comp.lang.ada
 help / color / mirror / Atom feed
From: Simon Wright <simon@pushface.org>
Subject: Re: unknown error mesage
Date: Wed, 17 Mar 2021 12:06:50 +0000	[thread overview]
Message-ID: <ly35wuoug5.fsf@pushface.org> (raw)
In-Reply-To: 6051b1b4$0$31158$e4fe514c@news.kpn.nl

ldries46 <bertus.dries@planet.nl> writes:

> Op 16-3-2021 om 19:21 schreef John McCabe:
>> On Tue, 16 Mar 2021 18:49:21 +0100, ldries46 <bertus.dries@planet.nl>
>> wrote:
>>
>>> fatal error: file gtk.ali is incorrectly formatted
>>> make sure you are using consistent versions of gcc/gnatbind

>>> I have already deleted all .ali files in the build directory, deleted
>>> the Datum.exe file, reinstalled GtkAda en even reinstalled GNAT Studio
>>> Community version 2020.
>>>
>>> Who can tell me what is message means and how to eliminate that
>> I presume you've read this question with a similar issue?
>>
>> https://stackoverflow.com/questions/7576994/ada-strange-error-consistent-version-of-gcc-gnatbind
>>
>> The suggestion from that would appear to be that your gtk may have
>> been built with a different version of GNAT, hence you might need to
>> recompile the GtkAda library with whatever version of GNAT you've now
>> got.
>>
>> However, did you download the GtkAda version matching the GNAT
>> Community version from here - https://www.adacore.com/download/more?

> Yesterday I downloaded and installed GtkAda and GNAT Studio again from
> the site you mentioned the error did not disappear.
> In the mean time I tried to rebuild another project and it created a
> simular problem but with glib.ali
> I read the stack overflow item but I think that is not the problem.
> For reasons of tracebility I did not build GtkAda in the original
> directory but in C:/GNAT/GtkAda.
> The last uodate of the OS was only a few days ago but after that it
> still worhed

The complaint about mismatched versions is almost certainly derived from
the version info in the first line of an ALI file.

Here I see (from a collection of old StackOverflow-related
investigations)

First line              Compiler
-----------------       ------------
V "GNAT Lib v10"        FSF GCC 10
V "GNAT Lib v11"        FSF GCC 11
V "GNAT Lib v2019"      GNAT CE 2019
V "GNAT Lib v2020"      GNAT CE 2020
V "GNAT Lib v7"         FSF GCC 7
V "GNAT Lib v8"         FSF GCC 8
V "GNAT Lib v9"         FSF GCC 9

Using a wrong ADA_PROJECT_PATH/GPR_PROJECT_PATH can lead to this. Using
gprbuild -v can help.

Also,

* Compile unit U with A.gpr which puts the ALI in directory A_OBJ
* Move unit U into the source tree for B.gpr which put the ALI in
  directory B_OBJ
* Run gprclean on A.gpr, which doesn't delete U.ali because U is no logner
  one of its sources
* Change compiler!

  reply	other threads:[~2021-03-17 12:06 UTC|newest]

Thread overview: 12+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-03-16 17:49 unknown error mesage ldries46
2021-03-16 18:21 ` John McCabe
2021-03-17  7:37   ` ldries46
2021-03-17 12:06     ` Simon Wright [this message]
2021-03-17 13:17       ` ldries46
2021-03-17 20:29         ` Simon Wright
2021-03-18  6:08         ` ldries46
2021-03-18  9:11         ` ldries46
2021-03-18 11:40           ` Stephen Leake
2021-03-18 13:55             ` ldries46
2021-03-18 17:34           ` Simon Wright
2021-03-18 10:54         ` ldries46
replies disabled

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox