Thanks St_W
I had sifted through line endings also, but missed the #define ending.
Either we must type very strictly #defines() --- or use #macro() --- #endmacro instead.
Thank you for all these auto builds.
FBC 1.00.0
Re: FBC 1.00.0
Well done St_W!
Yes adding a trailinge space at the end of the #define line induces the compiler error, but only with fbc version 1.01.0:
- compilation OK for fbc_win32_mingw_0155_2014-09-27.zip 27-Sep-2014 02:09
- compilation NOK for fbc_win32_mingw_0157_2014-09-27.zip 28-Sep-2014 02:00
This revolves around 27-Sep-2014, but I have not found the change that can induce this bad behavior!
Yes adding a trailinge space at the end of the #define line induces the compiler error, but only with fbc version 1.01.0:
- compilation OK for fbc_win32_mingw_0155_2014-09-27.zip 27-Sep-2014 02:09
- compilation NOK for fbc_win32_mingw_0157_2014-09-27.zip 28-Sep-2014 02:00
This revolves around 27-Sep-2014, but I have not found the change that can induce this bad behavior!
Re: FBC 1.00.0
That's already a very good discovery. I tried both build #155 (2014-09-27 01:56:57) and build #156 (2014-09-27 03:41:59) and the latter introduced the compiler error. Thus it's probably something from these two commits:fxm wrote:Yes adding a trailinge space at the end of the #define line induces the compiler error, but only with fbc version 1.01.0:
- compilation OK for fbc_win32_mingw_0155_2014-09-27.zip 27-Sep-2014 02:09
- compilation NOK for fbc_win32_mingw_0157_2014-09-27.zip 28-Sep-2014 02:00
This revolves around 27-Sep-2014, but I have not found the change that can induce this bad behavior!
- Fix self-assignment type mismatch detection https://github.com/freebasic/fbc/commit ... 4434a0f202
- parser: Check for self-assignments when parsing BOPs for disambiguation https://github.com/freebasic/fbc/commit ... 2904d4c54e
That's only one possible reason and there are probably a lot of such reasons, that I currently can't think of.
So now it's dkl's turn :-)
(btw: #156 was not uploaded to freebasic-portal, as only the latest build each morning at 02:00 (GMT+1) is uploaded for space reasons)
Re: FBC 1.00.0
@Admin:St_W wrote:So now it's dkl's turn :-)
Should we file a bug report for that?
(I can do it)
-
- Site Admin
- Posts: 6323
- Joined: Jul 05, 2005 17:32
- Location: Manchester, Lancs
Re: FBC 1.00.0
Good idea.
Re: FBC 1.00.0
Thanks for figuring out the test case that triggered the problem, it was most useful. The problem should be fixed in Git now. Apparently the macro recursion check was broken in combination with parser look-ahead, which is used more extensively since the recent changes in fbc 1.01 expression parsing.
Re: FBC 1.00.0
May be a bit offtopic:
The idea of releasing more often, warmed up again ...
Would it be to big an effort to: release a "bug-fixed" fbc (and maybe fixed "main" lib's,
like fbgfx) at a rate of let's say a 3 month frame, because not everybody here is comfy
with dev-builds ...
As a side effect, the recent changes whould be "more extensively" tested, too.
The idea of releasing more often, warmed up again ...
Would it be to big an effort to: release a "bug-fixed" fbc (and maybe fixed "main" lib's,
like fbgfx) at a rate of let's say a 3 month frame, because not everybody here is comfy
with dev-builds ...
As a side effect, the recent changes whould be "more extensively" tested, too.