[m-users.] Signal 253 error on MinGW32 during compilation of module with a large table

[Dirk Ziegemeyer]

> Am 19.02.2015 um 19:47 schrieb Dirk Ziegemeyer <dirk at ziegemeyer.de>:
> 
> Hi Julien,
> 
>> Am 19.02.2015 um 07:15 schrieb Julien Fischer <jfischer at opturion.com>:
>> 
>> The problem is a stack overflow in the Mercury compiler.  Everything is
>> fine in the current rotd builds and I don't think we've adjusted the
>> cstack reserve size we set in compiler/Mmakefile, so presuambly one of
>> the changes between 14.01 and now has fixed this.
>> 
>> I'll try to isolate what has changed, but unless it's relatively small I
>> don't think we'll be able to merge into on the release branch without a
>> lot of fuss.  The best workaround for users encountering this issue at
>> the moment is to use a recent rotd build.
> 
> Thank you for your help.
> I'll try the recent rotd build.

I confirm that the problematic fact table compiles now with rotd-2015-03-10 in a Windows-7/MinGW environment.

It took me some time to migrate my application to rotd, because the improved mode checker showed me some inconsistencies/errors in my code that I had to correct first.

Thank you very much for your assistance.

In addition to that, Julien inspired me to remove my use of pragma fact_table because the support for large tables of facts was improved. I still need to test this with rotd on Windows-7/MinGW and I’ll give feedback for that separately.