilityArray<POINT>(mName.get(), DC_PAPERSIZE,
paperIds.Length());
// Check that we have papers and that the array lengths match.
if (!paperNames.Length() || paperNames.Length() != paperIds.Length() ||