Patches Proposed So Microsoft Debuggers Can Deal With GCC-Built MinGW Executables

[phoronix]

Phoronix: Patches Proposed So Microsoft Debuggers Can Deal With GCC-Built MinGW Executables

Patches have been proposed for the GCC compiler to ultimately allow MinGW Windows executables to be debugged with Microsoft's debuggers...

https://www.phoronix.com/scan.php?page=news_item&px=GCC-PE-Specific-CodeView

[ddriver]

So the goal here is to bloat debug binaries for the sake of convincing a greedy, privacy intruding, ultra rich, long term monopoly exploiting, anti-competitive corporation?

Who could object to that?

[fkelava]

So the goal here is to bloat debug binaries for the sake of convincing a greedy, privacy intruding, ultra rich, long term monopoly exploiting, anti-competitive corporation?

Who could object to that?

Your point is absolutely valid, but I don't see how this, in particular, is harmful to anyone. MinGW Windows executables- correct me if I'm wrong- seem like something only tangentially (if that much!) related to a vast majority of projects, which will go on as if this switch never existed.

If I had to choose between this change being kept private to Microsoft or included for all in GCC, I'd unhesitatingly pick the latter. The diff also indicates it's small enough to trivially excise, in case the nuclear option has to be exercised someday.

[skeevy420]

Your point is absolutely valid, but I don't see how this, in particular, is harmful to anyone. MinGW Windows executables- correct me if I'm wrong- seem like something only tangentially (if that much!) related to a vast majority of projects, which will go on as if this switch never existed.

If I had to choose between this change being kept private to Microsoft or included for all in GCC, I'd unhesitatingly pick the latter. The diff also indicates it's small enough to trivially excise, in case the nuclear option has to be exercised someday.

The was I see it, it can only be a net benefit to be able to debug Wine, Proton, etc with more debuggers like the native Windows tools themselves. Both GDB and Windows telling a dev "well that could be why it's acting fucky" has to be a good thing...Right???

[Thaodan]

Your point is absolutely valid, but I don't see how this, in particular, is harmful to anyone. MinGW Windows executables- correct me if I'm wrong- seem like something only tangentially (if that much!) related to a vast majority of projects, which will go on as if this switch never existed.

If I had to choose between this change being kept private to Microsoft or included for all in GCC, I'd unhesitatingly pick the latter. The diff also indicates it's small enough to trivially excise, in case the nuclear option has to be exercised someday.

Most cross compile using mingw or mingw-w. For example the Sailfish SDK is build using that.

[fkelava]

The was I see it, it can only be a net benefit to be able to debug Wine, Proton, etc with more debuggers like the native Windows tools themselves. Both GDB and Windows telling a dev "well that could be why it's acting fucky" has to be a good thing...Right???

I agree for a couple of reasons:

• This change, as I understand it, is atomic. Disinterested parties can revert it.
• It doesn't disadvantage existing non-proprietary formats, or entice you to use PDB instead.
• It doesn't fuck with your debug binaries if you don't include the switch.
• The PDB format has had source from the VC++ compiler published under the MIT license. As I understand it, anyone else could've feasibly done the legwork too.
• It lets people accustomed to using Windows debugging tools get to the bottom of quirks in a meaningful way.

It doesn't seem harmful at first, second, or hundredth glance. I'm all for it.

[fkelava]

Most cross compile using mingw or mingw-w. For example the Sailfish SDK is build using that.

I didn't know this was the case. Thank you for clueing me in.

[ddriver]

[/LIST]It doesn't seem harmful at first, second, or hundredth glance. I'm all for it.

Embrace - done
Extend - done
Extinguish - ongoing

Also, how do you skip so many glances, and what could you be missing along?

[fkelava]

Embrace - done
Extend - done
Extinguish - ongoing

Right. In general, I'd entertain the idea, but this is about the change that permits GCC to emit PDB debug information. I am decidedly arguing about this change.

3. Extinguish: When extensions become a de facto standard [...]

The PDB format has had the relevant source from the Visual C++ compiler published six years ago. No source-available shenanigans- it's MIT'd. A format being introduced as an option years after it could have been introduced, with the format itself targeting a specific, targeted subset of situations in which GCC as a whole is used, is rather obviously just another option instead of the next dominant standard. It does not compete with the existing options, devalue them, or disadvantage them.

EEE is a valid concern. EEE is something to be wary of. Changes obviously intended to subvert the ecosystem should be fought against bitterly. But it also does not take a huge amount of scrutiny to deduce that not every change invokes EEE. I maintain that this does not.

[ddriver]

EEE is a valid concern. EEE is something to be wary of. Changes obviously intended to subvert the ecosystem should be fought against bitterly. But it also does not take a huge amount of scrutiny to deduce that not every change invokes EEE. I maintain that this does not.

What matters is the driving intent and the overall direction of the agenda. Even if M$ engages in truly non-harmful discrete contributions, if that is only in the name of gaining trust and traction to conduct harmful actions, it is still ultimately a bad thing.

Call me suspicious, but I strongly doubt M$ is driven by an interest, different from continuing its malignant growth and expanding its ability to extract profit from controlling and selectively stifling progress.

Need I remind you that the company achieved quite the monopoly on actions most people wouldn't identify as harmful, and even identify as beneficial to them?

So maybe the problem is not in the count of glances you make, but that it takes more of an in-depth examination than a superficial momentary glance.