Re: [mpir-devel] cannot compile mpir on windows xp - several errors

[Bill Hart]

Hi,

is this using Cygwin, MinGW or MSVC?

If you are getting random hangs, it could be due to an antivirus program or an old version of MinGW.

Bill.

On 2 May 2013 12:51, developerchris <cha...@gmail.com> wrote:

The main error is a particularly odd one

when using the emerge system to compile mpir on windows xp it just hangs at seemingly random places while configuring yasm

Sometimes it will fail complaining that python is too low a version. it requires > 2.4 and my version is only 3.3

I am pretty sure the fault is the test in yasm\configure

line 7890 and line 7921 declare sys.exit(sys.hexversion < minverhex)

I think that needs to be sys.exit(sys.hexversion > minverhex)

Not sure how this could ever compile with that bad test in there?

Still that is fixable. the strange hangs leave me lost I dont know where to look. the various make logs don't shed any light on the issue at all.

It just hung once again at this point

config.status: linking /r/build/win32libs/mpir-20120201/work/mpir-2.5.0/mpn/generic/mod_1_2.c to mpn/mod_1_2.c

config log shows
config.status:1647: linking /r/build/win32libs/mpir-20120201/work/mpir-2.5.0/mpn/generic/mod_1_2.c to mpn/mod_1_2.c

This is after linking many files.

more often it will hang at configure: creating ./config.status

When it hangs I have to kill the shell its running in ^C does not work

Any ideas what I can do to track this issue down?




DC

--
You received this message because you are subscribed to the Google Groups "mpir-devel" group.
To unsubscribe from this group and stop receiving emails from it, send an email to mpir-devel+...@googlegroups.com.
To post to this group, send email to mpir-...@googlegroups.com.
Visit this group at http://groups.google.com/group/mpir-devel?hl=en.
For more options, visit https://groups.google.com/groups/opt_out.

[gmail]

I am using mingw 4.4.

I disabled virus protection

this time it froze at
configure: creating ./config.status

the config.log ends at
configure:28571: creating ./config.status

I'll try and compile it outside of the kde environment and see what happens...

DC

[gmail]

When attempting to configure using the latest mingw install i get the following error...

checking for assembler local label prefix... L
checking for assembler byte directive... .byte
checking how to define a 32-bit word... configure: error: cannot determine how to define a 32-bit word

At that point configure quits

DC


On 02/05/2013 9:54 PM, Bill Hart wrote:

[Bill Hart]

I'm afraid I haven't seen this before. If it is randomly doing this, then there must be something wrong with the environment. Sorry I don't think I have any specific advice.

Bill.

[Chris Hamono]

I have realised that mpir is as far as I am aware the only project I have compiled so far that uses configure. At least the only one that has caused issues thus bringing it to my attention.

It may be an issue with msys and sh.exe although the last isssue (cannot determine how to define a 32-bit word) appears to be a compiler issue. its probably the test thats bad

With the python detection issue. I realised that configure may be using the exit value the opposite to the way I assumed. i.e. exit code 0 is actually a successful test.

Is that the case?

If so I need to workout why python detection fails.

[Bill Hart]

Hi Chris,

configure is automatically generated by autotools from configure.in. There is no need to go digging through the tests that it runs. If configure doesn't run on your machine, then something is very wrong. 

Bill.

[sisy...@optusnet.com.au]

From: gmail

> checking for assembler local label prefix... L
> checking for assembler byte directive... .byte
> checking how to define a 32-bit word... configure: error: cannot determine
> how to define a 32-bit word
>
> At that point configure quits

There should be an error message in the config.log that throws some light on
why the determination failed. (You need to find out why it failed.)

Search the config.log for the string "checking how to define a 32-bit word"
and read over the next few lines until you come to the actual error
message.

Cheers,
Rob

[gmail]

As far as I recall there was nothing in the logs that indicated what the
error was with the 32 bit word test. it just stopped at the line that said
it was testing it. I have tried rebuilding again so the log was deleted. so
I cant be 100% sure.

By running in safe mode and totally disabling Microsoft security essentials
(as opposed to turning off realtime protection) I was once again able to get
to the python check

That still fails.

So microsofts antivirus is the most likely culprit although some other
service may be the cause.

I just need to track down why the python test fails


Damn! Spoke too soon. I tried running it again and it hung at creating
config.status

The file appears to hang after partially writing config.status
the next line is
cat >>$CONFIG_STATUS <<_ACEOF || ac_write_fail=1

It has to be an issue with msys/bin/sh.exe it must be failing at random
times for random reasons like returning the wrong exit code from python



Chris

[Bill Hart]

Either way, this just doesn't sound like an MPIR problem but a MinGW/msys one.

You might want to ask on their list about random stalls.

On later versions of Windows you must disable all antivirus software and run the MinGW/msys shell in administrator mode to avoid this. But to my knowledge the latter happens automatically on XP, which is much less secure by default.

Bill.

[developerchris]

I managed to get mpir to compile.

The sh issue is either as you say a problem with msys or something peculiar to my machine. I could get past it by locating the rogue sh process and killing it. I am working on resolving the cause of it.

The problem with python failing the version test magically disappeared. I have no idea what that means! but once I found how to work past the sh problem the python tests passed.

I really appreciate your patience with me It helped.

DC