* A user is reporting that, on one of their platforms, Rufus is writing to the wrong target during the file-copy
phase and using their existing Y: local drive instead of the drive associated to the USB, despite the fact
that Rufus is passing the right volume name to GetVolumePathNamesForVolumeName().
* Here's the PowerShell wmic output, confirming that the volume GUID obtained by Rufus is the right one:
DriveLetter : Y:
DeviceId : \\?\Volume{000349b1-17d0-69f6-c13f-f31162930600}\
Capacity : 118540464128
FileSystem : NTFS
Label : Y-DISK
DriveLetter : H:
DeviceId : \\?\Volume{b150ff4a-d62b-11ea-86e3-f49634660e54}\
Capacity : 15791824896
FileSystem : FAT32
Label : ADATA16GB
* And here's the Rufus log demonstrating that GetVolumePathNamesForVolumeName() is returning the *WRONG* letter:
Found volume \\?\Volume{b150ff4a-d62b-11ea-86e3-f49634660e54}\
\\?\Volume{b150ff4a-d62b-11ea-86e3-f49634660e54}\ is already mounted as Y: instead of H: - Will now use this target instead...
* The last line shows, without the shadow of a doubt, that we did feed "\\?\Volume{b150ff4a-d62b-11ea-86e3-f49634660e54}\" to
GetVolumePathNamesForVolumeName() and that this API call was successful (returned a non zero size) but ultimately returned
the wrong letter (Y: instead of H:)...
* Therefore, Windows is BUGGY and the use of GetVolumePathNamesForVolumeName() must be avoided.
* ISOs with tons of Rock Ridge deep directory entries (such as OPNsense)
can be very slow to scan due to the nature of deep directory parsing,
which requires processing the whole ISO9660 fs, for each deep directory
file, in order to find the relevant LSN entry.
* Since we don't expect much of the content we care about to reside in a
deep directory entry, we amend the code to cut short the scan of any
directory that contains such elements.
* Note that this only applies for ISO scan and it does nothing to speed
up the ISO extraction process.
* Related to issue #1575
* Closes#1467
* Also ensure that previous element is set when repositioning a
control to preserve tabbing order
* Also fix x64 version being able to download x86 BETA
* This is mostly aimed at Debian 11 netinst on the Raspberry Pi 4
* Only available for regular UEFI ISOs if GPT and FAT are selected (no MBR ESPs).
* Also fix a MinGW warning in GetUnusedDriveLetter()
* Now use version 1.6 of the EfiFs drivers that enables firmwares that
don't provide EFI_DEVICE_PATH_TO_TEXT_PROTOCOL to mount NTFS partitions
regardless.
* Also use the latest version of UEFI:NTFS that displays additional info
about the system.
* Closes#1213.
* Also update version to Rufus-next.
* ASLR is enabled by default for Visual Studio builds but that isn't the case
for MinGW builds. Fix that and also add -Wformat-security while we're at it.
* Closes#1518
* Also ensure that we'll never write protective MBR message for non-bootable
GPT drives, even as we are not calling WriteSBR() for those anyway.
* Also fix SBR message not being written for bootable images
* Also add an extra partition refresh after deleting partitions to try
to further force Windows take its stinking paws off our drive.
* Fix RTL location of "ISO" in the "Copying ISO files" translation for Arabic and Persian
* Fix whitespace/message ending issues for various translations
* Sync all .po's with .loc to avoid another German progress update is missing issue
* We distractedly chose to populate the message from our protective MBR
for GPT/UEFI-only boot media into the 4KB that directly followed the
MBR, which of course is space that is being used by the primary GPT.
* This resulted on systems having to fall back to using the secondary
GPT, which not all appear to be designed to do.
* Alter the code to ensure the protective message is written at LBA 34,
after the primary GPT.
* Closes#1507
* If a converted label contains mostly underscore, the proposed
label is used for FAT32 instead. However this label still has
the KB/MB/GB symbols localized so it may be invalid.
* Ensure that we use a non-localized version of the size when
using such a label.
* Closes#1506.
* Also fix a VS2019 static analysis warning in net.c.
* The upcoming Ubuntu 20.04 comes with MD5 validation turned on by default.
* When creating persistent boot media, we may update some of the validated files
to add persistence, update the search labels, etc.
* Make sure that the files we modify get their MD5 updated where needed.
* Also add 'loopback.cfg' to the list of config files we can add persistence to.
