unprintable characters in source listing

[LRaim]

When developing a new module I usually request the compiler to generate the source/errors listing. In many cases I found generated warning/error messages likes the following: 

=============================================================

Page 1           Source Listing                                     ifort: warning #10145: no action performed for file 'error'
ifort: warning #10145: no action performed for file 'warn'
7712: This variable has not been used.   [E14] (... here are not printable chars ... )

     &          E11/111/, E12/112/, E13/113/, E14/114/
----------------------------------------------^

========================================================

with non printable characters. This is at  least not elegant.

Such characters are not accepted in submission due to ::

"Your submission contains invalid characters and will not be accepted"

Currently used compiler: COMPILER: Intel(R) Visual Fortran 15.0-1750

[Steven_L_Intel1]

Would you please attach a ZIP of a project that creates such errors?

[LRaim]

Attached is the complete listing of the compilation: file fortout.lst

[Steven_L_Intel1]

I need to be able to compile the source. If you want, use "Send Author A Message" and attach a ZIP of the source file (and any modules or includes it needs) there.

[LRaim]

Here is the source file.

[mecej4]

I suspect that some options to the compiler were used in an incorrect format. File fortout.lst contains the following lines:

ifort: warning #10145: no action performed for file 'error'
ifort: warning #10145: no action performed for file 'warn'

These warnings suggest that "error" and "warn" are missing the "/" prefix, so the compiler driver is trying to process these as file name arguments instead of compiler options.

Please examine your build commands and, if the error persists, post the exact command used (along with all arguments).

I do not have the slightest notion about the origin of notices such as "Such characters are not accepted in submission due to...". Do you use some sort of remote job submission facility?

[Steven_L_Intel1]

What is the exact command line you used to get this listing? (mecej4, it was probably an error from the forum). If I compile this source with /warn:all, I get a bunch of diagnostics (understandable), but no garbage characters. 

[LRaim]

Sorry to disagree.

Message line:
"7712: This variable has not been used.   [E14]"
is repeated at the top of the listing as in a 'summary' section.
In this section is followed by 'dirty' characters.
This can be easily seen by browsing the listing file using a hexadecimal view.

Whichever option I use, the file is generated by the compiler. As I said this is a very minor problem.
Pass the listing to the developers and they will catch the problem at once.

[andrew_4619]

So what causes "Page 1           Source Listing                                     ifort: warning #10145: no action performed for file 'error'
ifort: warning #10145: no action performed for file 'warn'"

at the top of the source listing. I suggest to show the build log for a file compile that creates this source listing. Are there errors/warnings on the command line also?

To identify and fix a problem it is necessary to be able to reproduce it.

[Steven_L_Intel1]

What causes the 10145 warning is not using the right syntax on the command line.

I can reproduce the problem now - thanks. Issue ID is DPD200378600.

[mecej4]

Steve, I see the error, but not always, with IFort 16.0 32-bit. Please the spurious characters after the closing bracket on Lines 6 and 10.

s:\lang>ifort /c /list /warn:unused /Od v0iini54.for 2> nul & findstr "used.*UOPNAME" V0IINI54.lst
v0iini54.for(459): remark #7712: This variable has not been used.   [UOPNAME]
(2) This variable has not been used.   [UOPNAME]

s:\lang>ifort /c /list /warn:unused /Od v0iini54.for 2> nul & findstr "used.*UOPNAME" V0IINI54.lst
v0iini54.for(459): remark #7712: This variable has not been used.   [UOPNAME]a
(2) This variable has not been used.   [UOPNAME]

s:\lang>ifort /c /list /warn:unused /Od v0iini54.for 2> nul & findstr "used.*UOPNAME" V0IINI54.lst
v0iini54.for(459): remark #7712: This variable has not been used.   [UOPNAME]¿
(2) This variable has not been used.   [UOPNAME]

[Steven_L_Intel1]

I can see the error - it was my misreading earlier that prevented it. Interestingly I find it worse in 15.0 than 16.0 but 16.0 still has garbage characters at the end of the closing bracket.  

You are also seeing this in the stderr output? I didn't.

[mecej4]

Steve Lionel (Intel) wrote:

You are also seeing this in the stderr output? I didn't.

No, only in the listing, if one is requested.

[LRaim]

Without opening a new discussion topic, another issue can be identified using the same fortran file.

By deleting the declation of one character variable (e.g. UPRES0) a correct error message is issued, but now messages as:

v0iini54.for(35): remark #7712: This variable has not been used.   [NCARD]
      CHARACTER*8  NCARD,YC,  ACODE,YT
-------------------^

These remark messages are not correct in this context.

Regards

[Steven_L_Intel1]

Interesting - thanks.

[Steven_L_Intel1]

This has been fixed for the next major release, coming out in the second half of 2016.