The inline assembler is a surprisingly compact piece of code for what it does, but is overly complex. It's also not used that much, so it has tended to be neglected, and bug reports have slowly accumulated for it. Some were 10 years old.

But, once one figures out how it works, might as well go through the bug list for iasm and see what can be done about them.

I'm happy to report that it has been whittled down to 3:

https://issues.dlang.org/show_bug.cgi?id=12635 - the description is a little vague and I can't determine what the issue actually is

https://issues.dlang.org/show_bug.cgi?id=12878 - this one is a bit complicated to fix, but fortunately isn't critical

https://issues.dlang.org/show_bug.cgi?id=15242 - an enhancement request, not a bug

The list:

https://issues.dlang.org/buglist.cgi?bug_status=NEW&bug_status=ASSIGNED&bug_status=REOPENED&keywords=iasm&keywords_type=allwords&list_id=232881&query_format=advanced

My apologies to those who have been suffering waiting for these fixes.

Happy Assembling!

On Saturday, 29 August 2020 at 06:36:03 UTC, Walter Bright wrote:
> The inline assembler is a surprisingly compact piece of code for what it does, but is overly complex. It's also not used that much, so it has tended to be neglected, and bug reports have slowly accumulated for it. Some were 10 years old.
>
> But, once one figures out how it works, might as well go through the bug list for iasm and see what can be done about them.
>
> I'm happy to report that it has been whittled down to 3:
>
> https://issues.dlang.org/show_bug.cgi?id=12635 - the description is a little vague and I can't determine what the issue actually is
>
> https://issues.dlang.org/show_bug.cgi?id=12878 - this one is a bit complicated to fix, but fortunately isn't critical
>
> https://issues.dlang.org/show_bug.cgi?id=15242 - an enhancement request, not a bug
>
> The list:
>
> https://issues.dlang.org/buglist.cgi?bug_status=NEW&bug_status=ASSIGNED&bug_status=REOPENED&keywords=iasm&keywords_type=allwords&list_id=232881&query_format=advanced
>
> My apologies to those who have been suffering waiting for these fixes.
>
> Happy Assembling!

Thanks, that was a nice initiative.

There's another remaining one actually : https://issues.dlang.org/show_bug.cgi?id=16045.
It's about accessing global variable with -fPIC enabled. I believe this should work using RIP (https://godbolt.org/z/a18vnj) but this is not the case.

Nice!

On Saturday, 29 August 2020 at 06:36:03 UTC, Walter Bright wrote:
> The inline assembler is a surprisingly compact piece of code for what it does, but is overly complex. It's also not used that much, so it has tended to be neglected, and bug reports have slowly accumulated for it. Some were 10 years old.
>
> [...]

Please fix this one as well:
https://issues.dlang.org/show_bug.cgi?id=18826

On Saturday, 29 August 2020 at 06:36:03 UTC, Walter Bright wrote:
> But, once one figures out how it works, might as well go through the bug list for iasm and see what can be done about them.
>
> I'm happy to report that it has been whittled down to 3:
>
> https://issues.dlang.org/show_bug.cgi?id=12635 - the description is a little vague and I can't determine what the issue actually is
>
> https://issues.dlang.org/show_bug.cgi?id=12878 - this one is a bit complicated to fix, but fortunately isn't critical
>
> https://issues.dlang.org/show_bug.cgi?id=15242 - an enhancement request, not a bug
>
> The list:
>
> https://issues.dlang.org/buglist.cgi?bug_status=NEW&bug_status=ASSIGNED&bug_status=REOPENED&keywords=iasm&keywords_type=allwords&list_id=232881&query_format=advanced

I discovered that the VCVTPS2PH instruction is broken a while back, but didn't know to tag my issue "iasm" so it didn't show up in your search:

https://issues.dlang.org/show_bug.cgi?id=20111

On Saturday, 29 August 2020 at 06:36:03 UTC, Walter Bright wrote:
> I'm happy to report that it has been whittled down to 3:
>
> ...
>
> https://issues.dlang.org/buglist.cgi?bug_status=NEW&bug_status=ASSIGNED&bug_status=REOPENED&keywords=iasm&keywords_type=allwords&list_id=232881&query_format=advanced
>
> My apologies to those who have been suffering waiting for these fixes.
>
> Happy Assembling!

I count 13. What'd you say about Bugzilla not being a problem before? That if someone wanted to they could easily find the issues they were looking for? I hope the irony isn't lost on you.

On Sunday, 30 August 2020 at 12:39:10 UTC, Avrina wrote:
> On Saturday, 29 August 2020 at 06:36:03 UTC, Walter Bright wrote:
>> I'm happy to report that it has been whittled down to 3:
>>
>> ...
>>
>> https://issues.dlang.org/buglist.cgi?bug_status=NEW&bug_status=ASSIGNED&bug_status=REOPENED&keywords=iasm&keywords_type=allwords&list_id=232881&query_format=advanced
>>
>> My apologies to those who have been suffering waiting for these fixes.
>>
>> Happy Assembling!
>
> I count 13. What'd you say about Bugzilla not being a problem before? That if someone wanted to they could easily find the issues they were looking for? I hope the irony isn't lost on you.

There was indeed only 3 when he wrote his announce... but bugs are not always reviewed after their creation so often the keywords are missing. For the 10 news, the keyword has been added after the announce.

On Saturday, 29 August 2020 at 06:36:03 UTC, Walter Bright wrote:
> https://issues.dlang.org/show_bug.cgi?id=15242 - an enhancement request, not a bug

FWIW this can be closed, I've needed this in the past but since `intel-intrinsics` the amount of assembly code I maintain has reduced (but wasn't totally eliminated).

On Saturday, 29 August 2020 at 06:36:03 UTC, Walter Bright wrote:
> The inline assembler is a surprisingly compact piece of code for what it does, but is overly complex. It's also not used that much, so it has tended to be neglected, and bug reports have slowly accumulated for it. Some were 10 years old.
>
> [...]
>
> My apologies to those who have been suffering waiting for these fixes.
>
> Happy Assembling!

Thanks Walter. This kind of targeted work aiming to plug all holes in one feature / part of the language is really useful!

On 8/30/2020 5:39 AM, Avrina wrote:
>> https://issues.dlang.org/buglist.cgi?bug_status=NEW&bug_status=ASSIGNED&bug_status=REOPENED&keywords=iasm&keywords_type=allwords&list_id=232881&query_format=advanced 
> I count 13.

That's because people have responded to my post by reclassifying unclassified bugs as iasm bugs. Take a look at the URL above, it only counts bugs with the 'iasm' keyword. Properly classifying bugs when posting them works well.