[fasmg] small error in line continuation with comments

Index > Compiler Internals > [fasmg] small error in line continuation with comments

Goto page Previous 1, 2

Author

Thread

revolution
When all else fails, read the source

Joined: 24 Aug 2004
Posts: 20530
Location: In your JS exploiting you and your system

revolution 21 Dec 2018, 10:00

Yay, I think I understand.

21 Dec 2018, 10:00

bitRAKE

Joined: 21 Jul 2003
Posts: 4180
Location: vpcmpistri

bitRAKE 21 Dec 2018, 23:13

Would it be accurate to say:

Labeling can only navigate the symbol space at the base of the current namespace. Although, the dot notation allows the creation of a symbolic hierarchy it doesn't change the symbolic base.

Additionally, when entering a namespace we don't know what the last base label was.

Code:

BASE.y:
BASE.r:

namespace BASE
  .x = 1 ; BASE.r.x = 1
end namespace

I'm sure I'll be back to read this again.

_________________
¯\(°_o)/¯ “languages are not safe - uses can be” Bjarne Stroustrup

21 Dec 2018, 23:13

Tomasz Grysztar

Joined: 16 Jun 2003
Posts: 8367
Location: Kraków, Poland

Tomasz Grysztar 16 Feb 2019, 08:28

I'm testing a modification to make this even more compatible with fasm 1. The fasmg engine allows the "last label" to be from a different namespace than the current one, so I could probably force this to use any label that was not started with a dot, regardless of the namespace where its final symbol resides.

The way initially I designed it for fasmg seemed simple and clear from theoretical point of view, but the parts where it differs from fasm 1 are not very usable in practice. So making it backwards-compatible might be a better choice, even if a little bit more clunky in internal implementation.

16 Feb 2019, 08:28

Tomasz Grysztar

Joined: 16 Jun 2003
Posts: 8367
Location: Kraków, Poland

Tomasz Grysztar 16 Feb 2019, 11:54

I'm releasing it, it seems to work well. Is it very likely that the parts of fasmg's specification that were so non-intuitive in practical use have never been actually used by anyone, so this change may in fact not break anything. But please let me know if you find anything that stopped working because of this change.

The compatibility with fasm 1 is now stronger, even though underlying structures are very different:

Code:

BASE:
.x = 1 ; this is BASE.x

BASE.y:
.x = 2 ; this is BASE.y.x

Additionally, I made NAMESPACE command nullify the "last label" setting and END NAMESPACE restore it:

Code:

BASE.x:

namespace BASE
  .y = 1 ; BASE..y = 1
  sub:
  .y = 2 ; BASE.sub.y = 2
end namespace

.y = 3 ; BASE.x.y = 3

This allows to use NAMESPACE block to insert any code without disturbing the flow of labels in the surrounding source.

16 Feb 2019, 11:54

revolution
When all else fails, read the source

Joined: 24 Aug 2004
Posts: 20530
Location: In your JS exploiting you and your system

revolution 16 Feb 2019, 12:12

Tomasz Grysztar wrote:

Code:

namespace BASE
  .y = 1 ; BASE..y = 1

Hmm. That looks weird to me. Those double dots! Confused

16 Feb 2019, 12:12

Lost_Ghost123

Joined: 12 Feb 2019
Posts: 29

Lost_Ghost123 16 Feb 2019, 12:59

This change seems to be for the better. Will try to keep this in mind.

16 Feb 2019, 12:59

Tomasz Grysztar

Joined: 16 Jun 2003
Posts: 8367
Location: Kraków, Poland

Tomasz Grysztar 16 Feb 2019, 13:35

revolution wrote:

Hmm. That looks weird to me. Those double dots!

This is another aspect of how the new world of NAMESPACE needs to cooperate with the legacy "starting dot" syntax. You can ease it on the eyes with help of "#" operator, as in this excerpt from the manual

Quote:

[...] The following example demonstrates the two methods of identifying such symbol:
Code:
        namespace base
                ..other = 1
        end namespace

        result = base.#..other    
The "#" character has been inserted into the last identifier for a better readability, but the plain sequence of three dots would do the same.

16 Feb 2019, 13:35

revolution
When all else fails, read the source

Joined: 24 Aug 2004
Posts: 20530
Location: In your JS exploiting you and your system

revolution 16 Feb 2019, 13:39

It looked like an empty namespace.

Code:

namespace BASE
  .y = 1 ; BASE.<empty>.y = 1

16 Feb 2019, 13:39

Tomasz Grysztar

Joined: 16 Jun 2003
Posts: 8367
Location: Kraków, Poland

Tomasz Grysztar 16 Feb 2019, 14:12

revolution wrote:

It looked like an empty namespace.
Code:
namespace BASE
  .y = 1 ; BASE.<empty>.y = 1    

Yes, you can also understand it that way. To quote the manual again:

Quote:

When an identifier starts with a dot, but there is no label that would be a parent for it, the identifier refers to the descendant of a special symbol that resides in the current namespace but has no name.

Both interpretations conclude the same.

16 Feb 2019, 14:12

Tomasz Grysztar

Joined: 16 Jun 2003
Posts: 8367
Location: Kraków, Poland

Tomasz Grysztar 16 Feb 2019, 17:29

By making it possible to have a current base label that is not from current namespace, this new system allows for some pretty interesting tricks. I'm not going to encourage them, though - because mixing the two systems too much would be most likely not a good habit, considering clarity of the source text.

16 Feb 2019, 17:29

Goto page Previous 1, 2

< Last Thread | Next Thread >

Forum Rules:

You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum
You cannot attach files in this forum
You can download files in this forum