* Part of #1499
* Among other nefarious things, ubuntu 20.04 added a $casper_flavour suffix
to their grub.cfg /casper/vmlinuz kernel option, so we can no longer rely
on 'persistent' being inserted in a proper location.
* Switch to latching on file=/cdrom/preseed and hope that it will work for
all of Ubuntu & derivatives.
* Part of #1499.
* Commit 4c5adf092e moved us away from using CreateFile()
when extracting a file on the target media, and as such the error code returned when
failing to create an 'autorun.inf' due to a security solution has shifted.
* Make sure we handle the new error and don't bail out on 'autorun.inf' creation.
* Also update the actual name of the RtlDosPathNameToNtPathNameXXX function we use.
* Closes#1496
* Recent versions of Windows can set the deafult locale to codepage 65001 (UTF-8).
* This produces an assert due to a missing entry in cp_hr_list[], so fix that.
* However, this fix alone is not enough, as a GetOEMCP() that returns 65001 means
that any systems set to UTF-8 will fall back to codepage 437 for DOS, which is
definitely not what we want => Add an extra call to determine the actual OEM
codepage when UTF-8 is detected.
* Closes#1468
* Commit [e522ef6c55] (PR #1426) regressed the '%s'
progress messages back to '%0.1f%%' which results in the percentage remaining at
zero when the UI is in German.
* Surround macro params to ensure expected results
* Fix copy-paste errors
* Fix a potential buffer overflow in SetSectionHeaders()
* Add const modifier where relevant
* Use GetWindowLongPtr() everywhere
* Use proper sprintf format for unsigned int
* Use %s for printf-like funcs (https://www.viva64.com/en/w/v618/print/)
* Closes#1464
* Status code assignation was removed when the original code
was altered to use pfNtFsControlFile(). Fix that and also
make the code more similar to other calls.
* Closes#1459
* msg.S now reads an ASCII message (with escaped colour sequences)
from the following blocks, which is both more flexible and allows
for more content to be displayed.
* Also adds Bochs testing to the MBR build facility
* Hopefully using DICS_FLAG_CONFIGSPECIFIC instead of DICS_FLAG_GLOBAL is all that was needed
to get device disabling/re-enabling work without creating zombie devices, because we sure
need to force Windows' hand when it comes to detecting logical volumes...
* Implement CreatePreallocatedFile() which uses NtCreateFile() to create files with preallocated sizes.
This is used during ISO extraction to improve performance.
* Remove now-unused preallocate_filesize which was called after CreateFileU().
* Closes#1445
* ClearMBRGPT() attempts to write WRITE_RETRIES times, even if all those times succeed.
* Instead, skip the remaining retries on success.
* Also improve code readability.
* Closes#1454
* So, as it happens, when assigning the product of two 32-bit variables into a 64-bit one,
compilers default to being *DUMB* and, against all reasonable expectations, do not perform
that multiplication as a 64-bit operation (even when the code is compiled as x64). Wow,
that's really great decision making by compiler designers if I ever saw some... Whoever
decided that C developers would much rather want truncation and 32-bit overflows, instead
of the expected *LOGICAL* behaviour of conducting arithmetic operations as 64-bit when the
result will be assigned to a 64-bit variable, need to be condemned to a lifetime of trying
to help elderly folks trying to conduct simple computing tasks as a punishment...
Anyhoo, nt_write_blk()'s offset.QuadPart = block * channel->block_size + nt_data->offset
was overflowing 32-bit as soon as block * channel->block_size went over the 4 GB mark,
with the disastrous results one can expect. Considering that this is code we practically
lifted verbatim from e2fsprogs, I guess e2fsprogs' NT I/O manager was never properly
tested with anything larger than a 4 GB. Awesome!
* We fix the above by doing what unix_io.c does and setting the 32-bit read/write_blk()
calls to be wrappers around their 64-bit counterpart (since, once you deal with a 64-bit
block variable, the computation is conducted as 64-bit).
* Also remove a bunch of stuff we don't need from config.h
* Closes#1396
* Fix use of EXT2_BLOCK_SIZE() instead of EXT2_INODE_SIZE() during inode
initialization, that made us zero way many more blocks than was needed.
* Also disable sparse_super feature and improve block setup.
* Also explicitly use IS_POWER_OF_2 macro where required.
* Most distros (Debian, Ubuntu) have moved to using Sylinus 6.04 even
as it has NOT officially been released, so we want our fallback to
work against this too.
* pre1 since the Syslinux folks advise against using pre2 or later...
* Closes#1444
